Generated v2 prefixes

This commit is contained in:
geky-bot
2023-05-04 18:31:30 +00:00
67 changed files with 26771 additions and 5203 deletions
+1430
View File
File diff suppressed because it is too large Load Diff
+181
View File
@@ -0,0 +1,181 @@
#!/usr/bin/env python3
#
# Change prefixes in files/filenames. Useful for creating different versions
# of a codebase that don't conflict at compile time.
#
# Example:
# $ ./scripts/changeprefix.py lfs2 lfs23
#
# Copyright (c) 2022, The littlefs authors.
# Copyright (c) 2019, Arm Limited. All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
#
import glob
import itertools
import os
import os.path
import re
import shlex
import shutil
import subprocess
import tempfile
GIT_PATH = ['git']
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def changeprefix(from_prefix, to_prefix, line):
line, count1 = re.subn(
'\\b'+from_prefix,
to_prefix,
line)
line, count2 = re.subn(
'\\b'+from_prefix.upper(),
to_prefix.upper(),
line)
line, count3 = re.subn(
'\\B-D'+from_prefix.upper(),
'-D'+to_prefix.upper(),
line)
return line, count1+count2+count3
def changefile(from_prefix, to_prefix, from_path, to_path, *,
no_replacements=False):
# rename any prefixes in file
count = 0
# create a temporary file to avoid overwriting ourself
if from_path == to_path and to_path != '-':
to_path_temp = tempfile.NamedTemporaryFile('w', delete=False)
to_path = to_path_temp.name
else:
to_path_temp = None
with openio(from_path) as from_f:
with openio(to_path, 'w') as to_f:
for line in from_f:
if not no_replacements:
line, n = changeprefix(from_prefix, to_prefix, line)
count += n
to_f.write(line)
if from_path != '-' and to_path != '-':
shutil.copystat(from_path, to_path)
if to_path_temp:
os.rename(to_path, from_path)
elif from_path != '-':
os.remove(from_path)
# Summary
print('%s: %d replacements' % (
'%s -> %s' % (from_path, to_path) if not to_path_temp else from_path,
count))
def main(from_prefix, to_prefix, paths=[], *,
verbose=False,
output=None,
no_replacements=False,
no_renames=False,
git=False,
no_stage=False,
git_path=GIT_PATH):
if not paths:
if git:
cmd = git_path + ['ls-tree', '-r', '--name-only', 'HEAD']
if verbose:
print(' '.join(shlex.quote(c) for c in cmd))
paths = subprocess.check_output(cmd, encoding='utf8').split()
else:
print('no paths?', file=sys.stderr)
sys.exit(1)
for from_path in paths:
# rename filename?
if output:
to_path = output
elif no_renames:
to_path = from_path
else:
to_path = os.path.join(
os.path.dirname(from_path),
changeprefix(from_prefix, to_prefix,
os.path.basename(from_path))[0])
# rename contents
changefile(from_prefix, to_prefix, from_path, to_path,
no_replacements=no_replacements)
# stage?
if git and not no_stage:
if from_path != to_path:
cmd = git_path + ['rm', '-q', from_path]
if verbose:
print(' '.join(shlex.quote(c) for c in cmd))
subprocess.check_call(cmd)
cmd = git_path + ['add', to_path]
if verbose:
print(' '.join(shlex.quote(c) for c in cmd))
subprocess.check_call(cmd)
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Change prefixes in files/filenames. Useful for creating "
"different versions of a codebase that don't conflict at compile "
"time.",
allow_abbrev=False)
parser.add_argument(
'from_prefix',
help="Prefix to replace.")
parser.add_argument(
'to_prefix',
help="Prefix to replace with.")
parser.add_argument(
'paths',
nargs='*',
help="Files to operate on.")
parser.add_argument(
'-v', '--verbose',
action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument(
'-o', '--output',
help="Output file.")
parser.add_argument(
'-N', '--no-replacements',
action='store_true',
help="Don't change prefixes in files")
parser.add_argument(
'-R', '--no-renames',
action='store_true',
help="Don't rename files")
parser.add_argument(
'--git',
action='store_true',
help="Use git to find/update files.")
parser.add_argument(
'--no-stage',
action='store_true',
help="Don't stage changes with git.")
parser.add_argument(
'--git-path',
type=lambda x: x.split(),
default=GIT_PATH,
help="Path to git executable, may include flags. "
"Defaults to %r." % GIT_PATH)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
+649 -226
View File
@@ -1,42 +1,188 @@
#!/usr/bin/env python3
#
# Script to find code size at the function level. Basically just a bit wrapper
# Script to find code size at the function level. Basically just a big wrapper
# around nm with some extra conveniences for comparing builds. Heavily inspired
# by Linux's Bloat-O-Meter.
#
# Example:
# ./scripts/code.py lfs2.o lfs2_util.o -Ssize
#
# Copyright (c) 2022, The littlefs authors.
# Copyright (c) 2020, Arm Limited. All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
#
import os
import glob
import itertools as it
import subprocess as sp
import shlex
import re
import csv
import collections as co
import csv
import difflib
import itertools as it
import math as m
import os
import re
import shlex
import subprocess as sp
OBJ_PATHS = ['*.o']
NM_PATH = ['nm']
NM_TYPES = 'tTrRdD'
OBJDUMP_PATH = ['objdump']
def collect(paths, **args):
results = co.defaultdict(lambda: 0)
pattern = re.compile(
# integer fields
class Int(co.namedtuple('Int', 'x')):
__slots__ = ()
def __new__(cls, x=0):
if isinstance(x, Int):
return x
if isinstance(x, str):
try:
x = int(x, 0)
except ValueError:
# also accept +-∞ and +-inf
if re.match('^\s*\+?\s*(?:∞|inf)\s*$', x):
x = m.inf
elif re.match('^\s*-\s*(?:∞|inf)\s*$', x):
x = -m.inf
else:
raise
assert isinstance(x, int) or m.isinf(x), x
return super().__new__(cls, x)
def __str__(self):
if self.x == m.inf:
return ''
elif self.x == -m.inf:
return '-∞'
else:
return str(self.x)
def __int__(self):
assert not m.isinf(self.x)
return self.x
def __float__(self):
return float(self.x)
none = '%7s' % '-'
def table(self):
return '%7s' % (self,)
diff_none = '%7s' % '-'
diff_table = table
def diff_diff(self, other):
new = self.x if self else 0
old = other.x if other else 0
diff = new - old
if diff == +m.inf:
return '%7s' % '+∞'
elif diff == -m.inf:
return '%7s' % '-∞'
else:
return '%+7d' % diff
def ratio(self, other):
new = self.x if self else 0
old = other.x if other else 0
if m.isinf(new) and m.isinf(old):
return 0.0
elif m.isinf(new):
return +m.inf
elif m.isinf(old):
return -m.inf
elif not old and not new:
return 0.0
elif not old:
return 1.0
else:
return (new-old) / old
def __add__(self, other):
return self.__class__(self.x + other.x)
def __sub__(self, other):
return self.__class__(self.x - other.x)
def __mul__(self, other):
return self.__class__(self.x * other.x)
# code size results
class CodeResult(co.namedtuple('CodeResult', [
'file', 'function',
'size'])):
_by = ['file', 'function']
_fields = ['size']
_sort = ['size']
_types = {'size': Int}
__slots__ = ()
def __new__(cls, file='', function='', size=0):
return super().__new__(cls, file, function,
Int(size))
def __add__(self, other):
return CodeResult(self.file, self.function,
self.size + other.size)
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def collect(obj_paths, *,
nm_path=NM_PATH,
nm_types=NM_TYPES,
objdump_path=OBJDUMP_PATH,
sources=None,
everything=False,
**args):
size_pattern = re.compile(
'^(?P<size>[0-9a-fA-F]+)' +
' (?P<type>[%s])' % re.escape(args['type']) +
' (?P<type>[%s])' % re.escape(nm_types) +
' (?P<func>.+?)$')
for path in paths:
# note nm-tool may contain extra args
cmd = args['nm_tool'] + ['--size-sort', path]
line_pattern = re.compile(
'^\s+(?P<no>[0-9]+)'
'(?:\s+(?P<dir>[0-9]+))?'
'\s+.*'
'\s+(?P<path>[^\s]+)$')
info_pattern = re.compile(
'^(?:.*(?P<tag>DW_TAG_[a-z_]+).*'
'|.*DW_AT_name.*:\s*(?P<name>[^:\s]+)\s*'
'|.*DW_AT_decl_file.*:\s*(?P<file>[0-9]+)\s*)$')
results = []
for path in obj_paths:
# guess the source, if we have debug-info we'll replace this later
file = re.sub('(\.o)?$', '.c', path, 1)
# find symbol sizes
results_ = []
# note nm-path may contain extra args
cmd = nm_path + ['--size-sort', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
errors='replace',
close_fds=False)
for line in proc.stdout:
m = pattern.match(line)
m = size_pattern.match(line)
if m:
results[(path, m.group('func'))] += int(m.group('size'), 16)
func = m.group('func')
# discard internal functions
if not everything and func.startswith('__'):
continue
results_.append(CodeResult(
file, func,
int(m.group('size'), 16)))
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
@@ -44,241 +190,518 @@ def collect(paths, **args):
sys.stdout.write(line)
sys.exit(-1)
flat_results = []
for (file, func), size in results.items():
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
# discard internal functions
if not args.get('everything'):
if func.startswith('__'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
flat_results.append((file, func, size))
# try to figure out the source file if we have debug-info
dirs = {}
files = {}
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=rawline', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace',
close_fds=False)
for line in proc.stdout:
# note that files contain references to dirs, which we
# dereference as soon as we see them as each file table follows a
# dir table
m = line_pattern.match(line)
if m:
if not m.group('dir'):
# found a directory entry
dirs[int(m.group('no'))] = m.group('path')
else:
# found a file entry
dir = int(m.group('dir'))
if dir in dirs:
files[int(m.group('no'))] = os.path.join(
dirs[dir],
m.group('path'))
else:
files[int(m.group('no'))] = m.group('path')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
# do nothing on error, we don't need objdump to work, source files
# may just be inaccurate
pass
return flat_results
defs = {}
is_func = False
f_name = None
f_file = None
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=info', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace',
close_fds=False)
for line in proc.stdout:
# state machine here to find definitions
m = info_pattern.match(line)
if m:
if m.group('tag'):
if is_func:
defs[f_name] = files.get(f_file, '?')
is_func = (m.group('tag') == 'DW_TAG_subprogram')
elif m.group('name'):
f_name = m.group('name')
elif m.group('file'):
f_file = int(m.group('file'))
if is_func:
defs[f_name] = files.get(f_file, '?')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
# do nothing on error, we don't need objdump to work, source files
# may just be inaccurate
pass
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
for r in results_:
# find best matching debug symbol, this may be slightly different
# due to optimizations
if defs:
# exact match? avoid difflib if we can for speed
if r.function in defs:
file = defs[r.function]
else:
_, file = max(
defs.items(),
key=lambda d: difflib.SequenceMatcher(None,
d[0],
r.function, False).ratio())
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
file = r.file
# find sizes
if not args.get('use', None):
# find .o files
paths = []
for path in args['obj_paths']:
if os.path.isdir(path):
path = path + '/*.o'
# ignore filtered sources
if sources is not None:
if not any(
os.path.abspath(file) == os.path.abspath(s)
for s in sources):
continue
else:
# default to only cwd
if not everything and not os.path.commonpath([
os.getcwd(),
os.path.abspath(file)]) == os.getcwd():
continue
for path in glob.glob(path):
paths.append(path)
# simplify path
if os.path.commonpath([
os.getcwd(),
os.path.abspath(file)]) == os.getcwd():
file = os.path.relpath(file)
else:
file = os.path.abspath(file)
if not paths:
print('no .obj files found in %r?' % args['obj_paths'])
results.append(r._replace(file=file))
return results
def fold(Result, results, *,
by=None,
defines=None,
**_):
if by is None:
by = Result._by
for k in it.chain(by or [], (k for k, _ in defines or [])):
if k not in Result._by and k not in Result._fields:
print("error: could not find field %r?" % k)
sys.exit(-1)
results = collect(paths, **args)
# filter by matching defines
if defines is not None:
results_ = []
for r in results:
if all(getattr(r, k) in vs for k, vs in defines):
results_.append(r)
results = results_
# organize results into conflicts
folding = co.OrderedDict()
for r in results:
name = tuple(getattr(r, k) for k in by)
if name not in folding:
folding[name] = []
folding[name].append(r)
# merge conflicts
folded = []
for name, rs in folding.items():
folded.append(sum(rs[1:], start=rs[0]))
return folded
def table(Result, results, diff_results=None, *,
by=None,
fields=None,
sort=None,
summary=False,
all=False,
percent=False,
**_):
all_, all = all, __builtins__.all
if by is None:
by = Result._by
if fields is None:
fields = Result._fields
types = Result._types
# fold again
results = fold(Result, results, by=by)
if diff_results is not None:
diff_results = fold(Result, diff_results, by=by)
# organize by name
table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in results}
diff_table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in diff_results or []}
names = list(table.keys() | diff_table.keys())
# sort again, now with diff info, note that python's sort is stable
names.sort()
if diff_results is not None:
names.sort(key=lambda n: tuple(
types[k].ratio(
getattr(table.get(n), k, None),
getattr(diff_table.get(n), k, None))
for k in fields),
reverse=True)
if sort:
for k, reverse in reversed(sort):
names.sort(
key=lambda n: tuple(
(getattr(table[n], k),)
if getattr(table.get(n), k, None) is not None else ()
for k in ([k] if k else [
k for k in Result._sort if k in fields])),
reverse=reverse ^ (not k or k in Result._fields))
# build up our lines
lines = []
# header
header = []
header.append('%s%s' % (
','.join(by),
' (%d added, %d removed)' % (
sum(1 for n in table if n not in diff_table),
sum(1 for n in diff_table if n not in table))
if diff_results is not None and not percent else '')
if not summary else '')
if diff_results is None:
for k in fields:
header.append(k)
elif percent:
for k in fields:
header.append(k)
else:
for k in fields:
header.append('o'+k)
for k in fields:
header.append('n'+k)
for k in fields:
header.append('d'+k)
header.append('')
lines.append(header)
def table_entry(name, r, diff_r=None, ratios=[]):
entry = []
entry.append(name)
if diff_results is None:
for k in fields:
entry.append(getattr(r, k).table()
if getattr(r, k, None) is not None
else types[k].none)
elif percent:
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
else:
for k in fields:
entry.append(getattr(diff_r, k).diff_table()
if getattr(diff_r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(types[k].diff_diff(
getattr(r, k, None),
getattr(diff_r, k, None)))
if diff_results is None:
entry.append('')
elif percent:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios))
else:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios
if t)
if any(ratios) else '')
return entry
# entries
if not summary:
for name in names:
r = table.get(name)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = diff_table.get(name)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
if not all_ and not any(ratios):
continue
lines.append(table_entry(name, r, diff_r, ratios))
# total
r = next(iter(fold(Result, results, by=[])), None)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = next(iter(fold(Result, diff_results, by=[])), None)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
lines.append(table_entry('TOTAL', r, diff_r, ratios))
# find the best widths, note that column 0 contains the names and column -1
# the ratios, so those are handled a bit differently
widths = [
((max(it.chain([w], (len(l[i]) for l in lines)))+1+4-1)//4)*4-1
for w, i in zip(
it.chain([23], it.repeat(7)),
range(len(lines[0])-1))]
# print our table
for line in lines:
print('%-*s %s%s' % (
widths[0], line[0],
' '.join('%*s' % (w, x)
for w, x in zip(widths[1:], line[1:-1])),
line[-1]))
def main(obj_paths, *,
by=None,
fields=None,
defines=None,
sort=None,
**args):
# find sizes
if not args.get('use', None):
results = collect(obj_paths, **args)
else:
results = []
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['code_size']))
for result in r
if result.get('code_size') not in {None, ''}]
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('code_'+k in r and r['code_'+k].strip()
for k in CodeResult._fields):
continue
try:
results.append(CodeResult(
**{k: r[k] for k in CodeResult._by
if k in r and r[k].strip()},
**{k: r['code_'+k] for k in CodeResult._fields
if 'code_'+k in r and r['code_'+k].strip()}))
except TypeError:
pass
total = 0
for _, _, size in results:
total += size
# fold
results = fold(CodeResult, results, by=by, defines=defines)
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['code_size']))
for result in r
if result.get('code_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
prev_total += size
# sort, note that python's sort is stable
results.sort()
if sort:
for k, reverse in reversed(sort):
results.sort(
key=lambda r: tuple(
(getattr(r, k),) if getattr(r, k) is not None else ()
for k in ([k] if k else CodeResult._sort)),
reverse=reverse ^ (not k or k in CodeResult._fields))
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('code_size', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, size in results:
merged_results[(file, func)]['code_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'code_size'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
writer = csv.DictWriter(f,
(by if by is not None else CodeResult._by)
+ ['code_'+k for k in (
fields if fields is not None else CodeResult._fields)])
writer.writeheader()
for r in results:
writer.writerow(
{k: getattr(r, k) for k in (
by if by is not None else CodeResult._by)}
| {'code_'+k: getattr(r, k) for k in (
fields if fields is not None else CodeResult._fields)})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, func, size in results:
entry = (file if by == 'file' else func)
entries[entry] += size
return entries
# find previous results?
if args.get('diff'):
diff_results = []
try:
with openio(args['diff']) as f:
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('code_'+k in r and r['code_'+k].strip()
for k in CodeResult._fields):
continue
try:
diff_results.append(CodeResult(
**{k: r[k] for k in CodeResult._by
if k in r and r[k].strip()},
**{k: r['code_'+k] for k in CodeResult._fields
if 'code_'+k in r and r['code_'+k].strip()}))
except TypeError:
pass
except FileNotFoundError:
pass
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0))
for name, new in news.items():
diff[name] = (0, new, new, 1.0)
for name, old in olds.items():
_, new, _, _ = diff[name]
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
# fold
diff_results = fold(CodeResult, diff_results, by=by, defines=defines)
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
# print table
if not args.get('quiet'):
table(CodeResult, results,
diff_results if args.get('diff') else None,
by=by if by is not None else ['function'],
fields=fields,
sort=sort,
**args)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find code size at the function level.")
parser.add_argument('obj_paths', nargs='*', default=OBJ_PATHS,
help="Description of where to find *.o files. May be a directory \
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
description="Find code size at the function level.",
allow_abbrev=False)
parser.add_argument(
'obj_paths',
nargs='*',
help="Input *.o files.")
