xref: /openbmc/openbmc/poky/bitbake/lib/bb/siggen.py (revision 1a4b7ee2)
1import hashlib
2import logging
3import os
4import re
5import tempfile
6import pickle
7import bb.data
8import difflib
9import simplediff
10from bb.checksum import FileChecksumCache
11
12logger = logging.getLogger('BitBake.SigGen')
13
14def init(d):
15    siggens = [obj for obj in globals().values()
16                      if type(obj) is type and issubclass(obj, SignatureGenerator)]
17
18    desired = d.getVar("BB_SIGNATURE_HANDLER") or "noop"
19    for sg in siggens:
20        if desired == sg.name:
21            return sg(d)
22            break
23    else:
24        logger.error("Invalid signature generator '%s', using default 'noop'\n"
25                     "Available generators: %s", desired,
26                     ', '.join(obj.name for obj in siggens))
27        return SignatureGenerator(d)
28
29class SignatureGenerator(object):
30    """
31    """
32    name = "noop"
33
34    def __init__(self, data):
35        self.basehash = {}
36        self.taskhash = {}
37        self.runtaskdeps = {}
38        self.file_checksum_values = {}
39        self.taints = {}
40
41    def finalise(self, fn, d, varient):
42        return
43
44    def get_taskhash(self, fn, task, deps, dataCache):
45        return "0"
46
47    def writeout_file_checksum_cache(self):
48        """Write/update the file checksum cache onto disk"""
49        return
50
51    def stampfile(self, stampbase, file_name, taskname, extrainfo):
52        return ("%s.%s.%s" % (stampbase, taskname, extrainfo)).rstrip('.')
53
54    def stampcleanmask(self, stampbase, file_name, taskname, extrainfo):
55        return ("%s.%s.%s" % (stampbase, taskname, extrainfo)).rstrip('.')
56
57    def dump_sigtask(self, fn, task, stampbase, runtime):
58        return
59
60    def invalidate_task(self, task, d, fn):
61        bb.build.del_stamp(task, d, fn)
62
63    def dump_sigs(self, dataCache, options):
64        return
65
66    def get_taskdata(self):
67        return (self.runtaskdeps, self.taskhash, self.file_checksum_values, self.taints, self.basehash)
68
69    def set_taskdata(self, data):
70        self.runtaskdeps, self.taskhash, self.file_checksum_values, self.taints, self.basehash = data
71
72    def reset(self, data):
73        self.__init__(data)
74
75
76class SignatureGeneratorBasic(SignatureGenerator):
77    """
78    """
79    name = "basic"
80
81    def __init__(self, data):
82        self.basehash = {}
83        self.taskhash = {}
84        self.taskdeps = {}
85        self.runtaskdeps = {}
86        self.file_checksum_values = {}
87        self.taints = {}
88        self.gendeps = {}
89        self.lookupcache = {}
90        self.pkgnameextract = re.compile("(?P<fn>.*)\..*")
91        self.basewhitelist = set((data.getVar("BB_HASHBASE_WHITELIST") or "").split())
92        self.taskwhitelist = None
93        self.init_rundepcheck(data)
94        checksum_cache_file = data.getVar("BB_HASH_CHECKSUM_CACHE_FILE")
95        if checksum_cache_file:
96            self.checksum_cache = FileChecksumCache()
97            self.checksum_cache.init_cache(data, checksum_cache_file)
98        else:
99            self.checksum_cache = None
100
101    def init_rundepcheck(self, data):
102        self.taskwhitelist = data.getVar("BB_HASHTASK_WHITELIST") or None
103        if self.taskwhitelist:
104            self.twl = re.compile(self.taskwhitelist)
105        else:
106            self.twl = None
107
108    def _build_data(self, fn, d):
109
110        ignore_mismatch = ((d.getVar("BB_HASH_IGNORE_MISMATCH") or '') == '1')
111        tasklist, gendeps, lookupcache = bb.data.generate_dependencies(d)
112
113        taskdeps, basehash = bb.data.generate_dependency_hash(tasklist, gendeps, lookupcache, self.basewhitelist, fn)
114
115        for task in tasklist:
116            k = fn + "." + task
117            if not ignore_mismatch and k in self.basehash and self.basehash[k] != basehash[k]:
118                bb.error("When reparsing %s, the basehash value changed from %s to %s. The metadata is not deterministic and this needs to be fixed." % (k, self.basehash[k], basehash[k]))
119            self.basehash[k] = basehash[k]
120
121        self.taskdeps[fn] = taskdeps
122        self.gendeps[fn] = gendeps
123        self.lookupcache[fn] = lookupcache
