• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1#!/usr/bin/env python3
2#
3# Parse and report coverage info from .info files generated by lcov
4#
5import os
6import glob
7import csv
8import re
9import collections as co
10import bisect as b
11
12
13INFO_PATHS = ['tests/*.toml.info']
14
15def collect(paths, **args):
16    file = None
17    funcs = []
18    lines = co.defaultdict(lambda: 0)
19    pattern = re.compile(
20        '^(?P<file>SF:/?(?P<file_name>.*))$'
21        '|^(?P<func>FN:(?P<func_lineno>[0-9]*),(?P<func_name>.*))$'
22        '|^(?P<line>DA:(?P<line_lineno>[0-9]*),(?P<line_hits>[0-9]*))$')
23    for path in paths:
24        with open(path) as f:
25            for line in f:
26                m = pattern.match(line)
27                if m and m.group('file'):
28                    file = m.group('file_name')
29                elif m and file and m.group('func'):
30                    funcs.append((file, int(m.group('func_lineno')),
31                        m.group('func_name')))
32                elif m and file and m.group('line'):
33                    lines[(file, int(m.group('line_lineno')))] += (
34                        int(m.group('line_hits')))
35
36    # map line numbers to functions
37    funcs.sort()
38    def func_from_lineno(file, lineno):
39        i = b.bisect(funcs, (file, lineno))
40        if i and funcs[i-1][0] == file:
41            return funcs[i-1][2]
42        else:
43            return None
44
45    # reduce to function info
46    reduced_funcs = co.defaultdict(lambda: (0, 0))
47    for (file, line_lineno), line_hits in lines.items():
48        func = func_from_lineno(file, line_lineno)
49        if not func:
50            continue
51        hits, count = reduced_funcs[(file, func)]
52        reduced_funcs[(file, func)] = (hits + (line_hits > 0), count + 1)
53
54    results = []
55    for (file, func), (hits, count) in reduced_funcs.items():
56        # discard internal/testing functions (test_* injected with
57        # internal testing)
58        if not args.get('everything'):
59            if func.startswith('__') or func.startswith('test_'):
60                continue
61        # discard .8449 suffixes created by optimizer
62        func = re.sub('\.[0-9]+', '', func)
63        results.append((file, func, hits, count))
64
65    return results
66
67
68def main(**args):
69    def openio(path, mode='r'):
70        if path == '-':
71            if 'r' in mode:
72                return os.fdopen(os.dup(sys.stdin.fileno()), 'r')
73            else:
74                return os.fdopen(os.dup(sys.stdout.fileno()), 'w')
75        else:
76            return open(path, mode)
77
78    # find coverage
79    if not args.get('use'):
80        # find *.info files
81        paths = []
82        for path in args['info_paths']:
83            if os.path.isdir(path):
84                path = path + '/*.gcov'
85
86            for path in glob.glob(path):
87                paths.append(path)
88
89        if not paths:
90            print('no .info files found in %r?' % args['info_paths'])
91            sys.exit(-1)
92
93        results = collect(paths, **args)
94    else:
95        with openio(args['use']) as f:
96            r = csv.DictReader(f)
97            results = [
98                (   result['file'],
99                    result['name'],
100                    int(result['coverage_hits']),
101                    int(result['coverage_count']))
102                for result in r
103                if result.get('coverage_hits') not in {None, ''}
104                if result.get('coverage_count') not in {None, ''}]
105
106    total_hits, total_count = 0, 0
107    for _, _, hits, count in results:
108        total_hits += hits
109        total_count += count
110
111    # find previous results?
112    if args.get('diff'):
113        try:
114            with openio(args['diff']) as f:
115                r = csv.DictReader(f)
116                prev_results = [
117                    (   result['file'],
118                        result['name'],
119                        int(result['coverage_hits']),
120                        int(result['coverage_count']))
121                    for result in r
122                    if result.get('coverage_hits') not in {None, ''}
123                    if result.get('coverage_count') not in {None, ''}]
124        except FileNotFoundError:
125            prev_results = []
126
127        prev_total_hits, prev_total_count = 0, 0
128        for _, _, hits, count in prev_results:
129            prev_total_hits += hits
130            prev_total_count += count
131
132    # write results to CSV
133    if args.get('output'):
134        merged_results = co.defaultdict(lambda: {})
135        other_fields = []
136
137        # merge?