parser.add_argument(
'-v', '--verbose',
action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
parser.add_argument(
'-q', '--quiet',
action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
parser.add_argument(
'-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find code sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff code size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
parser.add_argument(
'-u', '--use',
help="Don't parse anything, use this CSV file.")
parser.add_argument(
'-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument(
'-a', '--all',
action='store_true',
help="Show all, not just the ones that changed.")
parser.add_argument(
'-p', '--percent',
action='store_true',
help="Only show percentage change, not a full diff.")
parser.add_argument(
'-b', '--by',
action='append',
choices=CodeResult._by,
help="Group by this field.")
parser.add_argument(
'-f', '--field',
dest='fields',
action='append',
choices=CodeResult._fields,
help="Show this field.")
parser.add_argument(
'-D', '--define',
dest='defines',
action='append',
type=lambda x: (lambda k,v: (k, set(v.split(','))))(*x.split('=', 1)),
help="Only include results where this field is this value.")
class AppendSort(argparse.Action):
def __call__(self, parser, namespace, value, option):
if namespace.sort is None:
namespace.sort = []
namespace.sort.append((value, True if option == '-S' else False))
parser.add_argument(
'-s', '--sort',
nargs='?',
action=AppendSort,
help="Sort by this field.")
parser.add_argument(
'-S', '--reverse-sort',
nargs='?',
action=AppendSort,
help="Sort by this field, but backwards.")
parser.add_argument(
'-Y', '--summary',
action='store_true',
help="Only show the total.")
parser.add_argument(
'-F', '--source',
dest='sources',
action='append',
help="Only consider definitions in this file. Defaults to anything "
"in the current directory.")
parser.add_argument(
'--everything',
action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level code sizes. Note this does not include padding! "
"So sizes may differ from other tools.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total code size.")
parser.add_argument('--type', default='tTrRdD',
parser.add_argument(
'--nm-types',
default=NM_TYPES,
help="Type of symbols to report, this uses the same single-character "
"type-names emitted by nm. Defaults to %(default)r.")
parser.add_argument('--nm-tool', default=['nm'], type=lambda x: x.split(),
help="Path to the nm tool to use.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))
"type-names emitted by nm. Defaults to %r." % NM_TYPES)
parser.add_argument(
'--nm-path',
type=lambda x: x.split(),
default=NM_PATH,
help="Path to the nm executable, may include flags. "
"Defaults to %r." % NM_PATH)
parser.add_argument(
'--objdump-path',
type=lambda x: x.split(),
default=OBJDUMP_PATH,
help="Path to the objdump executable, may include flags. "
"Defaults to %r." % OBJDUMP_PATH)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
Executable
+828
View File
@@ -0,0 +1,828 @@
#!/usr/bin/env python3
#
# Script to find coverage info after running tests.
#
# Example:
# ./scripts/cov.py \
# lfs2.t.a.gcda lfs2_util.t.a.gcda \
# -Flfs.c -Flfs_util.c -slines
#
# Copyright (c) 2022, The littlefs authors.
# Copyright (c) 2020, Arm Limited. All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
#
import collections as co
import csv
import itertools as it
import json
import math as m
import os
import re
import shlex
import subprocess as sp
# TODO use explode_asserts to avoid counting assert branches?
# TODO use dwarf=info to find functions for inline functions?
GCOV_PATH = ['gcov']
# integer fields
class Int(co.namedtuple('Int', 'x')):
__slots__ = ()
def __new__(cls, x=0):
if isinstance(x, Int):
return x
if isinstance(x, str):
try:
x = int(x, 0)
except ValueError:
# also accept +-∞ and +-inf
if re.match('^\s*\+?\s*(?:∞|inf)\s*$', x):
x = m.inf
elif re.match('^\s*-\s*(?:∞|inf)\s*$', x):
x = -m.inf
else:
raise
assert isinstance(x, int) or m.isinf(x), x
return super().__new__(cls, x)
def __str__(self):
if self.x == m.inf:
return ''
elif self.x == -m.inf:
return '-∞'
else:
return str(self.x)
def __int__(self):
assert not m.isinf(self.x)
return self.x
def __float__(self):
return float(self.x)
none = '%7s' % '-'
def table(self):
return '%7s' % (self,)
diff_none = '%7s' % '-'
diff_table = table
def diff_diff(self, other):
new = self.x if self else 0
old = other.x if other else 0
diff = new - old
if diff == +m.inf:
return '%7s' % '+∞'
elif diff == -m.inf:
return '%7s' % '-∞'
else:
return '%+7d' % diff
def ratio(self, other):
new = self.x if self else 0
old = other.x if other else 0
if m.isinf(new) and m.isinf(old):
return 0.0
elif m.isinf(new):
return +m.inf
elif m.isinf(old):
return -m.inf
elif not old and not new:
return 0.0
elif not old:
return 1.0
else:
return (new-old) / old
def __add__(self, other):
return self.__class__(self.x + other.x)
def __sub__(self, other):
return self.__class__(self.x - other.x)
def __mul__(self, other):
return self.__class__(self.x * other.x)
# fractional fields, a/b
class Frac(co.namedtuple('Frac', 'a,b')):
__slots__ = ()
def __new__(cls, a=0, b=None):
if isinstance(a, Frac) and b is None:
return a
if isinstance(a, str) and b is None:
a, b = a.split('/', 1)
if b is None:
b = a
return super().__new__(cls, Int(a), Int(b))
def __str__(self):
return '%s/%s' % (self.a, self.b)
def __float__(self):
return float(self.a)
none = '%11s %7s' % ('-', '-')
def table(self):
t = self.a.x/self.b.x if self.b.x else 1.0
return '%11s %7s' % (
self,
'%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%.1f%%' % (100*t))
diff_none = '%11s' % '-'
def diff_table(self):
return '%11s' % (self,)
def diff_diff(self, other):
new_a, new_b = self if self else (Int(0), Int(0))
old_a, old_b = other if other else (Int(0), Int(0))
return '%11s' % ('%s/%s' % (
new_a.diff_diff(old_a).strip(),
new_b.diff_diff(old_b).strip()))
def ratio(self, other):
new_a, new_b = self if self else (Int(0), Int(0))
old_a, old_b = other if other else (Int(0), Int(0))
new = new_a.x/new_b.x if new_b.x else 1.0
old = old_a.x/old_b.x if old_b.x else 1.0
return new - old
def __add__(self, other):
return self.__class__(self.a + other.a, self.b + other.b)
def __sub__(self, other):
return self.__class__(self.a - other.a, self.b - other.b)
def __mul__(self, other):
return self.__class__(self.a * other.a, self.b + other.b)
def __lt__(self, other):
self_t = self.a.x/self.b.x if self.b.x else 1.0
other_t = other.a.x/other.b.x if other.b.x else 1.0
return (self_t, self.a.x) < (other_t, other.a.x)
def __gt__(self, other):
return self.__class__.__lt__(other, self)
def __le__(self, other):
return not self.__gt__(other)
def __ge__(self, other):
return not self.__lt__(other)
# coverage results
class CovResult(co.namedtuple('CovResult', [
'file', 'function', 'line',
'calls', 'hits', 'funcs', 'lines', 'branches'])):
_by = ['file', 'function', 'line']
_fields = ['calls', 'hits', 'funcs', 'lines', 'branches']
_sort = ['funcs', 'lines', 'branches', 'hits', 'calls']
_types = {
'calls': Int, 'hits': Int,
'funcs': Frac, 'lines': Frac, 'branches': Frac}
__slots__ = ()
def __new__(cls, file='', function='', line=0,
calls=0, hits=0, funcs=0, lines=0, branches=0):
return super().__new__(cls, file, function, int(Int(line)),
Int(calls), Int(hits), Frac(funcs), Frac(lines), Frac(branches))
def __add__(self, other):
return CovResult(self.file, self.function, self.line,
max(self.calls, other.calls),
max(self.hits, other.hits),
self.funcs + other.funcs,
self.lines + other.lines,
self.branches + other.branches)
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def collect(gcda_paths, *,
gcov_path=GCOV_PATH,
sources=None,
everything=False,
**args):
results = []
for path in gcda_paths:
# get coverage info through gcov's json output
# note, gcov-path may contain extra args
cmd = GCOV_PATH + ['-b', '-t', '--json-format', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace',
close_fds=False)
data = json.load(proc.stdout)
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
sys.exit(-1)
# collect line/branch coverage
for file in data['files']:
# ignore filtered sources
if sources is not None:
if not any(
os.path.abspath(file['file']) == os.path.abspath(s)
for s in sources):
continue
else:
# default to only cwd
if not everything and not os.path.commonpath([
os.getcwd(),
os.path.abspath(file['file'])]) == os.getcwd():
continue
# simplify path
if os.path.commonpath([
os.getcwd(),
os.path.abspath(file['file'])]) == os.getcwd():
file_name = os.path.relpath(file['file'])
else:
file_name = os.path.abspath(file['file'])
for func in file['functions']:
func_name = func.get('name', '(inlined)')
# discard internal functions (this includes injected test cases)
if not everything:
if func_name.startswith('__'):
continue
# go ahead and add functions, later folding will merge this if
# there are other hits on this line
results.append(CovResult(
file_name, func_name, func['start_line'],
func['execution_count'], 0,
Frac(1 if func['execution_count'] > 0 else 0, 1),
0,
0))
for line in file['lines']:
func_name = line.get('function_name', '(inlined)')
# discard internal function (this includes injected test cases)
if not everything:
if func_name.startswith('__'):
continue
# go ahead and add lines, later folding will merge this if
# there are other hits on this line
results.append(CovResult(
file_name, func_name, line['line_number'],
0, line['count'],
0,
Frac(1 if line['count'] > 0 else 0, 1),
Frac(
sum(1 if branch['count'] > 0 else 0
for branch in line['branches']),
len(line['branches']))))
return results
def fold(Result, results, *,
by=None,
defines=None,
**_):
if by is None:
by = Result._by
for k in it.chain(by or [], (k for k, _ in defines or [])):
if k not in Result._by and k not in Result._fields:
print("error: could not find field %r?" % k)
sys.exit(-1)
# filter by matching defines
if defines is not None:
results_ = []
for r in results:
if all(getattr(r, k) in vs for k, vs in defines):
results_.append(r)
results = results_
# organize results into conflicts
folding = co.OrderedDict()
for r in results:
name = tuple(getattr(r, k) for k in by)
if name not in folding:
folding[name] = []
folding[name].append(r)
# merge conflicts
folded = []
for name, rs in folding.items():
folded.append(sum(rs[1:], start=rs[0]))
return folded
def table(Result, results, diff_results=None, *,
by=None,
fields=None,
sort=None,
summary=False,
all=False,
percent=False,
**_):
all_, all = all, __builtins__.all
if by is None:
by = Result._by
if fields is None:
fields = Result._fields
types = Result._types
# fold again
results = fold(Result, results, by=by)
if diff_results is not None:
diff_results = fold(Result, diff_results, by=by)
# organize by name
table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in results}
diff_table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in diff_results or []}
names = list(table.keys() | diff_table.keys())
# sort again, now with diff info, note that python's sort is stable
names.sort()
if diff_results is not None:
names.sort(key=lambda n: tuple(
types[k].ratio(
getattr(table.get(n), k, None),
getattr(diff_table.get(n), k, None))
for k in fields),
reverse=True)
if sort:
for k, reverse in reversed(sort):
names.sort(
key=lambda n: tuple(
(getattr(table[n], k),)
if getattr(table.get(n), k, None) is not None else ()
for k in ([k] if k else [
k for k in Result._sort if k in fields])),
reverse=reverse ^ (not k or k in Result._fields))
# build up our lines
lines = []
# header
header = []
header.append('%s%s' % (
','.join(by),
' (%d added, %d removed)' % (
sum(1 for n in table if n not in diff_table),
sum(1 for n in diff_table if n not in table))
if diff_results is not None and not percent else '')
if not summary else '')
if diff_results is None:
for k in fields:
header.append(k)
elif percent:
for k in fields:
header.append(k)
else:
for k in fields:
header.append('o'+k)
for k in fields:
header.append('n'+k)
for k in fields:
header.append('d'+k)
header.append('')
lines.append(header)
def table_entry(name, r, diff_r=None, ratios=[]):
entry = []
entry.append(name)
if diff_results is None:
for k in fields:
entry.append(getattr(r, k).table()
if getattr(r, k, None) is not None
else types[k].none)
elif percent:
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
else:
for k in fields:
entry.append(getattr(diff_r, k).diff_table()
if getattr(diff_r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(types[k].diff_diff(
getattr(r, k, None),
getattr(diff_r, k, None)))
if diff_results is None:
entry.append('')
elif percent:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios))
else:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios
if t)
if any(ratios) else '')
return entry
# entries
if not summary:
for name in names:
r = table.get(name)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = diff_table.get(name)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
if not all_ and not any(ratios):
continue
lines.append(table_entry(name, r, diff_r, ratios))
# total
r = next(iter(fold(Result, results, by=[])), None)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = next(iter(fold(Result, diff_results, by=[])), None)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
lines.append(table_entry('TOTAL', r, diff_r, ratios))
# find the best widths, note that column 0 contains the names and column -1
# the ratios, so those are handled a bit differently
widths = [
((max(it.chain([w], (len(l[i]) for l in lines)))+1+4-1)//4)*4-1
for w, i in zip(
it.chain([23], it.repeat(7)),
range(len(lines[0])-1))]
# print our table
for line in lines:
print('%-*s %s%s' % (
widths[0], line[0],
' '.join('%*s' % (w, x)
for w, x in zip(widths[1:], line[1:-1])),
line[-1]))
def annotate(Result, results, *,
annotate=False,
lines=False,
branches=False,
**args):
# if neither branches/lines specified, color both
if annotate and not lines and not branches:
lines, branches = True, True
for path in co.OrderedDict.fromkeys(r.file for r in results).keys():
# flatten to line info
results = fold(Result, results, by=['file', 'line'])
table = {r.line: r for r in results if r.file == path}
# calculate spans to show
if not annotate:
spans = []
last = None
func = None
for line, r in sorted(table.items()):
if ((lines and int(r.hits) == 0)
or (branches and r.branches.a < r.branches.b)):
if last is not None and line - last.stop <= args['context']:
last = range(
last.start,
line+1+args['context'])
else:
if last is not None:
spans.append((last, func))
last = range(
line-args['context'],
line+1+args['context'])
func = r.function
if last is not None:
spans.append((last, func))
with open(path) as f:
skipped = False
for i, line in enumerate(f):