124
125        return taskdeps
126
127    def finalise(self, fn, d, variant):
128
129        mc = d.getVar("__BBMULTICONFIG", False) or ""
130        if variant or mc:
131            fn = bb.cache.realfn2virtual(fn, variant, mc)
132
133        try:
134            taskdeps = self._build_data(fn, d)
135        except bb.parse.SkipRecipe:
136            raise
137        except:
138            bb.warn("Error during finalise of %s" % fn)
139            raise
140
141        #Slow but can be useful for debugging mismatched basehashes
142        #for task in self.taskdeps[fn]:
143        #    self.dump_sigtask(fn, task, d.getVar("STAMP"), False)
144
145        for task in taskdeps:
146            d.setVar("BB_BASEHASH_task-%s" % task, self.basehash[fn + "." + task])
147
148    def rundep_check(self, fn, recipename, task, dep, depname, dataCache):
149        # Return True if we should keep the dependency, False to drop it
150        # We only manipulate the dependencies for packages not in the whitelist
151        if self.twl and not self.twl.search(recipename):
152            # then process the actual dependencies
153            if self.twl.search(depname):
154                return False
155        return True
156
157    def read_taint(self, fn, task, stampbase):
158        taint = None
159        try:
160            with open(stampbase + '.' + task + '.taint', 'r') as taintf:
161                taint = taintf.read()
162        except IOError:
163            pass
164        return taint
165
166    def get_taskhash(self, fn, task, deps, dataCache):
167
168        mc = ''
169        if fn.startswith('multiconfig:'):
170            mc = fn.split(':')[1]
171        k = fn + "." + task
172
173        data = dataCache.basetaskhash[k]
174        self.basehash[k] = data
175        self.runtaskdeps[k] = []
176        self.file_checksum_values[k] = []
177        recipename = dataCache.pkg_fn[fn]
178        for dep in sorted(deps, key=clean_basepath):
179            pkgname = self.pkgnameextract.search(dep).group('fn')
180            if mc:
181                depmc = pkgname.split(':')[1]
182                if mc != depmc:
183                    continue
184            depname = dataCache.pkg_fn[pkgname]
185            if not self.rundep_check(fn, recipename, task, dep, depname, dataCache):
186                continue
187            if dep not in self.taskhash:
188                bb.fatal("%s is not in taskhash, caller isn't calling in dependency order?" % dep)
189            data = data + self.taskhash[dep]
190            self.runtaskdeps[k].append(dep)
191
192        if task in dataCache.file_checksums[fn]:
193            if self.checksum_cache:
194                checksums = self.checksum_cache.get_checksums(dataCache.file_checksums[fn][task], recipename)
195            else:
196                checksums = bb.fetch2.get_file_checksums(dataCache.file_checksums[fn][task], recipename)
197            for (f,cs) in checksums:
198                self.file_checksum_values[k].append((f,cs))
199                if cs:
200                    data = data + cs
201
202        taskdep = dataCache.task_deps[fn]
203        if 'nostamp' in taskdep and task in taskdep['nostamp']:
204            # Nostamp tasks need an implicit taint so that they force any dependent tasks to run
205            import uuid
206            taint = str(uuid.uuid4())
207            data = data + taint
208            self.taints[k] = "nostamp:" + taint
209
210        taint = self.read_taint(fn, task, dataCache.stamp[fn])
211        if taint:
212            data = data + taint
213            self.taints[k] = taint
214            logger.warning("%s is tainted from a forced run" % k)
215
216        h = hashlib.md5(data.encode("utf-8")).hexdigest()
217        self.taskhash[k] = h
218        #d.setVar("BB_TASKHASH_task-%s" % task, taskhash[task])
219        return h
220
221    def writeout_file_checksum_cache(self):
222        """Write/update the file checksum cache onto disk"""
223        if self.checksum_cache:
224            self.checksum_cache.save_extras()
225            self.checksum_cache.save_merge()
226        else:
227            bb.fetch2.fetcher_parse_save()
228            bb.fetch2.fetcher_parse_done()
229
230    def dump_sigtask(self, fn, task, stampbase, runtime):
231
232        k = fn + "." + task