138        if args.get('merge'):
139            try:
140                with openio(args['merge']) as f:
141                    r = csv.DictReader(f)
142                    for result in r:
143                        file = result.pop('file', '')
144                        func = result.pop('name', '')
145                        result.pop('coverage_hits', None)
146                        result.pop('coverage_count', None)
147                        merged_results[(file, func)] = result
148                        other_fields = result.keys()
149            except FileNotFoundError:
150                pass
151
152        for file, func, hits, count in results:
153            merged_results[(file, func)]['coverage_hits'] = hits
154            merged_results[(file, func)]['coverage_count'] = count
155
156        with openio(args['output'], 'w') as f:
157            w = csv.DictWriter(f, ['file', 'name', *other_fields, 'coverage_hits', 'coverage_count'])
158            w.writeheader()
159            for (file, func), result in sorted(merged_results.items()):
160                w.writerow({'file': file, 'name': func, **result})
161
162    # print results
163    def dedup_entries(results, by='name'):
164        entries = co.defaultdict(lambda: (0, 0))
165        for file, func, hits, count in results:
166            entry = (file if by == 'file' else func)
167            entry_hits, entry_count = entries[entry]
168            entries[entry] = (entry_hits + hits, entry_count + count)
169        return entries
170
171    def diff_entries(olds, news):
172        diff = co.defaultdict(lambda: (0, 0, 0, 0, 0, 0, 0))
173        for name, (new_hits, new_count) in news.items():
174            diff[name] = (
175                0, 0,
176                new_hits, new_count,
177                new_hits, new_count,
178                (new_hits/new_count if new_count else 1.0) - 1.0)
179        for name, (old_hits, old_count) in olds.items():
180            _, _, new_hits, new_count, _, _, _ = diff[name]
181            diff[name] = (
182                old_hits, old_count,
183                new_hits, new_count,
184                new_hits-old_hits, new_count-old_count,
185                ((new_hits/new_count if new_count else 1.0)
186                    - (old_hits/old_count if old_count else 1.0)))
187        return diff
188
189    def sorted_entries(entries):
190        if args.get('coverage_sort'):
191            return sorted(entries, key=lambda x: (-(x[1][0]/x[1][1] if x[1][1] else -1), x))
192        elif args.get('reverse_coverage_sort'):
193            return sorted(entries, key=lambda x: (+(x[1][0]/x[1][1] if x[1][1] else -1), x))
194        else:
195            return sorted(entries)
196
197    def sorted_diff_entries(entries):
198        if args.get('coverage_sort'):
199            return sorted(entries, key=lambda x: (-(x[1][2]/x[1][3] if x[1][3] else -1), x))
200        elif args.get('reverse_coverage_sort'):
201            return sorted(entries, key=lambda x: (+(x[1][2]/x[1][3] if x[1][3] else -1), x))
202        else:
203            return sorted(entries, key=lambda x: (-x[1][6], x))
204
205    def print_header(by=''):
206        if not args.get('diff'):
207            print('%-36s %19s' % (by, 'hits/line'))
208        else:
209            print('%-36s %19s %19s %11s' % (by, 'old', 'new', 'diff'))
210
211    def print_entry(name, hits, count):
212        print("%-36s %11s %7s" % (name,
213            '%d/%d' % (hits, count)
214                if count else '-',
215            '%.1f%%' % (100*hits/count)
216                if count else '-'))
217
218    def print_diff_entry(name,
219            old_hits, old_count,
220            new_hits, new_count,
221            diff_hits, diff_count,
222            ratio):
223        print("%-36s %11s %7s %11s %7s %11s%s" % (name,
224            '%d/%d' % (old_hits, old_count)
225                if old_count else '-',
226            '%.1f%%' % (100*old_hits/old_count)
227                if old_count else '-',
228            '%d/%d' % (new_hits, new_count)
229                if new_count else '-',
230            '%.1f%%' % (100*new_hits/new_count)
231                if new_count else '-',
232            '%+d/%+d' % (diff_hits, diff_count),