# skip lines not in spans?
if not annotate and not any(i+1 in s for s, _ in spans):
skipped = True
continue
if skipped:
skipped = False
print('%s@@ %s:%d: %s @@%s' % (
'\x1b[36m' if args['color'] else '',
path,
i+1,
next(iter(f for _, f in spans)),
'\x1b[m' if args['color'] else ''))
# build line
if line.endswith('\n'):
line = line[:-1]
if i+1 in table:
r = table[i+1]
line = '%-*s // %s hits%s' % (
args['width'],
line,
r.hits,
', %s branches' % (r.branches,)
if int(r.branches.b) else '')
if args['color']:
if lines and int(r.hits) == 0:
line = '\x1b[1;31m%s\x1b[m' % line
elif branches and r.branches.a < r.branches.b:
line = '\x1b[35m%s\x1b[m' % line
print(line)
def main(gcda_paths, *,
by=None,
fields=None,
defines=None,
sort=None,
hits=False,
**args):
# figure out what color should be
if args.get('color') == 'auto':
args['color'] = sys.stdout.isatty()
elif args.get('color') == 'always':
args['color'] = True
else:
args['color'] = False
# find sizes
if not args.get('use', None):
results = collect(gcda_paths, **args)
else:
results = []
with openio(args['use']) as f:
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('cov_'+k in r and r['cov_'+k].strip()
for k in CovResult._fields):
continue
try:
results.append(CovResult(
**{k: r[k] for k in CovResult._by
if k in r and r[k].strip()},
**{k: r['cov_'+k]
for k in CovResult._fields
if 'cov_'+k in r
and r['cov_'+k].strip()}))
except TypeError:
pass
# fold
results = fold(CovResult, results, by=by, defines=defines)
# sort, note that python's sort is stable
results.sort()
if sort:
for k, reverse in reversed(sort):
results.sort(
key=lambda r: tuple(
(getattr(r, k),) if getattr(r, k) is not None else ()
for k in ([k] if k else CovResult._sort)),
reverse=reverse ^ (not k or k in CovResult._fields))
# write results to CSV
if args.get('output'):
with openio(args['output'], 'w') as f:
writer = csv.DictWriter(f,
(by if by is not None else CovResult._by)
+ ['cov_'+k for k in (
fields if fields is not None else CovResult._fields)])
writer.writeheader()
for r in results:
writer.writerow(
{k: getattr(r, k) for k in (
by if by is not None else CovResult._by)}
| {'cov_'+k: getattr(r, k) for k in (
fields if fields is not None else CovResult._fields)})
# find previous results?
if args.get('diff'):
diff_results = []
try:
with openio(args['diff']) as f:
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('cov_'+k in r and r['cov_'+k].strip()
for k in CovResult._fields):
continue
try:
diff_results.append(CovResult(
**{k: r[k] for k in CovResult._by
if k in r and r[k].strip()},
**{k: r['cov_'+k]
for k in CovResult._fields
if 'cov_'+k in r
and r['cov_'+k].strip()}))
except TypeError:
pass
except FileNotFoundError:
pass
# fold
diff_results = fold(CovResult, diff_results,
by=by, defines=defines)
# print table
if not args.get('quiet'):
if (args.get('annotate')
or args.get('lines')
or args.get('branches')):
# annotate sources
annotate(CovResult, results, **args)
else:
# print table
table(CovResult, results,
diff_results if args.get('diff') else None,
by=by if by is not None else ['function'],
fields=fields if fields is not None
else ['lines', 'branches'] if not hits
else ['calls', 'hits'],
sort=sort,
**args)
# catch lack of coverage
if args.get('error_on_lines') and any(
r.lines.a < r.lines.b for r in results):
sys.exit(2)
elif args.get('error_on_branches') and any(
r.branches.a < r.branches.b for r in results):
sys.exit(3)
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find coverage info after running tests.",
allow_abbrev=False)
parser.add_argument(
'gcda_paths',
nargs='*',
help="Input *.gcda files.")
parser.add_argument(
'-v', '--verbose',
action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument(
'-q', '--quiet',
action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument(
'-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument(
'-u', '--use',
help="Don't parse anything, use this CSV file.")
parser.add_argument(
'-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument(
'-a', '--all',
action='store_true',
help="Show all, not just the ones that changed.")
parser.add_argument(
'-p', '--percent',
action='store_true',
help="Only show percentage change, not a full diff.")
parser.add_argument(
'-b', '--by',
action='append',
choices=CovResult._by,
help="Group by this field.")
parser.add_argument(
'-f', '--field',
dest='fields',
action='append',
choices=CovResult._fields,
help="Show this field.")
parser.add_argument(
'-D', '--define',
dest='defines',
action='append',
type=lambda x: (lambda k,v: (k, set(v.split(','))))(*x.split('=', 1)),
help="Only include results where this field is this value.")
class AppendSort(argparse.Action):
def __call__(self, parser, namespace, value, option):
if namespace.sort is None:
namespace.sort = []
namespace.sort.append((value, True if option == '-S' else False))
parser.add_argument(
'-s', '--sort',
nargs='?',
action=AppendSort,
help="Sort by this field.")
parser.add_argument(
'-S', '--reverse-sort',
nargs='?',
action=AppendSort,
help="Sort by this field, but backwards.")
parser.add_argument(
'-Y', '--summary',
action='store_true',
help="Only show the total.")
parser.add_argument(
'-F', '--source',
dest='sources',
action='append',
help="Only consider definitions in this file. Defaults to anything "
"in the current directory.")
parser.add_argument(
'--everything',
action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument(
'--hits',
action='store_true',
help="Show total hits instead of coverage.")
parser.add_argument(
'-A', '--annotate',
action='store_true',
help="Show source files annotated with coverage info.")
parser.add_argument(
'-L', '--lines',
action='store_true',
help="Show uncovered lines.")
parser.add_argument(
'-B', '--branches',
action='store_true',
help="Show uncovered branches.")
parser.add_argument(
'-c', '--context',
type=lambda x: int(x, 0),
default=3,
help="Show n additional lines of context. Defaults to 3.")
parser.add_argument(
'-W', '--width',
type=lambda x: int(x, 0),
default=80,
help="Assume source is styled with this many columns. Defaults to 80.")
parser.add_argument(
'--color',
choices=['never', 'always', 'auto'],
default='auto',
help="When to use terminal colors. Defaults to 'auto'.")
parser.add_argument(
'-e', '--error-on-lines',
action='store_true',
help="Error if any lines are not covered.")
parser.add_argument(
'-E', '--error-on-branches',
action='store_true',
help="Error if any branches are not covered.")
parser.add_argument(
'--gcov-path',
default=GCOV_PATH,
type=lambda x: x.split(),
help="Path to the gcov executable, may include paths. "
"Defaults to %r." % GCOV_PATH)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
-323
View File
@@ -1,323 +0,0 @@
#!/usr/bin/env python3
#
# Parse and report coverage info from .info files generated by lcov
#
import os
import glob
import csv
import re
import collections as co
import bisect as b
INFO_PATHS = ['tests/*.toml.info']
def collect(paths, **args):
file = None
funcs = []
lines = co.defaultdict(lambda: 0)
pattern = re.compile(
'^(?P<file>SF:/?(?P<file_name>.*))$'
'|^(?P<func>FN:(?P<func_lineno>[0-9]*),(?P<func_name>.*))$'
'|^(?P<line>DA:(?P<line_lineno>[0-9]*),(?P<line_hits>[0-9]*))$')
for path in paths:
with open(path) as f:
for line in f:
m = pattern.match(line)
if m and m.group('file'):
file = m.group('file_name')
elif m and file and m.group('func'):
funcs.append((file, int(m.group('func_lineno')),
m.group('func_name')))
elif m and file and m.group('line'):
lines[(file, int(m.group('line_lineno')))] += (
int(m.group('line_hits')))
# map line numbers to functions
funcs.sort()
def func_from_lineno(file, lineno):
i = b.bisect(funcs, (file, lineno))
if i and funcs[i-1][0] == file:
return funcs[i-1][2]
else:
return None
# reduce to function info
reduced_funcs = co.defaultdict(lambda: (0, 0))
for (file, line_lineno), line_hits in lines.items():
func = func_from_lineno(file, line_lineno)
if not func:
continue
hits, count = reduced_funcs[(file, func)]
reduced_funcs[(file, func)] = (hits + (line_hits > 0), count + 1)
results = []
for (file, func), (hits, count) in reduced_funcs.items():
# discard internal/testing functions (test_* injected with
# internal testing)
if not args.get('everything'):
if func.startswith('__') or func.startswith('test_'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
results.append((file, func, hits, count))
return results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# find coverage
if not args.get('use'):
# find *.info files
paths = []
for path in args['info_paths']:
if os.path.isdir(path):
path = path + '/*.gcov'
for path in glob.glob(path):
paths.append(path)
if not paths:
print('no .info files found in %r?' % args['info_paths'])
sys.exit(-1)
results = collect(paths, **args)
else:
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['coverage_hits']),
int(result['coverage_count']))
for result in r
if result.get('coverage_hits') not in {None, ''}
if result.get('coverage_count') not in {None, ''}]
total_hits, total_count = 0, 0
for _, _, hits, count in results:
total_hits += hits
total_count += count
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['coverage_hits']),
int(result['coverage_count']))
for result in r
if result.get('coverage_hits') not in {None, ''}
if result.get('coverage_count') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total_hits, prev_total_count = 0, 0
for _, _, hits, count in prev_results:
prev_total_hits += hits
prev_total_count += count
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('coverage_hits', None)
result.pop('coverage_count', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, hits, count in results:
merged_results[(file, func)]['coverage_hits'] = hits
merged_results[(file, func)]['coverage_count'] = count
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'coverage_hits', 'coverage_count'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: (0, 0))
for file, func, hits, count in results:
entry = (file if by == 'file' else func)
entry_hits, entry_count = entries[entry]
entries[entry] = (entry_hits + hits, entry_count + count)
return entries
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0, 0, 0, 0))
for name, (new_hits, new_count) in news.items():
diff[name] = (
0, 0,
new_hits, new_count,
new_hits, new_count,
(new_hits/new_count if new_count else 1.0) - 1.0)
for name, (old_hits, old_count) in olds.items():
_, _, new_hits, new_count, _, _, _ = diff[name]
diff[name] = (
old_hits, old_count,
new_hits, new_count,
new_hits-old_hits, new_count-old_count,
((new_hits/new_count if new_count else 1.0)
- (old_hits/old_count if old_count else 1.0)))
return diff
def sorted_entries(entries):
if args.get('coverage_sort'):
return sorted(entries, key=lambda x: (-(x[1][0]/x[1][1] if x[1][1] else -1), x))
elif args.get('reverse_coverage_sort'):
return sorted(entries, key=lambda x: (+(x[1][0]/x[1][1] if x[1][1] else -1), x))
else:
return sorted(entries)
def sorted_diff_entries(entries):
if args.get('coverage_sort'):
return sorted(entries, key=lambda x: (-(x[1][2]/x[1][3] if x[1][3] else -1), x))
elif args.get('reverse_coverage_sort'):
return sorted(entries, key=lambda x: (+(x[1][2]/x[1][3] if x[1][3] else -1), x))
else:
return sorted(entries, key=lambda x: (-x[1][6], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %19s' % (by, 'hits/line'))
else:
print('%-36s %19s %19s %11s' % (by, 'old', 'new', 'diff'))
def print_entry(name, hits, count):
print("%-36s %11s %7s" % (name,
'%d/%d' % (hits, count)
if count else '-',
'%.1f%%' % (100*hits/count)
if count else '-'))
def print_diff_entry(name,
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count,
ratio):
print("%-36s %11s %7s %11s %7s %11s%s" % (name,
'%d/%d' % (old_hits, old_count)
if old_count else '-',
'%.1f%%' % (100*old_hits/old_count)
if old_count else '-',
'%d/%d' % (new_hits, new_count)
if new_count else '-',
'%.1f%%' % (100*new_hits/new_count)
if new_count else '-',
'%+d/%+d' % (diff_hits, diff_count),
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, (hits, count) in sorted_entries(entries.items()):
print_entry(name, hits, count)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for _, old, _, _, _, _, _ in diff.values() if not old),
sum(1 for _, _, _, new, _, _, _ in diff.values() if not new)))
for name, (
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name,
old_hits, old_count,
new_hits, new_count,
diff_hits, diff_count,
ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total_hits, total_count)
else:
ratio = ((total_hits/total_count
if total_count else 1.0)
- (prev_total_hits/prev_total_count
if prev_total_count else 1.0))
print_diff_entry('TOTAL',
prev_total_hits, prev_total_count,
total_hits, total_count,
total_hits-prev_total_hits, total_count-prev_total_count,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Parse and report coverage info from .info files \
generated by lcov")
parser.add_argument('info_paths', nargs='*', default=INFO_PATHS,
help="Description of where to find *.info files. May be a directory \
or list of paths. *.info files will be merged to show the total \
coverage. Defaults to %r." % INFO_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't do any work, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff code size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--coverage-sort', action='store_true',
help="Sort by coverage.")
parser.add_argument('-S', '--reverse-coverage-sort', action='store_true',
help="Sort by coverage, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level coverage.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total coverage.")
parser.add_argument('-q', '--quiet', action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))
+647 -226
View File
@@ -1,42 +1,188 @@
#!/usr/bin/env python3
#
# Script to find data size at the function level. Basically just a bit wrapper
# Script to find data size at the function level. Basically just a big wrapper
# around nm with some extra conveniences for comparing builds. Heavily inspired
# by Linux's Bloat-O-Meter.
#
# Example:
# ./scripts/data.py lfs2.o lfs2_util.o -Ssize
#
# Copyright (c) 2022, The littlefs authors.
# Copyright (c) 2020, Arm Limited. All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
#
import os
import glob
import itertools as it
import subprocess as sp
import shlex
import re
import csv
import collections as co
import csv
import difflib
import itertools as it
import math as m
import os
import re
import shlex
import subprocess as sp
OBJ_PATHS = ['*.o']
NM_PATH = ['nm']
NM_TYPES = 'dDbB'
OBJDUMP_PATH = ['objdump']
def collect(paths, **args):
results = co.defaultdict(lambda: 0)
pattern = re.compile(
# integer fields
class Int(co.namedtuple('Int', 'x')):
__slots__ = ()
def __new__(cls, x=0):
if isinstance(x, Int):
return x
if isinstance(x, str):
try:
x = int(x, 0)
except ValueError:
# also accept +-∞ and +-inf
if re.match('^\s*\+?\s*(?:∞|inf)\s*$', x):
x = m.inf
elif re.match('^\s*-\s*(?:∞|inf)\s*$', x):
x = -m.inf
else:
raise
assert isinstance(x, int) or m.isinf(x), x
return super().__new__(cls, x)
def __str__(self):
if self.x == m.inf:
return ''
elif self.x == -m.inf:
return '-∞'
else:
return str(self.x)
def __int__(self):
assert not m.isinf(self.x)
return self.x
def __float__(self):
return float(self.x)
none = '%7s' % '-'
def table(self):
return '%7s' % (self,)
diff_none = '%7s' % '-'
diff_table = table
def diff_diff(self, other):
new = self.x if self else 0
old = other.x if other else 0
diff = new - old
if diff == +m.inf:
return '%7s' % '+∞'
elif diff == -m.inf:
return '%7s' % '-∞'
else:
return '%+7d' % diff
def ratio(self, other):
new = self.x if self else 0
old = other.x if other else 0
if m.isinf(new) and m.isinf(old):
return 0.0
elif m.isinf(new):
return +m.inf
elif m.isinf(old):
return -m.inf
elif not old and not new:
return 0.0
elif not old:
return 1.0
else:
return (new-old) / old
def __add__(self, other):
return self.__class__(self.x + other.x)
def __sub__(self, other):
return self.__class__(self.x - other.x)
def __mul__(self, other):
return self.__class__(self.x * other.x)
# data size results
class DataResult(co.namedtuple('DataResult', [
'file', 'function',
'size'])):
_by = ['file', 'function']
_fields = ['size']
_sort = ['size']
_types = {'size': Int}
__slots__ = ()
def __new__(cls, file='', function='', size=0):
return super().__new__(cls, file, function,
Int(size))
def __add__(self, other):
return DataResult(self.file, self.function,
self.size + other.size)
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def collect(obj_paths, *,
nm_path=NM_PATH,
nm_types=NM_TYPES,
objdump_path=OBJDUMP_PATH,
sources=None,
everything=False,
**args):
size_pattern = re.compile(
'^(?P<size>[0-9a-fA-F]+)' +
' (?P<type>[%s])' % re.escape(args['type']) +
' (?P<type>[%s])' % re.escape(nm_types) +
' (?P<func>.+?)$')
for path in paths:
# note nm-tool may contain extra args
cmd = args['nm_tool'] + ['--size-sort', path]
line_pattern = re.compile(
'^\s+(?P<no>[0-9]+)'
'(?:\s+(?P<dir>[0-9]+))?'