233        referencestamp = stampbase
234        if isinstance(runtime, str) and runtime.startswith("customfile"):
235            sigfile = stampbase
236            referencestamp = runtime[11:]
237        elif runtime and k in self.taskhash:
238            sigfile = stampbase + "." + task + ".sigdata" + "." + self.taskhash[k]
239        else:
240            sigfile = stampbase + "." + task + ".sigbasedata" + "." + self.basehash[k]
241
242        bb.utils.mkdirhier(os.path.dirname(sigfile))
243
244        data = {}
245        data['task'] = task
246        data['basewhitelist'] = self.basewhitelist
247        data['taskwhitelist'] = self.taskwhitelist
248        data['taskdeps'] = self.taskdeps[fn][task]
249        data['basehash'] = self.basehash[k]
250        data['gendeps'] = {}
251        data['varvals'] = {}
252        data['varvals'][task] = self.lookupcache[fn][task]
253        for dep in self.taskdeps[fn][task]:
254            if dep in self.basewhitelist:
255                continue
256            data['gendeps'][dep] = self.gendeps[fn][dep]
257            data['varvals'][dep] = self.lookupcache[fn][dep]
258
259        if runtime and k in self.taskhash:
260            data['runtaskdeps'] = self.runtaskdeps[k]
261            data['file_checksum_values'] = [(os.path.basename(f), cs) for f,cs in self.file_checksum_values[k]]
262            data['runtaskhashes'] = {}
263            for dep in data['runtaskdeps']:
264                data['runtaskhashes'][dep] = self.taskhash[dep]
265            data['taskhash'] = self.taskhash[k]
266
267        taint = self.read_taint(fn, task, referencestamp)
268        if taint:
269            data['taint'] = taint
270
271        if runtime and k in self.taints:
272            if 'nostamp:' in self.taints[k]:
273                data['taint'] = self.taints[k]
274
275        computed_basehash = calc_basehash(data)
276        if computed_basehash != self.basehash[k]:
277            bb.error("Basehash mismatch %s versus %s for %s" % (computed_basehash, self.basehash[k], k))
278        if runtime and k in self.taskhash:
279            computed_taskhash = calc_taskhash(data)
280            if computed_taskhash != self.taskhash[k]:
281                bb.error("Taskhash mismatch %s versus %s for %s" % (computed_taskhash, self.taskhash[k], k))
282                sigfile = sigfile.replace(self.taskhash[k], computed_taskhash)
283
284        fd, tmpfile = tempfile.mkstemp(dir=os.path.dirname(sigfile), prefix="sigtask.")
285        try:
286            with os.fdopen(fd, "wb") as stream:
287                p = pickle.dump(data, stream, -1)
288                stream.flush()
289            os.chmod(tmpfile, 0o664)
290            os.rename(tmpfile, sigfile)
291        except (OSError, IOError) as err:
292            try:
293                os.unlink(tmpfile)
294            except OSError:
295                pass
296            raise err
297
298    def dump_sigfn(self, fn, dataCaches, options):
299        if fn in self.taskdeps:
300            for task in self.taskdeps[fn]:
301                tid = fn + ":" + task
302                (mc, _, _) = bb.runqueue.split_tid(tid)
303                k = fn + "." + task
304                if k not in self.taskhash:
305                    continue
306                if dataCaches[mc].basetaskhash[k] != self.basehash[k]:
307                    bb.error("Bitbake's cached basehash does not match the one we just generated (%s)!" % k)
308                    bb.error("The mismatched hashes were %s and %s" % (dataCaches[mc].basetaskhash[k], self.basehash[k]))
309                self.dump_sigtask(fn, task, dataCaches[mc].stamp[fn], True)
310
311class SignatureGeneratorBasicHash(SignatureGeneratorBasic):
312    name = "basichash"
313
314    def stampfile(self, stampbase, fn, taskname, extrainfo, clean=False):
315        if taskname != "do_setscene" and taskname.endswith("_setscene"):
316            k = fn + "." + taskname[:-9]
317        else:
318            k = fn + "." + taskname
319        if clean:
320            h = "*"
321        elif k in self.taskhash:
322            h = self.taskhash[k]
323        else:
324            # If k is not in basehash, then error
325            h = self.basehash[k]
326        return ("%s.%s.%s.%s" % (stampbase, taskname, h, extrainfo)).rstrip('.')