233            ' (%+.1f%%)' % (100*ratio) if ratio else ''))
234
235    def print_entries(by='name'):
236        entries = dedup_entries(results, by=by)
237
238        if not args.get('diff'):
239            print_header(by=by)
240            for name, (hits, count) in sorted_entries(entries.items()):
241                print_entry(name, hits, count)
242        else:
243            prev_entries = dedup_entries(prev_results, by=by)
244            diff = diff_entries(prev_entries, entries)
245            print_header(by='%s (%d added, %d removed)' % (by,
246                sum(1 for _, old, _, _, _, _, _ in diff.values() if not old),
247                sum(1 for _, _, _, new, _, _, _ in diff.values() if not new)))
248            for name, (
249                    old_hits, old_count,
250                    new_hits, new_count,
251                    diff_hits, diff_count, ratio) in sorted_diff_entries(
252                        diff.items()):
253                if ratio or args.get('all'):
254                    print_diff_entry(name,
255                        old_hits, old_count,
256                        new_hits, new_count,
257                        diff_hits, diff_count,
258                        ratio)
259
260    def print_totals():
261        if not args.get('diff'):
262            print_entry('TOTAL', total_hits, total_count)
263        else:
264            ratio = ((total_hits/total_count
265                    if total_count else 1.0)
266                - (prev_total_hits/prev_total_count
267                    if prev_total_count else 1.0))
268            print_diff_entry('TOTAL',
269                prev_total_hits, prev_total_count,
270                total_hits, total_count,
271                total_hits-prev_total_hits, total_count-prev_total_count,
272                ratio)
273
274    if args.get('quiet'):
275        pass
276    elif args.get('summary'):
277        print_header()
278        print_totals()
279    elif args.get('files'):
280        print_entries(by='file')
281        print_totals()
282    else:
283        print_entries(by='name')
284        print_totals()
285
286if __name__ == "__main__":
287    import argparse
288    import sys
289    parser = argparse.ArgumentParser(
290        description="Parse and report coverage info from .info files \
291            generated by lcov")
292    parser.add_argument('info_paths', nargs='*', default=INFO_PATHS,
293        help="Description of where to find *.info files. May be a directory \
294            or list of paths. *.info files will be merged to show the total \
295            coverage. Defaults to %r." % INFO_PATHS)
296    parser.add_argument('-v', '--verbose', action='store_true',
297        help="Output commands that run behind the scenes.")
298    parser.add_argument('-o', '--output',
299        help="Specify CSV file to store results.")
300    parser.add_argument('-u', '--use',
301        help="Don't do any work, instead use this CSV file.")
302    parser.add_argument('-d', '--diff',
303        help="Specify CSV file to diff code size against.")
304    parser.add_argument('-m', '--merge',
305        help="Merge with an existing CSV file when writing to output.")
306    parser.add_argument('-a', '--all', action='store_true',
307        help="Show all functions, not just the ones that changed.")
308    parser.add_argument('-A', '--everything', action='store_true',
309        help="Include builtin and libc specific symbols.")
310    parser.add_argument('-s', '--coverage-sort', action='store_true',
311        help="Sort by coverage.")
312    parser.add_argument('-S', '--reverse-coverage-sort', action='store_true',
313        help="Sort by coverage, but backwards.")
314    parser.add_argument('-F', '--files', action='store_true',
315        help="Show file-level coverage.")
316    parser.add_argument('-Y', '--summary', action='store_true',
317        help="Only show the total coverage.")
318    parser.add_argument('-q', '--quiet', action='store_true',
319        help="Don't show anything, useful with -o.")
320    parser.add_argument('--build-dir',
321        help="Specify the relative build directory. Used to map object files \
322            to the correct source files.")
323    sys.exit(main(**vars(parser.parse_args())))
324