'\s+.*'
'\s+(?P<path>[^\s]+)$')
info_pattern = re.compile(
'^(?:.*(?P<tag>DW_TAG_[a-z_]+).*'
'|.*DW_AT_name.*:\s*(?P<name>[^:\s]+)\s*'
'|.*DW_AT_decl_file.*:\s*(?P<file>[0-9]+)\s*)$')
results = []
for path in obj_paths:
# guess the source, if we have debug-info we'll replace this later
file = re.sub('(\.o)?$', '.c', path, 1)
# find symbol sizes
results_ = []
# note nm-path may contain extra args
cmd = nm_path + ['--size-sort', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
errors='replace',
close_fds=False)
for line in proc.stdout:
m = pattern.match(line)
m = size_pattern.match(line)
if m:
results[(path, m.group('func'))] += int(m.group('size'), 16)
func = m.group('func')
# discard internal functions
if not everything and func.startswith('__'):
continue
results_.append(DataResult(
file, func,
int(m.group('size'), 16)))
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
@@ -44,240 +190,515 @@ def collect(paths, **args):
sys.stdout.write(line)
sys.exit(-1)
flat_results = []
for (file, func), size in results.items():
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
# discard internal functions
if not args.get('everything'):
if func.startswith('__'):
continue
# discard .8449 suffixes created by optimizer
func = re.sub('\.[0-9]+', '', func)
flat_results.append((file, func, size))
return flat_results
# try to figure out the source file if we have debug-info
dirs = {}
files = {}
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=rawline', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace',
close_fds=False)
for line in proc.stdout:
# note that files contain references to dirs, which we
# dereference as soon as we see them as each file table follows a
# dir table
m = line_pattern.match(line)
if m:
if not m.group('dir'):
# found a directory entry
dirs[int(m.group('no'))] = m.group('path')
else:
# found a file entry
dir = int(m.group('dir'))
if dir in dirs:
files[int(m.group('no'))] = os.path.join(
dirs[dir],
m.group('path'))
else:
files[int(m.group('no'))] = m.group('path')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
# do nothing on error, we don't need objdump to work, source files
# may just be inaccurate
pass
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
defs = {}
is_func = False
f_name = None
f_file = None
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=info', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace',
close_fds=False)
for line in proc.stdout:
# state machine here to find definitions
m = info_pattern.match(line)
if m:
if m.group('tag'):
if is_func:
defs[f_name] = files.get(f_file, '?')
is_func = (m.group('tag') == 'DW_TAG_subprogram')
elif m.group('name'):
f_name = m.group('name')
elif m.group('file'):
f_file = int(m.group('file'))
if is_func:
defs[f_name] = files.get(f_file, '?')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
for line in proc.stderr:
sys.stdout.write(line)
# do nothing on error, we don't need objdump to work, source files
# may just be inaccurate
pass
for r in results_:
# find best matching debug symbol, this may be slightly different
# due to optimizations
if defs:
# exact match? avoid difflib if we can for speed
if r.function in defs:
file = defs[r.function]
else:
_, file = max(
defs.items(),
key=lambda d: difflib.SequenceMatcher(None,
d[0],
r.function, False).ratio())
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
file = r.file
# find sizes
if not args.get('use', None):
# find .o files
paths = []
for path in args['obj_paths']:
if os.path.isdir(path):
path = path + '/*.o'
# ignore filtered sources
if sources is not None:
if not any(
os.path.abspath(file) == os.path.abspath(s)
for s in sources):
continue
else:
# default to only cwd
if not everything and not os.path.commonpath([
os.getcwd(),
os.path.abspath(file)]) == os.getcwd():
continue
for path in glob.glob(path):
paths.append(path)
# simplify path
if os.path.commonpath([
os.getcwd(),
os.path.abspath(file)]) == os.getcwd():
file = os.path.relpath(file)
else:
file = os.path.abspath(file)
if not paths:
print('no .obj files found in %r?' % args['obj_paths'])
results.append(r._replace(file=file))
return results
def fold(Result, results, *,
by=None,
defines=None,
**_):
if by is None:
by = Result._by
for k in it.chain(by or [], (k for k, _ in defines or [])):
if k not in Result._by and k not in Result._fields:
print("error: could not find field %r?" % k)
sys.exit(-1)
results = collect(paths, **args)
# filter by matching defines
if defines is not None:
results_ = []
for r in results:
if all(getattr(r, k) in vs for k, vs in defines):
results_.append(r)
results = results_
# organize results into conflicts
folding = co.OrderedDict()
for r in results:
name = tuple(getattr(r, k) for k in by)
if name not in folding:
folding[name] = []
folding[name].append(r)
# merge conflicts
folded = []
for name, rs in folding.items():
folded.append(sum(rs[1:], start=rs[0]))
return folded
def table(Result, results, diff_results=None, *,
by=None,
fields=None,
sort=None,
summary=False,
all=False,
percent=False,
**_):
all_, all = all, __builtins__.all
if by is None:
by = Result._by
if fields is None:
fields = Result._fields
types = Result._types
# fold again
results = fold(Result, results, by=by)
if diff_results is not None:
diff_results = fold(Result, diff_results, by=by)
# organize by name
table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in results}
diff_table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in diff_results or []}
names = list(table.keys() | diff_table.keys())
# sort again, now with diff info, note that python's sort is stable
names.sort()
if diff_results is not None:
names.sort(key=lambda n: tuple(
types[k].ratio(
getattr(table.get(n), k, None),
getattr(diff_table.get(n), k, None))
for k in fields),
reverse=True)
if sort:
for k, reverse in reversed(sort):
names.sort(
key=lambda n: tuple(
(getattr(table[n], k),)
if getattr(table.get(n), k, None) is not None else ()
for k in ([k] if k else [
k for k in Result._sort if k in fields])),
reverse=reverse ^ (not k or k in Result._fields))
# build up our lines
lines = []
# header
header = []
header.append('%s%s' % (
','.join(by),
' (%d added, %d removed)' % (
sum(1 for n in table if n not in diff_table),
sum(1 for n in diff_table if n not in table))
if diff_results is not None and not percent else '')
if not summary else '')
if diff_results is None:
for k in fields:
header.append(k)
elif percent:
for k in fields:
header.append(k)
else:
for k in fields:
header.append('o'+k)
for k in fields:
header.append('n'+k)
for k in fields:
header.append('d'+k)
header.append('')
lines.append(header)
def table_entry(name, r, diff_r=None, ratios=[]):
entry = []
entry.append(name)
if diff_results is None:
for k in fields:
entry.append(getattr(r, k).table()
if getattr(r, k, None) is not None
else types[k].none)
elif percent:
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
else:
for k in fields:
entry.append(getattr(diff_r, k).diff_table()
if getattr(diff_r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(types[k].diff_diff(
getattr(r, k, None),
getattr(diff_r, k, None)))
if diff_results is None:
entry.append('')
elif percent:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios))
else:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios
if t)
if any(ratios) else '')
return entry
# entries
if not summary:
for name in names:
r = table.get(name)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = diff_table.get(name)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
if not all_ and not any(ratios):
continue
lines.append(table_entry(name, r, diff_r, ratios))
# total
r = next(iter(fold(Result, results, by=[])), None)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = next(iter(fold(Result, diff_results, by=[])), None)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
lines.append(table_entry('TOTAL', r, diff_r, ratios))
# find the best widths, note that column 0 contains the names and column -1
# the ratios, so those are handled a bit differently
widths = [
((max(it.chain([w], (len(l[i]) for l in lines)))+1+4-1)//4)*4-1
for w, i in zip(
it.chain([23], it.repeat(7)),
range(len(lines[0])-1))]
# print our table
for line in lines:
print('%-*s %s%s' % (
widths[0], line[0],
' '.join('%*s' % (w, x)
for w, x in zip(widths[1:], line[1:-1])),
line[-1]))
def main(obj_paths, *,
by=None,
fields=None,
defines=None,
sort=None,
**args):
# find sizes
if not args.get('use', None):
results = collect(obj_paths, **args)
else:
results = []
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['data_size']))
for result in r
if result.get('data_size') not in {None, ''}]
reader = csv.DictReader(f, restval='')
for r in reader:
try:
results.append(DataResult(
**{k: r[k] for k in DataResult._by
if k in r and r[k].strip()},
**{k: r['data_'+k] for k in DataResult._fields
if 'data_'+k in r and r['data_'+k].strip()}))
except TypeError:
pass
total = 0
for _, _, size in results:
total += size
# fold
results = fold(DataResult, results, by=by, defines=defines)
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['data_size']))
for result in r
if result.get('data_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
prev_total += size
# sort, note that python's sort is stable
results.sort()
if sort:
for k, reverse in reversed(sort):
results.sort(
key=lambda r: tuple(
(getattr(r, k),) if getattr(r, k) is not None else ()
for k in ([k] if k else DataResult._sort)),
reverse=reverse ^ (not k or k in DataResult._fields))
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
func = result.pop('name', '')
result.pop('data_size', None)
merged_results[(file, func)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, func, size in results:
merged_results[(file, func)]['data_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'data_size'])
w.writeheader()
for (file, func), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': func, **result})
writer = csv.DictWriter(f,
(by if by is not None else DataResult._by)
+ ['data_'+k for k in (
fields if fields is not None else DataResult._fields)])
writer.writeheader()
for r in results:
writer.writerow(
{k: getattr(r, k) for k in (
by if by is not None else DataResult._by)}
| {'data_'+k: getattr(r, k) for k in (
fields if fields is not None else DataResult._fields)})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, func, size in results:
entry = (file if by == 'file' else func)
entries[entry] += size
return entries
# find previous results?
if args.get('diff'):
diff_results = []
try:
with openio(args['diff']) as f:
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('data_'+k in r and r['data_'+k].strip()
for k in DataResult._fields):
continue
try:
diff_results.append(DataResult(
**{k: r[k] for k in DataResult._by
if k in r and r[k].strip()},
**{k: r['data_'+k] for k in DataResult._fields
if 'data_'+k in r and r['data_'+k].strip()}))
except TypeError:
pass
except FileNotFoundError:
pass
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0))
for name, new in news.items():
diff[name] = (0, new, new, 1.0)
for name, old in olds.items():
_, new, _, _ = diff[name]
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
# fold
diff_results = fold(DataResult, diff_results, by=by, defines=defines)
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
# print table
if not args.get('quiet'):
table(DataResult, results,
diff_results if args.get('diff') else None,
by=by if by is not None else ['function'],
fields=fields,
sort=sort,
**args)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find data size at the function level.")
parser.add_argument('obj_paths', nargs='*', default=OBJ_PATHS,
help="Description of where to find *.o files. May be a directory \
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
description="Find data size at the function level.",
allow_abbrev=False)
parser.add_argument(
'obj_paths',
nargs='*',
help="Input *.o files.")
parser.add_argument(
'-v', '--verbose',
action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
parser.add_argument(
'-q', '--quiet',
action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
parser.add_argument(
'-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find data sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff data size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
parser.add_argument(
'-u', '--use',
help="Don't parse anything, use this CSV file.")
parser.add_argument(
'-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument(
'-a', '--all',
action='store_true',
help="Show all, not just the ones that changed.")
parser.add_argument(
'-p', '--percent',
action='store_true',
help="Only show percentage change, not a full diff.")
parser.add_argument(
'-b', '--by',
action='append',
choices=DataResult._by,
help="Group by this field.")
parser.add_argument(
'-f', '--field',
dest='fields',
action='append',
choices=DataResult._fields,
help="Show this field.")
parser.add_argument(
'-D', '--define',
dest='defines',
action='append',
type=lambda x: (lambda k,v: (k, set(v.split(','))))(*x.split('=', 1)),
help="Only include results where this field is this value.")
class AppendSort(argparse.Action):
def __call__(self, parser, namespace, value, option):
if namespace.sort is None:
namespace.sort = []
namespace.sort.append((value, True if option == '-S' else False))
parser.add_argument(
'-s', '--sort',
nargs='?',
action=AppendSort,
help="Sort by this field.")
parser.add_argument(
'-S', '--reverse-sort',
nargs='?',
action=AppendSort,
help="Sort by this field, but backwards.")
parser.add_argument(
'-Y', '--summary',
action='store_true',
help="Only show the total.")
parser.add_argument(
'-F', '--source',
dest='sources',
action='append',
help="Only consider definitions in this file. Defaults to anything "
"in the current directory.")
parser.add_argument(
'--everything',
action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level data sizes. Note this does not include padding! "
"So sizes may differ from other tools.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total data size.")
parser.add_argument('--type', default='dDbB',
parser.add_argument(
'--nm-types',
default=NM_TYPES,
help="Type of symbols to report, this uses the same single-character "
"type-names emitted by nm. Defaults to %(default)r.")