327
328    def stampcleanmask(self, stampbase, fn, taskname, extrainfo):
329        return self.stampfile(stampbase, fn, taskname, extrainfo, clean=True)
330
331    def invalidate_task(self, task, d, fn):
332        bb.note("Tainting hash to force rebuild of task %s, %s" % (fn, task))
333        bb.build.write_taint(task, d, fn)
334
335def dump_this_task(outfile, d):
336    import bb.parse
337    fn = d.getVar("BB_FILENAME")
338    task = "do_" + d.getVar("BB_CURRENTTASK")
339    referencestamp = bb.build.stamp_internal(task, d, None, True)
340    bb.parse.siggen.dump_sigtask(fn, task, outfile, "customfile:" + referencestamp)
341
342def init_colors(enable_color):
343    """Initialise colour dict for passing to compare_sigfiles()"""
344    # First set up the colours
345    colors = {'color_title':   '\033[1m',
346              'color_default': '\033[0m',
347              'color_add':     '\033[0;32m',
348              'color_remove':  '\033[0;31m',
349             }
350    # Leave all keys present but clear the values
351    if not enable_color:
352        for k in colors.keys():
353            colors[k] = ''
354    return colors
355
356def worddiff_str(oldstr, newstr, colors=None):
357    if not colors:
358        colors = init_colors(False)
359    diff = simplediff.diff(oldstr.split(' '), newstr.split(' '))
360    ret = []
361    for change, value in diff:
362        value = ' '.join(value)
363        if change == '=':
364            ret.append(value)
365        elif change == '+':
366            item = '{color_add}{{+{value}+}}{color_default}'.format(value=value, **colors)
367            ret.append(item)
368        elif change == '-':
369            item = '{color_remove}[-{value}-]{color_default}'.format(value=value, **colors)
370            ret.append(item)
371    whitespace_note = ''
372    if oldstr != newstr and ' '.join(oldstr.split()) == ' '.join(newstr.split()):
373        whitespace_note = ' (whitespace changed)'
374    return '"%s"%s' % (' '.join(ret), whitespace_note)
375
376def list_inline_diff(oldlist, newlist, colors=None):
377    if not colors:
378        colors = init_colors(False)
379    diff = simplediff.diff(oldlist, newlist)
380    ret = []
381    for change, value in diff:
382        value = ' '.join(value)
383        if change == '=':
384            ret.append("'%s'" % value)
385        elif change == '+':
386            item = '{color_add}+{value}{color_default}'.format(value=value, **colors)
387            ret.append(item)
388        elif change == '-':
389            item = '{color_remove}-{value}{color_default}'.format(value=value, **colors)
390            ret.append(item)
391    return '[%s]' % (', '.join(ret))
392
393def clean_basepath(a):
394    mc = None
395    if a.startswith("multiconfig:"):
396        _, mc, a = a.split(":", 2)
397    b = a.rsplit("/", 2)[1] + '/' + a.rsplit("/", 2)[2]
398    if a.startswith("virtual:"):
399        b = b + ":" + a.rsplit(":", 1)[0]
400    if mc:
401        b = b + ":multiconfig:" + mc
402    return b
403
404def clean_basepaths(a):
405    b = {}
406    for x in a:
407        b[clean_basepath(x)] = a[x]
408    return b
409
410def clean_basepaths_list(a):
411    b = []
412    for x in a:
413        b.append(clean_basepath(x))
414    return b
415
416def compare_sigfiles(a, b, recursecb=None, color=False, collapsed=False):
417    output = []
418
419    colors = init_colors(color)
420    def color_format(formatstr, **values):
421        """
