view i18n/posplit @ 32151:4d504e541d3d

rebase: use matcher to optimize manifestmerge

The old merge code would call manifestmerge and calculate the complete diff
between the source to the destination. In many cases, like rebase, the vast
majority of differences between the source and destination are irrelevant
because they are differences between the destination and the common ancestor
only, and therefore don't affect the merge. Since most actions are 'keep', all
the effort to compute them is wasted.

Instead, let's compute the difference between the source and the common ancestor
and only perform the diff of those files against the merge destination. When
using treemanifest, this lets us avoid loading almost the entire tree when
rebasing from a very old ancestor. This speeds up rebase of an old stack of 27
commits by 20x.

In mozilla-central, without treemanifest, when rebasing a commit from
default~100000 to default, this speeds up the manifestmerge step from 2.6s to
1.2s.  However, the additional diff adds an overhead to all manifestmerge calls,
especially for flat manifests. When rebasing a commit from default~1 to default
it appears to add 100ms in mozilla-central. While we could put this optimization
behind a flag, I think the fact that it makes merge O(number of changes being
applied) instead of O(number of changes between X and Y) justifies it.

#!/usr/bin/env python
#
# posplit - split messages in paragraphs on .po/.pot files
#
# license: MIT/X11/Expat
#

from __future__ import absolute_import, print_function

import polib
import re
import sys

def addentry(po, entry, cache):
    e = cache.get(entry.msgid)
    if e:
        e.occurrences.extend(entry.occurrences)
    else:
        po.append(entry)
        cache[entry.msgid] = entry

def mkentry(orig, delta, msgid, msgstr):
    entry = polib.POEntry()
    entry.merge(orig)
    entry.msgid = msgid or orig.msgid
    entry.msgstr = msgstr or orig.msgstr
    entry.occurrences = [(p, int(l) + delta) for (p, l) in orig.occurrences]
    return entry

if __name__ == "__main__":
    po = polib.pofile(sys.argv[1])

    cache = {}
    entries = po[:]
    po[:] = []
    findd = re.compile(r' *\.\. (\w+)::') # for finding directives
    for entry in entries:
        msgids = entry.msgid.split(u'\n\n')
        if entry.msgstr:
            msgstrs = entry.msgstr.split(u'\n\n')
        else:
            msgstrs = [u''] * len(msgids)

        if len(msgids) != len(msgstrs):
            # places the whole existing translation as a fuzzy
            # translation for each paragraph, to give the
            # translator a chance to recover part of the old
            # translation - erasing extra paragraphs is
            # probably better than retranslating all from start
            if 'fuzzy' not in entry.flags:
                entry.flags.append('fuzzy')
            msgstrs = [entry.msgstr] * len(msgids)

        delta = 0
        for msgid, msgstr in zip(msgids, msgstrs):
            if msgid and msgid != '::':
                newentry = mkentry(entry, delta, msgid, msgstr)
                mdirective = findd.match(msgid)
                if mdirective:
                    if not msgid[mdirective.end():].rstrip():
                        # only directive, nothing to translate here
                        delta += 2
                        continue
                    directive = mdirective.group(1)
                    if directive in ('container', 'include'):
                        if msgid.rstrip('\n').count('\n') == 0:
                            # only rst syntax, nothing to translate
                            delta += 2
                            continue
                        else:
                            # lines following directly, unexpected
                            print('Warning: text follows line with directive' \
                                  ' %s' % directive)
                    comment = 'do not translate: .. %s::' % directive
                    if not newentry.comment:
                        newentry.comment = comment
                    elif comment not in newentry.comment:
                        newentry.comment += '\n' + comment
                addentry(po, newentry, cache)
            delta += 2 + msgid.count('\n')
    po.save()