parser.add_argument('--nm-tool', default=['nm'], type=lambda x: x.split(),
help="Path to the nm tool to use.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))
"type-names emitted by nm. Defaults to %r." % NM_TYPES)
parser.add_argument(
'--nm-path',
type=lambda x: x.split(),
default=NM_PATH,
help="Path to the nm executable, may include flags. "
"Defaults to %r." % NM_PATH)
parser.add_argument(
'--objdump-path',
type=lambda x: x.split(),
default=OBJDUMP_PATH,
help="Path to the objdump executable, may include flags. "
"Defaults to %r." % OBJDUMP_PATH)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
-383
View File
@@ -1,383 +0,0 @@
#!/usr/bin/env python3
import re
import sys
PATTERN = ['LFS2_ASSERT', 'assert']
PREFIX = 'LFS2'
MAXWIDTH = 16
ASSERT = "__{PREFIX}_ASSERT_{TYPE}_{COMP}"
FAIL = """
__attribute__((unused))
static void __{prefix}_assert_fail_{type}(
const char *file, int line, const char *comp,
{ctype} lh, size_t lsize,
{ctype} rh, size_t rsize) {{
printf("%s:%d:assert: assert failed with ", file, line);
__{prefix}_assert_print_{type}(lh, lsize);
printf(", expected %s ", comp);
__{prefix}_assert_print_{type}(rh, rsize);
printf("\\n");
fflush(NULL);
raise(SIGABRT);
}}
"""
COMP = {
'==': 'eq',
'!=': 'ne',
'<=': 'le',
'>=': 'ge',
'<': 'lt',
'>': 'gt',
}
TYPE = {
'int': {
'ctype': 'intmax_t',
'fail': FAIL,
'print': """
__attribute__((unused))
static void __{prefix}_assert_print_{type}({ctype} v, size_t size) {{
(void)size;
printf("%"PRIiMAX, v);
}}
""",
'assert': """
#define __{PREFIX}_ASSERT_{TYPE}_{COMP}(file, line, lh, rh)
do {{
__typeof__(lh) _lh = lh;
__typeof__(lh) _rh = (__typeof__(lh))rh;
if (!(_lh {op} _rh)) {{
__{prefix}_assert_fail_{type}(file, line, "{comp}",
(intmax_t)_lh, 0, (intmax_t)_rh, 0);
}}
}} while (0)
"""
},
'bool': {
'ctype': 'bool',
'fail': FAIL,
'print': """
__attribute__((unused))
static void __{prefix}_assert_print_{type}({ctype} v, size_t size) {{
(void)size;
printf("%s", v ? "true" : "false");
}}
""",
'assert': """
#define __{PREFIX}_ASSERT_{TYPE}_{COMP}(file, line, lh, rh)
do {{
bool _lh = !!(lh);
bool _rh = !!(rh);
if (!(_lh {op} _rh)) {{
__{prefix}_assert_fail_{type}(file, line, "{comp}",
_lh, 0, _rh, 0);
}}
}} while (0)
"""
},
'mem': {
'ctype': 'const void *',
'fail': FAIL,
'print': """
__attribute__((unused))
static void __{prefix}_assert_print_{type}({ctype} v, size_t size) {{
const uint8_t *s = v;
printf("\\\"");
for (size_t i = 0; i < size && i < {maxwidth}; i++) {{
if (s[i] >= ' ' && s[i] <= '~') {{
printf("%c", s[i]);
}} else {{
printf("\\\\x%02x", s[i]);
}}
}}
if (size > {maxwidth}) {{
printf("...");
}}
printf("\\\"");
}}
""",
'assert': """
#define __{PREFIX}_ASSERT_{TYPE}_{COMP}(file, line, lh, rh, size)
do {{
const void *_lh = lh;
const void *_rh = rh;
if (!(memcmp(_lh, _rh, size) {op} 0)) {{
__{prefix}_assert_fail_{type}(file, line, "{comp}",
_lh, size, _rh, size);
}}
}} while (0)
"""
},
'str': {
'ctype': 'const char *',
'fail': FAIL,
'print': """
__attribute__((unused))
static void __{prefix}_assert_print_{type}({ctype} v, size_t size) {{
__{prefix}_assert_print_mem(v, size);
}}
""",
'assert': """
#define __{PREFIX}_ASSERT_{TYPE}_{COMP}(file, line, lh, rh)
do {{
const char *_lh = lh;
const char *_rh = rh;
if (!(strcmp(_lh, _rh) {op} 0)) {{
__{prefix}_assert_fail_{type}(file, line, "{comp}",
_lh, strlen(_lh), _rh, strlen(_rh));
}}
}} while (0)
"""
}
}
def mkdecls(outf, maxwidth=16):
outf.write("#include <stdio.h>\n")
outf.write("#include <stdbool.h>\n")
outf.write("#include <stdint.h>\n")
outf.write("#include <inttypes.h>\n")
outf.write("#include <signal.h>\n")
for type, desc in sorted(TYPE.items()):
format = {
'type': type.lower(), 'TYPE': type.upper(),
'ctype': desc['ctype'],
'prefix': PREFIX.lower(), 'PREFIX': PREFIX.upper(),
'maxwidth': maxwidth,
}
outf.write(re.sub('\s+', ' ',
desc['print'].strip().format(**format))+'\n')
outf.write(re.sub('\s+', ' ',
desc['fail'].strip().format(**format))+'\n')
for op, comp in sorted(COMP.items()):
format.update({
'comp': comp.lower(), 'COMP': comp.upper(),
'op': op,
})
outf.write(re.sub('\s+', ' ',
desc['assert'].strip().format(**format))+'\n')
def mkassert(type, comp, lh, rh, size=None):
format = {
'type': type.lower(), 'TYPE': type.upper(),
'comp': comp.lower(), 'COMP': comp.upper(),
'prefix': PREFIX.lower(), 'PREFIX': PREFIX.upper(),
'lh': lh.strip(' '),
'rh': rh.strip(' '),
'size': size,
}
if size:
return ((ASSERT + '(__FILE__, __LINE__, {lh}, {rh}, {size})')
.format(**format))
else:
return ((ASSERT + '(__FILE__, __LINE__, {lh}, {rh})')
.format(**format))
# simple recursive descent parser
LEX = {
'ws': [r'(?:\s|\n|#.*?\n|//.*?\n|/\*.*?\*/)+'],
'assert': PATTERN,
'string': [r'"(?:\\.|[^"])*"', r"'(?:\\.|[^'])\'"],
'arrow': ['=>'],
'paren': ['\(', '\)'],
'op': ['strcmp', 'memcmp', '->'],
'comp': ['==', '!=', '<=', '>=', '<', '>'],
'logic': ['\&\&', '\|\|'],
'sep': [':', ';', '\{', '\}', ','],
}
class ParseFailure(Exception):
def __init__(self, expected, found):
self.expected = expected
self.found = found
def __str__(self):
return "expected %r, found %s..." % (
self.expected, repr(self.found)[:70])
class Parse:
def __init__(self, inf, lexemes):
p = '|'.join('(?P<%s>%s)' % (n, '|'.join(l))
for n, l in lexemes.items())
p = re.compile(p, re.DOTALL)
data = inf.read()
tokens = []
while True:
m = p.search(data)
if m:
if m.start() > 0:
tokens.append((None, data[:m.start()]))
tokens.append((m.lastgroup, m.group()))
data = data[m.end():]
else:
tokens.append((None, data))
break
self.tokens = tokens
self.off = 0
def lookahead(self, *pattern):
if self.off < len(self.tokens):
token = self.tokens[self.off]
if token[0] in pattern or token[1] in pattern:
self.m = token[1]
return self.m
self.m = None
return self.m
def accept(self, *patterns):
m = self.lookahead(*patterns)
if m is not None:
self.off += 1
return m
def expect(self, *patterns):
m = self.accept(*patterns)
if not m:
raise ParseFailure(patterns, self.tokens[self.off:])
return m
def push(self):
return self.off
def pop(self, state):
self.off = state
def passert(p):
def pastr(p):
p.expect('assert') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
p.expect('strcmp') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
lh = pexpr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = pexpr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
comp = p.expect('comp') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
p.expect(')')
return mkassert('str', COMP[comp], lh, rh)
def pamem(p):
p.expect('assert') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
p.expect('memcmp') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
lh = pexpr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = pexpr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
size = pexpr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
comp = p.expect('comp') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
p.expect(')')
return mkassert('mem', COMP[comp], lh, rh, size)
def paint(p):
p.expect('assert') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
lh = pexpr(p) ; p.accept('ws')
comp = p.expect('comp') ; p.accept('ws')
rh = pexpr(p) ; p.accept('ws')
p.expect(')')
return mkassert('int', COMP[comp], lh, rh)
def pabool(p):
p.expect('assert') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
lh = pexprs(p) ; p.accept('ws')
p.expect(')')
return mkassert('bool', 'eq', lh, 'true')
def pa(p):
return p.expect('assert')
state = p.push()
lastf = None
for pa in [pastr, pamem, paint, pabool, pa]:
try:
return pa(p)
except ParseFailure as f:
p.pop(state)
lastf = f
else:
raise lastf
def pexpr(p):
res = []
while True:
if p.accept('('):
res.append(p.m)
while True:
res.append(pexprs(p))
if p.accept('sep'):
res.append(p.m)
else:
break
res.append(p.expect(')'))
elif p.lookahead('assert'):
res.append(passert(p))
elif p.accept('assert', 'ws', 'string', 'op', None):
res.append(p.m)
else:
return ''.join(res)
def pexprs(p):
res = []
while True:
res.append(pexpr(p))
if p.accept('comp', 'logic', ','):
res.append(p.m)
else:
return ''.join(res)
def pstmt(p):
ws = p.accept('ws') or ''
lh = pexprs(p)
if p.accept('=>'):
rh = pexprs(p)
return ws + mkassert('int', 'eq', lh, rh)
else:
return ws + lh
def main(args):
inf = open(args.input, 'r') if args.input else sys.stdin
outf = open(args.output, 'w') if args.output else sys.stdout
lexemes = LEX.copy()
if args.pattern:
lexemes['assert'] = args.pattern
p = Parse(inf, lexemes)
# write extra verbose asserts
mkdecls(outf, maxwidth=args.maxwidth)
if args.input:
outf.write("#line %d \"%s\"\n" % (1, args.input))
# parse and write out stmt at a time
try:
while True:
outf.write(pstmt(p))
if p.accept('sep'):
outf.write(p.m)
else:
break
except ParseFailure as f:
pass
for i in range(p.off, len(p.tokens)):
outf.write(p.tokens[i][1])
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(
description="Cpp step that increases assert verbosity")
parser.add_argument('input', nargs='?',
help="Input C file after cpp.")
parser.add_argument('-o', '--output', required=True,
help="Output C file.")
parser.add_argument('-p', '--pattern', action='append',
help="Patterns to search for starting an assert statement.")
parser.add_argument('--maxwidth', default=MAXWIDTH, type=int,
help="Maximum number of characters to display for strcmp and memcmp.")
main(parser.parse_args())
+1344
View File
File diff suppressed because it is too large Load Diff
+1276
View File
File diff suppressed because it is too large Load Diff
+1592
View File
File diff suppressed because it is too large Load Diff
+1262
View File
File diff suppressed because it is too large Load Diff
-61
View File
@@ -1,61 +0,0 @@
#!/usr/bin/env python2
# This script replaces prefixes of files, and symbols in that file.
# Useful for creating different versions of the codebase that don't
# conflict at compile time.
#
# example:
# $ ./scripts/prefix.py lfs22
import os
import os.path
import re
import glob
import itertools
import tempfile
import shutil
import subprocess
DEFAULT_PREFIX = "lfs2"
def subn(from_prefix, to_prefix, name):
name, count1 = re.subn('\\b'+from_prefix, to_prefix, name)
name, count2 = re.subn('\\b'+from_prefix.upper(), to_prefix.upper(), name)
name, count3 = re.subn('\\B-D'+from_prefix.upper(),
'-D'+to_prefix.upper(), name)
return name, count1+count2+count3
def main(from_prefix, to_prefix=None, files=None):
if not to_prefix:
from_prefix, to_prefix = DEFAULT_PREFIX, from_prefix
if not files:
files = subprocess.check_output([
'git', 'ls-tree', '-r', '--name-only', 'HEAD']).split()
for oldname in files:
# Rename any matching file names
newname, namecount = subn(from_prefix, to_prefix, oldname)
if namecount:
subprocess.check_call(['git', 'mv', oldname, newname])
# Rename any prefixes in file
count = 0
with open(newname+'~', 'w') as tempf:
with open(newname) as newf:
for line in newf:
line, n = subn(from_prefix, to_prefix, line)
count += n
tempf.write(line)
shutil.copystat(newname, newname+'~')
os.rename(newname+'~', newname)
subprocess.check_call(['git', 'add', newname])
# Summary
print '%s: %d replacements' % (
'%s -> %s' % (oldname, newname) if namecount else oldname,
count)
if __name__ == "__main__":
import sys
sys.exit(main(*sys.argv[1:]))
+452
View File
@@ -0,0 +1,452 @@
#!/usr/bin/env python3
#
# Preprocessor that makes asserts easier to debug.
#
# Example:
# ./scripts/prettyasserts.py -p LFS2_ASSERT lfs2.c -o lfs2.a.c
#
# Copyright (c) 2022, The littlefs authors.
# Copyright (c) 2020, Arm Limited. All rights reserved.
# SPDX-License-Identifier: BSD-3-Clause
#
import re
import sys
# NOTE the use of macros here helps keep a consistent stack depth which
# tools may rely on.
#
# If compilation errors are noisy consider using -ftrack-macro-expansion=0.
#
LIMIT = 16
CMP = {
'==': 'eq',
'!=': 'ne',
'<=': 'le',
'>=': 'ge',
'<': 'lt',
'>': 'gt',
}
LEXEMES = {
'ws': [r'(?:\s|\n|#.*?\n|//.*?\n|/\*.*?\*/)+'],
'assert': ['assert'],
'arrow': ['=>'],
'string': [r'"(?:\\.|[^"])*"', r"'(?:\\.|[^'])\'"],
'paren': ['\(', '\)'],
'cmp': CMP.keys(),
'logic': ['\&\&', '\|\|'],
'sep': [':', ';', '\{', '\}', ','],
'op': ['->'], # specifically ops that conflict with cmp
}
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def write_header(f, limit=LIMIT):
f.writeln("// Generated by %s:" % sys.argv[0])
f.writeln("//")
f.writeln("// %s" % ' '.join(sys.argv))
f.writeln("//")
f.writeln()
f.writeln("#include <stdbool.h>")
f.writeln("#include <stdint.h>")
f.writeln("#include <inttypes.h>")
f.writeln("#include <stdio.h>")
f.writeln("#include <string.h>")
f.writeln("#include <signal.h>")
# give source a chance to define feature macros
f.writeln("#undef _FEATURES_H")
f.writeln()
# write print macros
f.writeln("__attribute__((unused))")
f.writeln("static void __pretty_assert_print_bool(")
f.writeln(" const void *v, size_t size) {")
f.writeln(" (void)size;")
f.writeln(" printf(\"%s\", *(const bool*)v ? \"true\" : \"false\");")
f.writeln("}")
f.writeln()
f.writeln("__attribute__((unused))")
f.writeln("static void __pretty_assert_print_int(")
f.writeln(" const void *v, size_t size) {")
f.writeln(" (void)size;")
f.writeln(" printf(\"%\"PRIiMAX, *(const intmax_t*)v);")
f.writeln("}")
f.writeln()
f.writeln("__attribute__((unused))")
f.writeln("static void __pretty_assert_print_mem(")
f.writeln(" const void *v, size_t size) {")
f.writeln(" const uint8_t *v_ = v;")
f.writeln(" printf(\"\\\"\");")
f.writeln(" for (size_t i = 0; i < size && i < %d; i++) {" % limit)
f.writeln(" if (v_[i] >= ' ' && v_[i] <= '~') {")
f.writeln(" printf(\"%c\", v_[i]);")
f.writeln(" } else {")
f.writeln(" printf(\"\\\\x%02x\", v_[i]);")
f.writeln(" }")
f.writeln(" }")
f.writeln(" if (size > %d) {" % limit)
f.writeln(" printf(\"...\");")
f.writeln(" }")
f.writeln(" printf(\"\\\"\");")
f.writeln("}")
f.writeln()
f.writeln("__attribute__((unused))")
f.writeln("static void __pretty_assert_print_str(")
f.writeln(" const void *v, size_t size) {")
f.writeln(" __pretty_assert_print_mem(v, size);")
f.writeln("}")
f.writeln()
f.writeln("__attribute__((unused, noinline))")
f.writeln("static void __pretty_assert_fail(")
f.writeln(" const char *file, int line,")
f.writeln(" void (*type_print_cb)(const void*, size_t),")
f.writeln(" const char *cmp,")
f.writeln(" const void *lh, size_t lsize,")
f.writeln(" const void *rh, size_t rsize) {")
f.writeln(" printf(\"%s:%d:assert: assert failed with \", file, line);")
f.writeln(" type_print_cb(lh, lsize);")
f.writeln(" printf(\", expected %s \", cmp);")
f.writeln(" type_print_cb(rh, rsize);")
f.writeln(" printf(\"\\n\");")
f.writeln(" fflush(NULL);")
f.writeln(" raise(SIGABRT);")
f.writeln("}")
f.writeln()
# write assert macros
for op, cmp in sorted(CMP.items()):
f.writeln("#define __PRETTY_ASSERT_BOOL_%s(lh, rh) do { \\"
% cmp.upper())
f.writeln(" bool _lh = !!(lh); \\")
f.writeln(" bool _rh = !!(rh); \\")
f.writeln(" if (!(_lh %s _rh)) { \\" % op)
f.writeln(" __pretty_assert_fail( \\")
f.writeln(" __FILE__, __LINE__, \\")
f.writeln(" __pretty_assert_print_bool, \"%s\", \\"
% cmp)
f.writeln(" &_lh, 0, \\")
f.writeln(" &_rh, 0); \\")
f.writeln(" } \\")
f.writeln("} while (0)")
for op, cmp in sorted(CMP.items()):
f.writeln("#define __PRETTY_ASSERT_INT_%s(lh, rh) do { \\"
% cmp.upper())
f.writeln(" __typeof__(lh) _lh = lh; \\")
f.writeln(" __typeof__(lh) _rh = rh; \\")
f.writeln(" if (!(_lh %s _rh)) { \\" % op)
f.writeln(" __pretty_assert_fail( \\")
f.writeln(" __FILE__, __LINE__, \\")
f.writeln(" __pretty_assert_print_int, \"%s\", \\"
% cmp)
f.writeln(" &(intmax_t){_lh}, 0, \\")
f.writeln(" &(intmax_t){_rh}, 0); \\")
f.writeln(" } \\")
f.writeln("} while (0)")
for op, cmp in sorted(CMP.items()):
f.writeln("#define __PRETTY_ASSERT_MEM_%s(lh, rh, size) do { \\"
% cmp.upper())
f.writeln(" const void *_lh = lh; \\")
f.writeln(" const void *_rh = rh; \\")
f.writeln(" if (!(memcmp(_lh, _rh, size) %s 0)) { \\" % op)
f.writeln(" __pretty_assert_fail( \\")
f.writeln(" __FILE__, __LINE__, \\")
f.writeln(" __pretty_assert_print_mem, \"%s\", \\"
% cmp)
f.writeln(" _lh, size, \\")
f.writeln(" _rh, size); \\")
f.writeln(" } \\")
f.writeln("} while (0)")
for op, cmp in sorted(CMP.items()):
f.writeln("#define __PRETTY_ASSERT_STR_%s(lh, rh) do { \\"
% cmp.upper())
f.writeln(" const char *_lh = lh; \\")
f.writeln(" const char *_rh = rh; \\")
f.writeln(" if (!(strcmp(_lh, _rh) %s 0)) { \\" % op)
f.writeln(" __pretty_assert_fail( \\")
f.writeln(" __FILE__, __LINE__, \\")
f.writeln(" __pretty_assert_print_str, \"%s\", \\"
% cmp)
f.writeln(" _lh, strlen(_lh), \\")
f.writeln(" _rh, strlen(_rh)); \\")
f.writeln(" } \\")
f.writeln("} while (0)")
f.writeln()
f.writeln()
def mkassert(type, cmp, lh, rh, size=None):
if size is not None:
return ("__PRETTY_ASSERT_%s_%s(%s, %s, %s)"
% (type.upper(), cmp.upper(), lh, rh, size))
else:
return ("__PRETTY_ASSERT_%s_%s(%s, %s)"
% (type.upper(), cmp.upper(), lh, rh))
# simple recursive descent parser
class ParseFailure(Exception):
def __init__(self, expected, found):
self.expected = expected
self.found = found
def __str__(self):
return "expected %r, found %s..." % (
self.expected, repr(self.found)[:70])
class Parser:
def __init__(self, in_f, lexemes=LEXEMES):
p = '|'.join('(?P<%s>%s)' % (n, '|'.join(l))
for n, l in lexemes.items())
p = re.compile(p, re.DOTALL)
data = in_f.read()
tokens = []
line = 1
col = 0
while True:
m = p.search(data)
if m:
if m.start() > 0:
tokens.append((None, data[:m.start()], line, col))
tokens.append((m.lastgroup, m.group(), line, col))
data = data[m.end():]
else:
tokens.append((None, data, line, col))
break
self.tokens = tokens
self.off = 0
def lookahead(self, *pattern):
if self.off < len(self.tokens):
token = self.tokens[self.off]
if token[0] in pattern or token[1] in pattern:
self.m = token[1]
return self.m
self.m = None
return self.m
def accept(self, *patterns):
m = self.lookahead(*patterns)
if m is not None:
self.off += 1
return m
def expect(self, *patterns):
m = self.accept(*patterns)
if not m:
raise ParseFailure(patterns, self.tokens[self.off:])
return m
def push(self):
return self.off
def pop(self, state):
self.off = state
def p_assert(p):
state = p.push()