422        Return colour formatted string.
423        NOTE: call with the format string, not an already formatted string
424        containing values (otherwise you could have trouble with { and }
425        characters)
426        """
427        if not formatstr.endswith('{color_default}'):
428            formatstr += '{color_default}'
429        # In newer python 3 versions you can pass both of these directly,
430        # but we only require 3.4 at the moment
431        formatparams = {}
432        formatparams.update(colors)
433        formatparams.update(values)
434        return formatstr.format(**formatparams)
435
436    with open(a, 'rb') as f:
437        p1 = pickle.Unpickler(f)
438        a_data = p1.load()
439    with open(b, 'rb') as f:
440        p2 = pickle.Unpickler(f)
441        b_data = p2.load()
442
443    def dict_diff(a, b, whitelist=set()):
444        sa = set(a.keys())
445        sb = set(b.keys())
446        common = sa & sb
447        changed = set()
448        for i in common:
449            if a[i] != b[i] and i not in whitelist:
450                changed.add(i)
451        added = sb - sa
452        removed = sa - sb
453        return changed, added, removed
454
455    def file_checksums_diff(a, b):
456        from collections import Counter
457        # Handle old siginfo format
458        if isinstance(a, dict):
459            a = [(os.path.basename(f), cs) for f, cs in a.items()]
460        if isinstance(b, dict):
461            b = [(os.path.basename(f), cs) for f, cs in b.items()]
462        # Compare lists, ensuring we can handle duplicate filenames if they exist
463        removedcount = Counter(a)
464        removedcount.subtract(b)
465        addedcount = Counter(b)
466        addedcount.subtract(a)
467        added = []
468        for x in b:
469            if addedcount[x] > 0:
470                addedcount[x] -= 1
471                added.append(x)
472        removed = []
473        changed = []
474        for x in a:
475            if removedcount[x] > 0:
476                removedcount[x] -= 1
477                for y in added:
478                    if y[0] == x[0]:
479                        changed.append((x[0], x[1], y[1]))
480                        added.remove(y)
481                        break
482                else:
483                    removed.append(x)
484        added = [x[0] for x in added]
485        removed = [x[0] for x in removed]
486        return changed, added, removed
487
488    if 'basewhitelist' in a_data and a_data['basewhitelist'] != b_data['basewhitelist']:
489        output.append(color_format("{color_title}basewhitelist changed{color_default} from '%s' to '%s'") % (a_data['basewhitelist'], b_data['basewhitelist']))
490        if a_data['basewhitelist'] and b_data['basewhitelist']:
491            output.append("changed items: %s" % a_data['basewhitelist'].symmetric_difference(b_data['basewhitelist']))
492
493    if 'taskwhitelist' in a_data and a_data['taskwhitelist'] != b_data['taskwhitelist']:
494        output.append(color_format("{color_title}taskwhitelist changed{color_default} from '%s' to '%s'") % (a_data['taskwhitelist'], b_data['taskwhitelist']))
495        if a_data['taskwhitelist'] and b_data['taskwhitelist']:
496            output.append("changed items: %s" % a_data['taskwhitelist'].symmetric_difference(b_data['taskwhitelist']))
497
498    if a_data['taskdeps'] != b_data['taskdeps']:
499        output.append(color_format("{color_title}Task dependencies changed{color_default} from:\n%s\nto:\n%s") % (sorted(a_data['taskdeps']), sorted(b_data['taskdeps'])))
500
501    if a_data['basehash'] != b_data['basehash'] and not collapsed:
502        output.append(color_format("{color_title}basehash changed{color_default} from %s to %s") % (a_data['basehash'], b_data['basehash']))
503
504    changed, added, removed = dict_diff(a_data['gendeps'], b_data['gendeps'], a_data['basewhitelist'] & b_data['basewhitelist'])
505    if changed:
506        for dep in changed:
507            output.append(color_format("{color_title}List of dependencies for variable %s changed from '{color_default}%s{color_title}' to '{color_default}%s{color_title}'") % (dep, a_data['gendeps'][dep], b_data['gendeps'][dep]))
508            if a_data['gendeps'][dep] and b_data['gendeps'][dep]:
509                output.append("changed items: %s" % a_data['gendeps'][dep].symmetric_difference(b_data['gendeps'][dep]))
510    if added:
511        for dep in added:
512            output.append(color_format("{color_title}Dependency on variable %s was added") % (dep))
513    if removed:
514        for dep in removed:
515            output.append(color_format("{color_title}Dependency on Variable %s was removed") % (dep))
516
517
518    changed, added, removed = dict_diff(a_data['varvals'], b_data['varvals'])
519    if changed:
520        for dep in changed:
521            oldval = a_data['varvals'][dep]
522            newval = b_data['varvals'][dep]
523            if newval and oldval and ('\n' in oldval or '\n' in newval):
524                diff = difflib.unified_diff(oldval.splitlines(), newval.splitlines(), lineterm='')
525                # Cut off the first two lines, since we aren't interested in
526                # the old/new filename (they are blank anyway in this case)
527                difflines = list(diff)[2:]
528                if color:
529                    # Add colour to diff output
530                    for i, line in enumerate(difflines):
531                        if line.startswith('+'):
532                            line = color_format('{color_add}{line}', line=line)
533                            difflines[i] = line
534                        elif line.startswith('-'):
535                            line = color_format('{color_remove}{line}', line=line)
536                            difflines[i] = line
537                output.append(color_format("{color_title}Variable {var} value changed:{color_default}\n{diff}", var=dep, diff='\n'.join(difflines)))
538            elif newval and oldval and (' ' in oldval or ' ' in newval):
539                output.append(color_format("{color_title}Variable {var} value changed:{color_default}\n{diff}", var=dep, diff=worddiff_str(oldval, newval, colors)))
540            else:
541                output.append(color_format("{color_title}Variable {var} value changed from '{color_default}{oldval}{color_title}' to '{color_default}{newval}{color_title}'{color_default}", var=dep, oldval=oldval, newval=newval))
542
543    if not 'file_checksum_values' in a_data:
544         a_data['file_checksum_values'] = {}
545    if not 'file_checksum_values' in b_data:
546         b_data['file_checksum_values'] = {}
547
548    changed, added, removed = file_checksums_diff(a_data['file_checksum_values'], b_data['file_checksum_values'])
549    if changed:
550        for f, old, new in changed:
551            output.append(color_format("{color_title}Checksum for file %s changed{color_default} from %s to %s") % (f, old, new))
552    if added:
553        for f in added:
554            output.append(color_format("{color_title}Dependency on checksum of file %s was added") % (f))
555    if removed:
556        for f in removed:
557            output.append(color_format("{color_title}Dependency on checksum of file %s was removed") % (f))
558
559    if not 'runtaskdeps' in a_data:
560         a_data['runtaskdeps'] = {}
561    if not 'runtaskdeps' in b_data:
562         b_data['runtaskdeps'] = {}
563
564    if not collapsed:
565        if len(a_data['runtaskdeps']) != len(b_data['runtaskdeps']):
566            changed = ["Number of task dependencies changed"]
567        else:
568            changed = []
569            for idx, task in enumerate(a_data['runtaskdeps']):
570                a = a_data['runtaskdeps'][idx]
571                b = b_data['runtaskdeps'][idx]
572                if a_data['runtaskhashes'][a] != b_data['runtaskhashes'][b] and not collapsed:
573                    changed.append("%s with hash %s\n changed to\n%s with hash %s" % (clean_basepath(a), a_data['runtaskhashes'][a], clean_basepath(b), b_data['runtaskhashes'][b]))
574
575        if changed:
576            clean_a = clean_basepaths_list(a_data['runtaskdeps'])
577            clean_b = clean_basepaths_list(b_data['runtaskdeps'])
578            if clean_a != clean_b:
579                output.append(color_format("{color_title}runtaskdeps changed:{color_default}\n%s") % list_inline_diff(clean_a, clean_b, colors))
580            else:
581                output.append(color_format("{color_title}runtaskdeps changed:"))
582            output.append("\n".join(changed))
583
584
585    if 'runtaskhashes' in a_data and 'runtaskhashes' in b_data:
586        a = a_data['runtaskhashes']
587        b = b_data['runtaskhashes']
588        changed, added, removed = dict_diff(a, b)
589        if added:
590            for dep in added:
591                bdep_found = False
592                if removed:
593                    for bdep in removed:
594                        if b[dep] == a[bdep]:
595                            #output.append("Dependency on task %s was replaced by %s with same hash" % (dep, bdep))
596                            bdep_found = True
597                if not bdep_found:
598                    output.append(color_format("{color_title}Dependency on task %s was added{color_default} with hash %s") % (clean_basepath(dep), b[dep]))
599        if removed:
600            for dep in removed:
601                adep_found = False
602                if added:
603                    for adep in added:
604                        if b[adep] == a[dep]:
605                            #output.append("Dependency on task %s was replaced by %s with same hash" % (adep, dep))
606                            adep_found = True
607                if not adep_found:
608                    output.append(color_format("{color_title}Dependency on task %s was removed{color_default} with hash %s") % (clean_basepath(dep), a[dep]))
609        if changed:
610            for dep in changed:
611                if not collapsed:
612                    output.append(color_format("{color_title}Hash for dependent task %s changed{color_default} from %s to %s") % (clean_basepath(dep), a[dep], b[dep]))
613                if callable(recursecb):
614                    recout = recursecb(dep, a[dep], b[dep])
615                    if recout:
616                        if collapsed:
617                            output.extend(recout)
618                        else:
619                            # If a dependent hash changed, might as well print the line above and then defer to the changes in
620                            # that hash since in all likelyhood, they're the same changes this task also saw.