# assert(memcmp(a,b,size) cmp 0)?
try:
p.expect('assert') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
p.expect('memcmp') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
lh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
size = p_expr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
cmp = p.expect('cmp') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
p.expect(')')
return mkassert('mem', CMP[cmp], lh, rh, size)
except ParseFailure:
p.pop(state)
# assert(strcmp(a,b) cmp 0)?
try:
p.expect('assert') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
p.expect('strcmp') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
lh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = p_expr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
cmp = p.expect('cmp') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
p.expect(')')
return mkassert('str', CMP[cmp], lh, rh)
except ParseFailure:
p.pop(state)
# assert(a cmp b)?
try:
p.expect('assert') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
lh = p_expr(p) ; p.accept('ws')
cmp = p.expect('cmp') ; p.accept('ws')
rh = p_expr(p) ; p.accept('ws')
p.expect(')')
return mkassert('int', CMP[cmp], lh, rh)
except ParseFailure:
p.pop(state)
# assert(a)?
p.expect('assert') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
lh = p_exprs(p) ; p.accept('ws')
p.expect(')')
return mkassert('bool', 'eq', lh, 'true')
def p_expr(p):
res = []
while True:
if p.accept('('):
res.append(p.m)
while True:
res.append(p_exprs(p))
if p.accept('sep'):
res.append(p.m)
else:
break
res.append(p.expect(')'))
elif p.lookahead('assert'):
state = p.push()
try:
res.append(p_assert(p))
except ParseFailure:
p.pop(state)
res.append(p.expect('assert'))
elif p.accept('string', 'op', 'ws', None):
res.append(p.m)
else:
return ''.join(res)
def p_exprs(p):
res = []
while True:
res.append(p_expr(p))
if p.accept('cmp', 'logic', ','):
res.append(p.m)
else:
return ''.join(res)
def p_stmt(p):
ws = p.accept('ws') or ''
# memcmp(lh,rh,size) => 0?
if p.lookahead('memcmp'):
state = p.push()
try:
p.expect('memcmp') ; p.accept('ws')
p.expect('(') ; p.accept('ws')
lh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
size = p_expr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
p.expect('=>') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
return ws + mkassert('mem', 'eq', lh, rh, size)
except ParseFailure:
p.pop(state)
# strcmp(lh,rh) => 0?
if p.lookahead('strcmp'):
state = p.push()
try:
p.expect('strcmp') ; p.accept('ws') ; p.expect('(') ; p.accept('ws')
lh = p_expr(p) ; p.accept('ws')
p.expect(',') ; p.accept('ws')
rh = p_expr(p) ; p.accept('ws')
p.expect(')') ; p.accept('ws')
p.expect('=>') ; p.accept('ws')
p.expect('0') ; p.accept('ws')
return ws + mkassert('str', 'eq', lh, rh)
except ParseFailure:
p.pop(state)
# lh => rh?
lh = p_exprs(p)
if p.accept('=>'):
rh = p_exprs(p)
return ws + mkassert('int', 'eq', lh, rh)
else:
return ws + lh
def main(input=None, output=None, pattern=[], limit=LIMIT):
with openio(input or '-', 'r') as in_f:
# create parser
lexemes = LEXEMES.copy()
lexemes['assert'] += pattern
p = Parser(in_f, lexemes)
with openio(output or '-', 'w') as f:
def writeln(s=''):
f.write(s)
f.write('\n')
f.writeln = writeln
# write extra verbose asserts
write_header(f, limit=limit)
if input is not None:
f.writeln("#line %d \"%s\"" % (1, input))
# parse and write out stmt at a time
try:
while True:
f.write(p_stmt(p))
if p.accept('sep'):
f.write(p.m)
else:
break
except ParseFailure as e:
print('warning: %s' % e)
pass
for i in range(p.off, len(p.tokens)):
f.write(p.tokens[i][1])
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Preprocessor that makes asserts easier to debug.",
allow_abbrev=False)
parser.add_argument(
'input',
help="Input C file.")
parser.add_argument(
'-o', '--output',
required=True,
help="Output C file.")
parser.add_argument(
'-p', '--pattern',
action='append',
help="Regex patterns to search for starting an assert statement. This"
" implicitly includes \"assert\" and \"=>\".")
parser.add_argument(
'-l', '--limit',
type=lambda x: int(x, 0),
default=LIMIT,
help="Maximum number of characters to display in strcmp and memcmp. "
"Defaults to %r." % LIMIT)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
+47 -15
View File
@@ -24,6 +24,8 @@ TAG_TYPES = {
'gstate': (0x700, 0x700),
'movestate': (0x7ff, 0x7ff),
'crc': (0x700, 0x500),
'ccrc': (0x780, 0x500),
'fcrc': (0x7ff, 0x5ff),
}
class Tag:
@@ -99,7 +101,16 @@ class Tag:
return struct.unpack('b', struct.pack('B', self.chunk))[0]
def is_(self, type):
return (self.type & TAG_TYPES[type][0]) == TAG_TYPES[type][1]
try:
if ' ' in type:
type1, type3 = type.split()
return (self.is_(type1) and
(self.type & ~TAG_TYPES[type1][0]) == int(type3, 0))
return self.type == int(type, 0)
except (ValueError, KeyError):
return (self.type & TAG_TYPES[type][0]) == TAG_TYPES[type][1]
def mkmask(self):
return Tag(
@@ -109,14 +120,20 @@ class Tag:
def chid(self, nid):
ntag = Tag(self.type, nid, self.size)
if hasattr(self, 'off'): ntag.off = self.off
if hasattr(self, 'data'): ntag.data = self.data
if hasattr(self, 'crc'): ntag.crc = self.crc
if hasattr(self, 'off'): ntag.off = self.off
if hasattr(self, 'data'): ntag.data = self.data
if hasattr(self, 'ccrc'): ntag.crc = self.crc
if hasattr(self, 'erased'): ntag.erased = self.erased
return ntag
def typerepr(self):
if self.is_('crc') and getattr(self, 'crc', 0xffffffff) != 0xffffffff:
return 'crc (bad)'
if (self.is_('ccrc')
and getattr(self, 'ccrc', 0xffffffff) != 0xffffffff):
crc_status = ' (bad)'
elif self.is_('fcrc') and getattr(self, 'erased', False):
crc_status = ' (era)'
else:
crc_status = ''
reverse_types = {v: k for k, v in TAG_TYPES.items()}
for prefix in range(12):
@@ -124,12 +141,12 @@ class Tag:
if (mask, self.type & mask) in reverse_types:
type = reverse_types[mask, self.type & mask]
if prefix > 0:
return '%s %#0*x' % (
type, prefix//4, self.type & ((1 << prefix)-1))
return '%s %#x%s' % (
type, self.type & ((1 << prefix)-1), crc_status)
else:
return type
return '%s%s' % (type, crc_status)
else:
return '%02x' % self.type
return '%02x%s' % (self.type, crc_status)
def idrepr(self):
return repr(self.id) if self.id != 0x3ff else '.'
@@ -172,6 +189,8 @@ class MetadataPair:
self.rev, = struct.unpack('<I', block[0:4])
crc = binascii.crc32(block[0:4])
fcrctag = None
fcrcdata = None
# parse tags
corrupt = False
@@ -182,11 +201,11 @@ class MetadataPair:
while len(block) - off >= 4:
ntag, = struct.unpack('>I', block[off:off+4])
tag = Tag(int(tag) ^ ntag)
tag = Tag((int(tag) ^ ntag) & 0x7fffffff)
tag.off = off + 4
tag.data = block[off+4:off+tag.dsize]
if tag.is_('crc'):
crc = binascii.crc32(block[off:off+4+4], crc)
if tag.is_('ccrc'):
crc = binascii.crc32(block[off:off+2*4], crc)
else:
crc = binascii.crc32(block[off:off+tag.dsize], crc)
tag.crc = crc
@@ -194,16 +213,29 @@ class MetadataPair:
self.all_.append(tag)
if tag.is_('crc'):
if tag.is_('fcrc') and len(tag.data) == 8:
fcrctag = tag
fcrcdata = struct.unpack('<II', tag.data)
elif tag.is_('ccrc'):
# is valid commit?
if crc != 0xffffffff:
corrupt = True
if not corrupt:
self.log = self.all_.copy()
# end of commit?
if fcrcdata:
fcrcsize, fcrc = fcrcdata
fcrc_ = 0xffffffff ^ binascii.crc32(
block[off:off+fcrcsize])
if fcrc_ == fcrc:
fcrctag.erased = True
corrupt = True
# reset tag parsing
crc = 0
tag = Tag(int(tag) ^ ((tag.type & 1) << 31))
fcrctag = None
fcrcdata = None
# find active ids
self.ids = list(it.takewhile(
@@ -280,7 +312,7 @@ class MetadataPair:
f.write('\n')
for tag in tags:
f.write("%08x: %08x %-13s %4s %4s" % (
f.write("%08x: %08x %-14s %3s %4s" % (
tag.off, tag,
tag.typerepr(), tag.idrepr(), tag.sizerepr()))
if truncate:
+634 -329
View File
File diff suppressed because it is too large Load Diff
+580 -259
View File
@@ -2,49 +2,183 @@
#
# Script to find struct sizes.
#
# Example:
# ./scripts/structs.py lfs2.o lfs2_util.o -Ssize
#
# Copyright (c) 2022, The littlefs authors.
# SPDX-License-Identifier: BSD-3-Clause
#
import os
import glob
import itertools as it
import subprocess as sp
import shlex
import re
import csv
import collections as co
import csv
import difflib
import itertools as it
import math as m
import os
import re
import shlex
import subprocess as sp
OBJ_PATHS = ['*.o']
OBJDUMP_PATH = ['objdump']
def collect(paths, **args):
decl_pattern = re.compile(
# integer fields
class Int(co.namedtuple('Int', 'x')):
__slots__ = ()
def __new__(cls, x=0):
if isinstance(x, Int):
return x
if isinstance(x, str):
try:
x = int(x, 0)
except ValueError:
# also accept +-∞ and +-inf
if re.match('^\s*\+?\s*(?:∞|inf)\s*$', x):
x = m.inf
elif re.match('^\s*-\s*(?:∞|inf)\s*$', x):
x = -m.inf
else:
raise
assert isinstance(x, int) or m.isinf(x), x
return super().__new__(cls, x)
def __str__(self):
if self.x == m.inf:
return ''
elif self.x == -m.inf:
return '-∞'
else:
return str(self.x)
def __int__(self):
assert not m.isinf(self.x)
return self.x
def __float__(self):
return float(self.x)
none = '%7s' % '-'
def table(self):
return '%7s' % (self,)
diff_none = '%7s' % '-'
diff_table = table
def diff_diff(self, other):
new = self.x if self else 0
old = other.x if other else 0
diff = new - old
if diff == +m.inf:
return '%7s' % '+∞'
elif diff == -m.inf:
return '%7s' % '-∞'
else:
return '%+7d' % diff
def ratio(self, other):
new = self.x if self else 0
old = other.x if other else 0
if m.isinf(new) and m.isinf(old):
return 0.0
elif m.isinf(new):
return +m.inf
elif m.isinf(old):
return -m.inf
elif not old and not new:
return 0.0
elif not old:
return 1.0
else:
return (new-old) / old
def __add__(self, other):
return self.__class__(self.x + other.x)
def __sub__(self, other):
return self.__class__(self.x - other.x)
def __mul__(self, other):
return self.__class__(self.x * other.x)
# struct size results
class StructResult(co.namedtuple('StructResult', ['file', 'struct', 'size'])):
_by = ['file', 'struct']
_fields = ['size']
_sort = ['size']
_types = {'size': Int}
__slots__ = ()
def __new__(cls, file='', struct='', size=0):
return super().__new__(cls, file, struct,
Int(size))
def __add__(self, other):
return StructResult(self.file, self.struct,
self.size + other.size)
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def collect(obj_paths, *,
objdump_path=OBJDUMP_PATH,
sources=None,
everything=False,
internal=False,
**args):
line_pattern = re.compile(
'^\s+(?P<no>[0-9]+)'
'\s+(?P<dir>[0-9]+)'
'(?:\s+(?P<dir>[0-9]+))?'