621                            output = [output[-1]] + recout
622
623    a_taint = a_data.get('taint', None)
624    b_taint = b_data.get('taint', None)
625    if a_taint != b_taint:
626        output.append(color_format("{color_title}Taint (by forced/invalidated task) changed{color_default} from %s to %s") % (a_taint, b_taint))
627
628    return output
629
630
631def calc_basehash(sigdata):
632    task = sigdata['task']
633    basedata = sigdata['varvals'][task]
634
635    if basedata is None:
636        basedata = ''
637
638    alldeps = sigdata['taskdeps']
639    for dep in alldeps:
640        basedata = basedata + dep
641        val = sigdata['varvals'][dep]
642        if val is not None:
643            basedata = basedata + str(val)
644
645    return hashlib.md5(basedata.encode("utf-8")).hexdigest()
646
647def calc_taskhash(sigdata):
648    data = sigdata['basehash']
649
650    for dep in sigdata['runtaskdeps']:
651        data = data + sigdata['runtaskhashes'][dep]
652
653    for c in sigdata['file_checksum_values']:
654        if c[1]:
655            data = data + c[1]
656
657    if 'taint' in sigdata:
658        if 'nostamp:' in sigdata['taint']:
659            data = data + sigdata['taint'][8:]
660        else:
661            data = data + sigdata['taint']
662
663    return hashlib.md5(data.encode("utf-8")).hexdigest()
664
665
666def dump_sigfile(a):
667    output = []
668
669    with open(a, 'rb') as f:
670        p1 = pickle.Unpickler(f)
671        a_data = p1.load()
672
673    output.append("basewhitelist: %s" % (a_data['basewhitelist']))
674
675    output.append("taskwhitelist: %s" % (a_data['taskwhitelist']))
676
677    output.append("Task dependencies: %s" % (sorted(a_data['taskdeps'])))
678
679    output.append("basehash: %s" % (a_data['basehash']))
680
681    for dep in a_data['gendeps']:
682        output.append("List of dependencies for variable %s is %s" % (dep, a_data['gendeps'][dep]))
683
684    for dep in a_data['varvals']:
685        output.append("Variable %s value is %s" % (dep, a_data['varvals'][dep]))
686
687    if 'runtaskdeps' in a_data:
688        output.append("Tasks this task depends on: %s" % (a_data['runtaskdeps']))
689
690    if 'file_checksum_values' in a_data:
691        output.append("This task depends on the checksums of files: %s" % (a_data['file_checksum_values']))
692
693    if 'runtaskhashes' in a_data:
694        for dep in a_data['runtaskhashes']:
695            output.append("Hash for dependent task %s is %s" % (dep, a_data['runtaskhashes'][dep]))
696
697    if 'taint' in a_data:
698        output.append("Tainted (by forced/invalidated task): %s" % a_data['taint'])
699
700    if 'task' in a_data:
701        computed_basehash = calc_basehash(a_data)
702        output.append("Computed base hash is %s and from file %s" % (computed_basehash, a_data['basehash']))
703    else:
704        output.append("Unable to compute base hash")
705
706    computed_taskhash = calc_taskhash(a_data)
707    output.append("Computed task hash is %s" % computed_taskhash)
708
709    return output
710