'\s+.*'
'\s+(?P<file>[^\s]+)$')
struct_pattern = re.compile(
'^(?:.*DW_TAG_(?P<tag>[a-z_]+).*'
'|^.*DW_AT_name.*:\s*(?P<name>[^:\s]+)\s*'
'|^.*DW_AT_decl_file.*:\s*(?P<decl>[0-9]+)\s*'
'|^.*DW_AT_byte_size.*:\s*(?P<size>[0-9]+)\s*)$')
'\s+(?P<path>[^\s]+)$')
info_pattern = re.compile(
'^(?:.*(?P<tag>DW_TAG_[a-z_]+).*'
'|.*DW_AT_name.*:\s*(?P<name>[^:\s]+)\s*'
'|.*DW_AT_decl_file.*:\s*(?P<file>[0-9]+)\s*'
'|.*DW_AT_byte_size.*:\s*(?P<size>[0-9]+)\s*)$')
results = co.defaultdict(lambda: 0)
for path in paths:
# find decl, we want to filter by structs in .h files
decls = {}
# note objdump-tool may contain extra args
cmd = args['objdump_tool'] + ['--dwarf=rawline', path]
results = []
for path in obj_paths:
# find files, we want to filter by structs in .h files
dirs = {}
files = {}
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=rawline', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
errors='replace',
close_fds=False)
for line in proc.stdout:
# find file numbers
m = decl_pattern.match(line)
# note that files contain references to dirs, which we
# dereference as soon as we see them as each file table follows a
# dir table
m = line_pattern.match(line)
if m:
decls[int(m.group('no'))] = m.group('file')
if not m.group('dir'):
# found a directory entry
dirs[int(m.group('no'))] = m.group('path')
else:
# found a file entry
dir = int(m.group('dir'))
if dir in dirs:
files[int(m.group('no'))] = os.path.join(
dirs[dir],
m.group('path'))
else:
files[int(m.group('no'))] = m.group('path')
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
@@ -53,40 +187,39 @@ def collect(paths, **args):
sys.exit(-1)
# collect structs as we parse dwarf info
found = False
name = None
decl = None
size = None
# note objdump-tool may contain extra args
cmd = args['objdump_tool'] + ['--dwarf=info', path]
results_ = []
is_struct = False
s_name = None
s_file = None
s_size = None
# note objdump-path may contain extra args
cmd = objdump_path + ['--dwarf=info', path]
if args.get('verbose'):
print(' '.join(shlex.quote(c) for c in cmd))
proc = sp.Popen(cmd,
stdout=sp.PIPE,
stderr=sp.PIPE if not args.get('verbose') else None,
universal_newlines=True,
errors='replace')
errors='replace',
close_fds=False)
for line in proc.stdout:
# state machine here to find structs
m = struct_pattern.match(line)
m = info_pattern.match(line)
if m:
if m.group('tag'):
if (name is not None
and decl is not None
and size is not None):
decl = decls.get(decl, '?')
results[(decl, name)] = size
found = (m.group('tag') == 'structure_type')
name = None
decl = None
size = None
elif found and m.group('name'):
name = m.group('name')
elif found and name and m.group('decl'):
decl = int(m.group('decl'))
elif found and name and m.group('size'):
size = int(m.group('size'))
if is_struct:
file = files.get(s_file, '?')
results_.append(StructResult(file, s_name, s_size))
is_struct = (m.group('tag') == 'DW_TAG_structure_type')
elif m.group('name'):
s_name = m.group('name')
elif m.group('file'):
s_file = int(m.group('file'))
elif m.group('size'):
s_size = int(m.group('size'))
if is_struct:
file = files.get(s_file, '?')
results_.append(StructResult(file, s_name, s_size))
proc.wait()
if proc.returncode != 0:
if not args.get('verbose'):
@@ -94,238 +227,426 @@ def collect(paths, **args):
sys.stdout.write(line)
sys.exit(-1)
flat_results = []
for (file, struct), size in results.items():
# map to source files
if args.get('build_dir'):
file = re.sub('%s/*' % re.escape(args['build_dir']), '', file)
# only include structs declared in header files in the current
# directory, ignore internal-only # structs (these are represented
# in other measurements)
if not args.get('everything'):
if not file.endswith('.h'):
continue
# replace .o with .c, different scripts report .o/.c, we need to
# choose one if we want to deduplicate csv files
file = re.sub('\.o$', '.c', file)
flat_results.append((file, struct, size))
return flat_results
def main(**args):
def openio(path, mode='r'):
if path == '-':
if 'r' in mode:
return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
for r in results_:
# ignore filtered sources
if sources is not None:
if not any(
os.path.abspath(r.file) == os.path.abspath(s)
for s in sources):
continue
else:
return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
else:
return open(path, mode)
# default to only cwd
if not everything and not os.path.commonpath([
os.getcwd(),
os.path.abspath(r.file)]) == os.getcwd():
continue
# find sizes
if not args.get('use', None):
# find .o files
paths = []
for path in args['obj_paths']:
if os.path.isdir(path):
path = path + '/*.o'
# limit to .h files unless --internal
if not internal and not r.file.endswith('.h'):
continue
for path in glob.glob(path):
paths.append(path)
# simplify path
if os.path.commonpath([
os.getcwd(),
os.path.abspath(r.file)]) == os.getcwd():
file = os.path.relpath(r.file)
else:
file = os.path.abspath(r.file)
if not paths:
print('no .obj files found in %r?' % args['obj_paths'])
results.append(r._replace(file=file))
return results
def fold(Result, results, *,
by=None,
defines=None,
**_):
if by is None:
by = Result._by
for k in it.chain(by or [], (k for k, _ in defines or [])):
if k not in Result._by and k not in Result._fields:
print("error: could not find field %r?" % k)
sys.exit(-1)
results = collect(paths, **args)
# filter by matching defines
if defines is not None:
results_ = []
for r in results:
if all(getattr(r, k) in vs for k, vs in defines):
results_.append(r)
results = results_
# organize results into conflicts
folding = co.OrderedDict()
for r in results:
name = tuple(getattr(r, k) for k in by)
if name not in folding:
folding[name] = []
folding[name].append(r)
# merge conflicts
folded = []
for name, rs in folding.items():
folded.append(sum(rs[1:], start=rs[0]))
return folded
def table(Result, results, diff_results=None, *,
by=None,
fields=None,
sort=None,
summary=False,
all=False,
percent=False,
**_):
all_, all = all, __builtins__.all
if by is None:
by = Result._by
if fields is None:
fields = Result._fields
types = Result._types
# fold again
results = fold(Result, results, by=by)
if diff_results is not None:
diff_results = fold(Result, diff_results, by=by)
# organize by name
table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in results}
diff_table = {
','.join(str(getattr(r, k) or '') for k in by): r
for r in diff_results or []}
names = list(table.keys() | diff_table.keys())
# sort again, now with diff info, note that python's sort is stable
names.sort()
if diff_results is not None:
names.sort(key=lambda n: tuple(
types[k].ratio(
getattr(table.get(n), k, None),
getattr(diff_table.get(n), k, None))
for k in fields),
reverse=True)
if sort:
for k, reverse in reversed(sort):
names.sort(
key=lambda n: tuple(
(getattr(table[n], k),)
if getattr(table.get(n), k, None) is not None else ()
for k in ([k] if k else [
k for k in Result._sort if k in fields])),
reverse=reverse ^ (not k or k in Result._fields))
# build up our lines
lines = []
# header
header = []
header.append('%s%s' % (
','.join(by),
' (%d added, %d removed)' % (
sum(1 for n in table if n not in diff_table),
sum(1 for n in diff_table if n not in table))
if diff_results is not None and not percent else '')
if not summary else '')
if diff_results is None:
for k in fields:
header.append(k)
elif percent:
for k in fields:
header.append(k)
else:
for k in fields:
header.append('o'+k)
for k in fields:
header.append('n'+k)
for k in fields:
header.append('d'+k)
header.append('')
lines.append(header)
def table_entry(name, r, diff_r=None, ratios=[]):
entry = []
entry.append(name)
if diff_results is None:
for k in fields:
entry.append(getattr(r, k).table()
if getattr(r, k, None) is not None
else types[k].none)
elif percent:
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
else:
for k in fields:
entry.append(getattr(diff_r, k).diff_table()
if getattr(diff_r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(getattr(r, k).diff_table()
if getattr(r, k, None) is not None
else types[k].diff_none)
for k in fields:
entry.append(types[k].diff_diff(
getattr(r, k, None),
getattr(diff_r, k, None)))
if diff_results is None:
entry.append('')
elif percent:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios))
else:
entry.append(' (%s)' % ', '.join(
'+∞%' if t == +m.inf
else '-∞%' if t == -m.inf
else '%+.1f%%' % (100*t)
for t in ratios
if t)
if any(ratios) else '')
return entry
# entries
if not summary:
for name in names:
r = table.get(name)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = diff_table.get(name)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
if not all_ and not any(ratios):
continue
lines.append(table_entry(name, r, diff_r, ratios))
# total
r = next(iter(fold(Result, results, by=[])), None)
if diff_results is None:
diff_r = None
ratios = None
else:
diff_r = next(iter(fold(Result, diff_results, by=[])), None)
ratios = [
types[k].ratio(
getattr(r, k, None),
getattr(diff_r, k, None))
for k in fields]
lines.append(table_entry('TOTAL', r, diff_r, ratios))
# find the best widths, note that column 0 contains the names and column -1
# the ratios, so those are handled a bit differently
widths = [
((max(it.chain([w], (len(l[i]) for l in lines)))+1+4-1)//4)*4-1
for w, i in zip(
it.chain([23], it.repeat(7)),
range(len(lines[0])-1))]
# print our table
for line in lines:
print('%-*s %s%s' % (
widths[0], line[0],
' '.join('%*s' % (w, x)
for w, x in zip(widths[1:], line[1:-1])),
line[-1]))
def main(obj_paths, *,
by=None,
fields=None,
defines=None,
sort=None,
**args):
# find sizes
if not args.get('use', None):
results = collect(obj_paths, **args)
else:
results = []
with openio(args['use']) as f:
r = csv.DictReader(f)
results = [
( result['file'],
result['name'],
int(result['struct_size']))
for result in r
if result.get('struct_size') not in {None, ''}]
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('struct_'+k in r and r['struct_'+k].strip()
for k in StructResult._fields):
continue
try:
results.append(StructResult(
**{k: r[k] for k in StructResult._by
if k in r and r[k].strip()},
**{k: r['struct_'+k]
for k in StructResult._fields
if 'struct_'+k in r
and r['struct_'+k].strip()}))
except TypeError:
pass
total = 0
for _, _, size in results:
total += size
# fold
results = fold(StructResult, results, by=by, defines=defines)
# find previous results?
if args.get('diff'):
try:
with openio(args['diff']) as f:
r = csv.DictReader(f)
prev_results = [
( result['file'],
result['name'],
int(result['struct_size']))
for result in r
if result.get('struct_size') not in {None, ''}]
except FileNotFoundError:
prev_results = []
prev_total = 0
for _, _, size in prev_results:
prev_total += size
# sort, note that python's sort is stable
results.sort()
if sort:
for k, reverse in reversed(sort):
results.sort(
key=lambda r: tuple(
(getattr(r, k),) if getattr(r, k) is not None else ()
for k in ([k] if k else StructResult._sort)),
reverse=reverse ^ (not k or k in StructResult._fields))
# write results to CSV
if args.get('output'):
merged_results = co.defaultdict(lambda: {})
other_fields = []
# merge?
if args.get('merge'):
try:
with openio(args['merge']) as f:
r = csv.DictReader(f)
for result in r:
file = result.pop('file', '')
struct = result.pop('name', '')
result.pop('struct_size', None)
merged_results[(file, struct)] = result
other_fields = result.keys()
except FileNotFoundError:
pass
for file, struct, size in results:
merged_results[(file, struct)]['struct_size'] = size
with openio(args['output'], 'w') as f:
w = csv.DictWriter(f, ['file', 'name', *other_fields, 'struct_size'])
w.writeheader()
for (file, struct), result in sorted(merged_results.items()):
w.writerow({'file': file, 'name': struct, **result})
writer = csv.DictWriter(f,
(by if by is not None else StructResult._by)
+ ['struct_'+k for k in (
fields if fields is not None else StructResult._fields)])
writer.writeheader()
for r in results:
writer.writerow(
{k: getattr(r, k) for k in (
by if by is not None else StructResult._by)}
| {'struct_'+k: getattr(r, k) for k in (
fields if fields is not None else StructResult._fields)})
# print results
def dedup_entries(results, by='name'):
entries = co.defaultdict(lambda: 0)
for file, struct, size in results:
entry = (file if by == 'file' else struct)
entries[entry] += size
return entries
# find previous results?
if args.get('diff'):
diff_results = []
try:
with openio(args['diff']) as f:
reader = csv.DictReader(f, restval='')
for r in reader:
if not any('struct_'+k in r and r['struct_'+k].strip()
for k in StructResult._fields):
continue
try:
diff_results.append(StructResult(
**{k: r[k] for k in StructResult._by
if k in r and r[k].strip()},
**{k: r['struct_'+k]
for k in StructResult._fields
if 'struct_'+k in r
and r['struct_'+k].strip()}))
except TypeError:
pass
except FileNotFoundError:
pass
def diff_entries(olds, news):
diff = co.defaultdict(lambda: (0, 0, 0, 0))
for name, new in news.items():
diff[name] = (0, new, new, 1.0)
for name, old in olds.items():
_, new, _, _ = diff[name]
diff[name] = (old, new, new-old, (new-old)/old if old else 1.0)
return diff
# fold
diff_results = fold(StructResult, diff_results, by=by, defines=defines)
def sorted_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1], x))
else:
return sorted(entries)
# print table
if not args.get('quiet'):
table(StructResult, results,
diff_results if args.get('diff') else None,
by=by if by is not None else ['struct'],
fields=fields,
sort=sort,
**args)
def sorted_diff_entries(entries):
if args.get('size_sort'):
return sorted(entries, key=lambda x: (-x[1][1], x))
elif args.get('reverse_size_sort'):
return sorted(entries, key=lambda x: (+x[1][1], x))
else:
return sorted(entries, key=lambda x: (-x[1][3], x))
def print_header(by=''):
if not args.get('diff'):
print('%-36s %7s' % (by, 'size'))
else:
print('%-36s %7s %7s %7s' % (by, 'old', 'new', 'diff'))
def print_entry(name, size):
print("%-36s %7d" % (name, size))
def print_diff_entry(name, old, new, diff, ratio):
print("%-36s %7s %7s %+7d%s" % (name,
old or "-",
new or "-",
diff,
' (%+.1f%%)' % (100*ratio) if ratio else ''))
def print_entries(by='name'):
entries = dedup_entries(results, by=by)
if not args.get('diff'):
print_header(by=by)
for name, size in sorted_entries(entries.items()):
print_entry(name, size)
else:
prev_entries = dedup_entries(prev_results, by=by)
diff = diff_entries(prev_entries, entries)
print_header(by='%s (%d added, %d removed)' % (by,
sum(1 for old, _, _, _ in diff.values() if not old),
sum(1 for _, new, _, _ in diff.values() if not new)))
for name, (old, new, diff, ratio) in sorted_diff_entries(
diff.items()):
if ratio or args.get('all'):
print_diff_entry(name, old, new, diff, ratio)
def print_totals():
if not args.get('diff'):
print_entry('TOTAL', total)
else:
ratio = (0.0 if not prev_total and not total
else 1.0 if not prev_total
else (total-prev_total)/prev_total)
print_diff_entry('TOTAL',
prev_total, total,
total-prev_total,
ratio)
if args.get('quiet'):
pass
elif args.get('summary'):
print_header()
print_totals()
elif args.get('files'):
print_entries(by='file')
print_totals()
else:
print_entries(by='name')
print_totals()
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Find struct sizes.")
parser.add_argument('obj_paths', nargs='*', default=OBJ_PATHS,
help="Description of where to find *.o files. May be a directory \
or a list of paths. Defaults to %r." % OBJ_PATHS)
parser.add_argument('-v', '--verbose', action='store_true',
description="Find struct sizes.",
allow_abbrev=False)
parser.add_argument(
'obj_paths',
nargs='*',
help="Input *.o files.")
parser.add_argument(
'-v', '--verbose',
action='store_true',
help="Output commands that run behind the scenes.")
parser.add_argument('-q', '--quiet', action='store_true',
parser.add_argument(
'-q', '--quiet',
action='store_true',
help="Don't show anything, useful with -o.")
parser.add_argument('-o', '--output',
parser.add_argument(
'-o', '--output',
help="Specify CSV file to store results.")
parser.add_argument('-u', '--use',
help="Don't compile and find struct sizes, instead use this CSV file.")
parser.add_argument('-d', '--diff',
help="Specify CSV file to diff struct size against.")
parser.add_argument('-m', '--merge',
help="Merge with an existing CSV file when writing to output.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all functions, not just the ones that changed.")
parser.add_argument('-A', '--everything', action='store_true',
parser.add_argument(
'-u', '--use',
help="Don't parse anything, use this CSV file.")
parser.add_argument(
'-d', '--diff',
help="Specify CSV file to diff against.")
parser.add_argument(
'-a', '--all',
action='store_true',
help="Show all, not just the ones that changed.")
parser.add_argument(
'-p', '--percent',
action='store_true',
help="Only show percentage change, not a full diff.")
parser.add_argument(
'-b', '--by',
action='append',
choices=StructResult._by,
help="Group by this field.")
parser.add_argument(
'-f', '--field',
dest='fields',
action='append',
choices=StructResult._fields,
help="Show this field.")
parser.add_argument(
'-D', '--define',
dest='defines',
action='append',
type=lambda x: (lambda k,v: (k, set(v.split(','))))(*x.split('=', 1)),
help="Only include results where this field is this value.")
class AppendSort(argparse.Action):
def __call__(self, parser, namespace, value, option):
if namespace.sort is None:
namespace.sort = []
namespace.sort.append((value, True if option == '-S' else False))
parser.add_argument(
'-s', '--sort',
nargs='?',
action=AppendSort,
help="Sort by this field.")
parser.add_argument(
'-S', '--reverse-sort',
nargs='?',
action=AppendSort,
help="Sort by this field, but backwards.")
parser.add_argument(
'-Y', '--summary',
action='store_true',
help="Only show the total.")
parser.add_argument(
'-F', '--source',
dest='sources',
action='append',
help="Only consider definitions in this file. Defaults to anything "
"in the current directory.")
parser.add_argument(
'--everything',
action='store_true',
help="Include builtin and libc specific symbols.")
parser.add_argument('-s', '--size-sort', action='store_true',
help="Sort by size.")
parser.add_argument('-S', '--reverse-size-sort', action='store_true',
help="Sort by size, but backwards.")
parser.add_argument('-F', '--files', action='store_true',
help="Show file-level struct sizes.")
parser.add_argument('-Y', '--summary', action='store_true',
help="Only show the total struct size.")
parser.add_argument('--objdump-tool', default=['objdump'], type=lambda x: x.split(),
help="Path to the objdump tool to use.")
parser.add_argument('--build-dir',
help="Specify the relative build directory. Used to map object files \
to the correct source files.")
sys.exit(main(**vars(parser.parse_args())))
parser.add_argument(
'--internal',
action='store_true',
help="Also show structs in .c files.")
parser.add_argument(
'--objdump-path',
type=lambda x: x.split(),
default=OBJDUMP_PATH,
help="Path to the objdump executable, may include flags. "
"Defaults to %r." % OBJDUMP_PATH)
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
+793 -243
View File
File diff suppressed because it is too large Load Diff
+177
View File
@@ -0,0 +1,177 @@
#!/usr/bin/env python3
#
# Efficiently displays the last n lines of a file/pipe.
#
# Example:
# ./scripts/tailpipe.py trace -n5
#
# Copyright (c) 2022, The littlefs authors.
# SPDX-License-Identifier: BSD-3-Clause
#
import collections as co
import io
import os
import select
import shutil
import sys
import threading as th
import time
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
class LinesIO:
def __init__(self, maxlen=None):
self.maxlen = maxlen
self.lines = co.deque(maxlen=maxlen)
self.tail = io.StringIO()
# trigger automatic sizing
if maxlen == 0:
self.resize(0)
def write(self, s):
# note using split here ensures the trailing string has no newline
lines = s.split('\n')
if len(lines) > 1 and self.tail.getvalue():
self.tail.write(lines[0])
lines[0] = self.tail.getvalue()
self.tail = io.StringIO()
self.lines.extend(lines[:-1])
if lines[-1]:
self.tail.write(lines[-1])
def resize(self, maxlen):
self.maxlen = maxlen
if maxlen == 0:
maxlen = shutil.get_terminal_size((80, 5))[1]
if maxlen != self.lines.maxlen:
self.lines = co.deque(self.lines, maxlen=maxlen)
canvas_lines = 1
def draw(self):
# did terminal size change?
if self.maxlen == 0:
self.resize(0)
# first thing first, give ourself a canvas
while LinesIO.canvas_lines < len(self.lines):
sys.stdout.write('\n')
LinesIO.canvas_lines += 1
# clear the bottom of the canvas if we shrink
shrink = LinesIO.canvas_lines - len(self.lines)
if shrink > 0:
for i in range(shrink):
sys.stdout.write('\r')
if shrink-1-i > 0:
sys.stdout.write('\x1b[%dA' % (shrink-1-i))
sys.stdout.write('\x1b[K')
if shrink-1-i > 0:
sys.stdout.write('\x1b[%dB' % (shrink-1-i))
sys.stdout.write('\x1b[%dA' % shrink)
LinesIO.canvas_lines = len(self.lines)
for i, line in enumerate(self.lines):
# move cursor, clear line, disable/reenable line wrapping
sys.stdout.write('\r')
if len(self.lines)-1-i > 0:
sys.stdout.write('\x1b[%dA' % (len(self.lines)-1-i))
sys.stdout.write('\x1b[K')
sys.stdout.write('\x1b[?7l')
sys.stdout.write(line)
sys.stdout.write('\x1b[?7h')
if len(self.lines)-1-i > 0:
sys.stdout.write('\x1b[%dB' % (len(self.lines)-1-i))
sys.stdout.flush()
def main(path='-', *, lines=5, cat=False, sleep=None, keep_open=False):
if cat:
ring = sys.stdout
else:
ring = LinesIO(lines)
# if sleep print in background thread to avoid getting stuck in a read call
event = th.Event()
lock = th.Lock()
if not cat:
done = False
def background():
while not done:
event.wait()
event.clear()
with lock:
ring.draw()
time.sleep(sleep or 0.01)
th.Thread(target=background, daemon=True).start()
try:
while True:
with openio(path) as f:
for line in f:
with lock:
ring.write(line)
event.set()
if not keep_open:
break
# don't just flood open calls
time.sleep(sleep or 0.1)
except FileNotFoundError as e:
print("error: file not found %r" % path)
sys.exit(-1)
except KeyboardInterrupt:
pass
if not cat:
done = True
lock.acquire() # avoids https://bugs.python.org/issue42717
sys.stdout.write('\n')
if __name__ == "__main__":
import sys
import argparse
parser = argparse.ArgumentParser(
description="Efficiently displays the last n lines of a file/pipe.",
allow_abbrev=False)
parser.add_argument(
'path',
nargs='?',
help="Path to read from.")
parser.add_argument(
'-n', '--lines',
nargs='?',
type=lambda x: int(x, 0),
const=0,
help="Show this many lines of history. 0 uses the terminal height. "
"Defaults to 5.")
parser.add_argument(
'-z', '--cat',
action='store_true',
help="Pipe directly to stdout.")
parser.add_argument(
'-s', '--sleep',
type=float,
help="Seconds to sleep between reads. Defaults to 0.01.")
parser.add_argument(
'-k', '--keep-open',
action='store_true',
help="Reopen the pipe on EOF, useful when multiple "
"processes are writing.")
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
+73
View File
@@ -0,0 +1,73 @@
#!/usr/bin/env python3
#
# tee, but for pipes
#
# Example:
# ./scripts/tee.py in_pipe out_pipe1 out_pipe2
#
# Copyright (c) 2022, The littlefs authors.
# SPDX-License-Identifier: BSD-3-Clause
#
import os
import io
import time
import sys
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def main(in_path, out_paths, *, keep_open=False):
out_pipes = [openio(p, 'wb', 0) for p in out_paths]
try:
with openio(in_path, 'rb', 0) as f:
while True:
buf = f.read(io.DEFAULT_BUFFER_SIZE)
if not buf:
if not keep_open:
break
# don't just flood reads
time.sleep(0.1)
continue
for p in out_pipes:
try:
p.write(buf)
except BrokenPipeError:
pass
except FileNotFoundError as e:
print("error: file not found %r" % in_path)
sys.exit(-1)
except KeyboardInterrupt:
pass
if __name__ == "__main__":
import sys
import argparse
parser = argparse.ArgumentParser(
description="tee, but for pipes.",
allow_abbrev=False)
parser.add_argument(
'in_path',
help="Path to read from.")
parser.add_argument(
'out_paths',
nargs='+',
help="Path to write to.")
parser.add_argument(
'-k', '--keep-open',
action='store_true',
help="Reopen the pipe on EOF, useful when multiple "
"processes are writing.")
sys.exit(main(**{k: v
for k, v in vars(parser.parse_intermixed_args()).items()
if v is not None}))
+1390 -766
View File
File diff suppressed because it is too large Load Diff
+1002
View File
File diff suppressed because it is too large Load Diff
+265
View File
@@ -0,0 +1,265 @@
#!/usr/bin/env python3
#
# Traditional watch command, but with higher resolution updates and a bit
# different options/output format
#
# Example:
# ./scripts/watch.py -s0.1 date
#
# Copyright (c) 2022, The littlefs authors.
# SPDX-License-Identifier: BSD-3-Clause
#
import collections as co
import errno
import fcntl
import io
import os
import pty
import re
import shutil
import struct
import subprocess as sp
import sys
import termios
import time
try:
import inotify_simple
except ModuleNotFoundError:
inotify_simple = None
def openio(path, mode='r', buffering=-1):
# allow '-' for stdin/stdout
if path == '-':
if mode == 'r':
return os.fdopen(os.dup(sys.stdin.fileno()), mode, buffering)
else:
return os.fdopen(os.dup(sys.stdout.fileno()), mode, buffering)
else:
return open(path, mode, buffering)
def inotifywait(paths):
# wait for interesting events
inotify = inotify_simple.INotify()
flags = (inotify_simple.flags.ATTRIB
| inotify_simple.flags.CREATE
| inotify_simple.flags.DELETE
| inotify_simple.flags.DELETE_SELF
| inotify_simple.flags.MODIFY
| inotify_simple.flags.MOVED_FROM
| inotify_simple.flags.MOVED_TO
| inotify_simple.flags.MOVE_SELF)
# recurse into directories
for path in paths:
if os.path.isdir(path):
for dir, _, files in os.walk(path):
inotify.add_watch(dir, flags)
for f in files:
inotify.add_watch(os.path.join(dir, f), flags)
else:
inotify.add_watch(path, flags)
# wait for event
inotify.read()
class LinesIO:
def __init__(self, maxlen=None):
self.maxlen = maxlen
self.lines = co.deque(maxlen=maxlen)
self.tail = io.StringIO()
# trigger automatic sizing
if maxlen == 0:
self.resize(0)
def write(self, s):
# note using split here ensures the trailing string has no newline
lines = s.split('\n')
if len(lines) > 1 and self.tail.getvalue():
self.tail.write(lines[0])
lines[0] = self.tail.getvalue()
self.tail = io.StringIO()
self.lines.extend(lines[:-1])
if lines[-1]:
self.tail.write(lines[-1])
def resize(self, maxlen):
self.maxlen = maxlen
if maxlen == 0:
maxlen = shutil.get_terminal_size((80, 5))[1]
if maxlen != self.lines.maxlen:
self.lines = co.deque(self.lines, maxlen=maxlen)
canvas_lines = 1
def draw(self):
# did terminal size change?
if self.maxlen == 0:
self.resize(0)
# first thing first, give ourself a canvas
while LinesIO.canvas_lines < len(self.lines):
sys.stdout.write('\n')
LinesIO.canvas_lines += 1
# clear the bottom of the canvas if we shrink
shrink = LinesIO.canvas_lines - len(self.lines)
if shrink > 0:
for i in range(shrink):
sys.stdout.write('\r')
if shrink-1-i > 0:
sys.stdout.write('\x1b[%dA' % (shrink-1-i))
sys.stdout.write('\x1b[K')
if shrink-1-i > 0:
sys.stdout.write('\x1b[%dB' % (shrink-1-i))
sys.stdout.write('\x1b[%dA' % shrink)
LinesIO.canvas_lines = len(self.lines)
for i, line in enumerate(self.lines):
# move cursor, clear line, disable/reenable line wrapping
sys.stdout.write('\r')
if len(self.lines)-1-i > 0:
sys.stdout.write('\x1b[%dA' % (len(self.lines)-1-i))
sys.stdout.write('\x1b[K')
sys.stdout.write('\x1b[?7l')
sys.stdout.write(line)
sys.stdout.write('\x1b[?7h')
if len(self.lines)-1-i > 0:
sys.stdout.write('\x1b[%dB' % (len(self.lines)-1-i))
sys.stdout.flush()
def main(command, *,
lines=0,
cat=False,
sleep=None,
keep_open=False,
keep_open_paths=None,
exit_on_error=False):
returncode = 0
try:
while True:
# reset ring each run
if cat:
ring = sys.stdout
else:
ring = LinesIO(lines)
try:
# run the command under a pseudoterminal
mpty, spty = pty.openpty()
# forward terminal size
w, h = shutil.get_terminal_size((80, 5))
if lines:
h = lines
fcntl.ioctl(spty, termios.TIOCSWINSZ,
struct.pack('HHHH', h, w, 0, 0))
proc = sp.Popen(command,
stdout=spty,
stderr=spty,
close_fds=False)
os.close(spty)
mpty = os.fdopen(mpty, 'r', 1)
while True:
try:
line = mpty.readline()
except OSError as e:
if e.errno != errno.EIO:
raise
break
if not line:
break
ring.write(line)
if not cat:
ring.draw()
mpty.close()
proc.wait()
if exit_on_error and proc.returncode != 0:
returncode = proc.returncode
break
except OSError as e:
if e.errno != errno.ETXTBSY:
raise
pass
# try to inotifywait
if keep_open and inotify_simple is not None:
if keep_open_paths:
paths = set(keep_paths)
else:
# guess inotify paths from command
paths = set()
for p in command:
for p in {
p,
re.sub('^-.', '', p),
re.sub('^--[^=]+=', '', p)}:
if p and os.path.exists(p):
paths.add(p)
ptime = time.time()
inotifywait(paths)
# sleep for a minimum amount of time, this helps issues around
# rapidly updating files
time.sleep(max(0, (sleep or 0.1) - (time.time()-ptime)))
else:
time.sleep(sleep or 0.1)
except KeyboardInterrupt:
pass
if not cat:
sys.stdout.write('\n')
sys.exit(returncode)
if __name__ == "__main__":
import sys
import argparse
parser = argparse.ArgumentParser(
description="Traditional watch command, but with higher resolution "
"updates and a bit different options/output format.",
allow_abbrev=False)
parser.add_argument(
'command',
nargs=argparse.REMAINDER,
help="Command to run.")
parser.add_argument(
'-n', '--lines',
nargs='?',
type=lambda x: int(x, 0),
const=0,
help="Show this many lines of history. 0 uses the terminal height. "
"Defaults to 0.")
parser.add_argument(
'-z', '--cat',
action='store_true',
help="Pipe directly to stdout.")
parser.add_argument(
'-s', '--sleep',
type=float,
help="Seconds to sleep between runs. Defaults to 0.1.")
parser.add_argument(
'-k', '--keep-open',
action='store_true',
help="Try to use inotify to wait for changes.")
parser.add_argument(
'-K', '--keep-open-path',
dest='keep_open_paths',
action='append',
help="Use this path for inotify. Defaults to guessing.")
parser.add_argument(
'-e', '--exit-on-error',
action='store_true',
help="Exit on error.")
sys.exit(main(**{k: v
for k, v in vars(parser.parse_args()).items()
if v is not None}))