I have spotted the biggest bottleneck in "bdiff.c". Actually it was
pretty easy to find after I recompiled the python interpreter and
mercurial for profiling.
In "bdiff.c" function "equatelines" allocates the minimum hash table
size, which can lead to tons of collisions. I introduced an
"overcommit" factor of 16, this is, I allocate 16 times more memory
than the minimum value. Overcommiting 128 times does not improve the
performance over the 16-times case.
changeset = 'changeset: {rev}:{node|short}\n{branches}{tags}{parents}user: {author}\ndate: {date|date}\nsummary: {desc|firstline}\n\n'
changeset_quiet = '{rev}:{node|short}\n'
changeset_verbose = 'changeset: {rev}:{node|short}\n{branches}{tags}{parents}{manifest}user: {author}\ndate: {date|date}\n{files}{file_adds}{file_dels}{file_copies}description:\n{desc|strip}\n\n\n'
changeset_debug = 'changeset: {rev}:{node}\n{branches}{tags}{parents}{manifest}user: {author}\ndate: {date|date}\n{files}{file_adds}{file_dels}{file_copies}{extras}description:\n{desc|strip}\n\n\n'
start_files = 'files: '
file = ' {file}'
end_files = '\n'
start_file_adds = 'files+: '
file_add = ' {file_add}'
end_file_adds = '\n'
start_file_dels = 'files-: '
file_del = ' {file_del}'
end_file_dels = '\n'
start_file_copies = 'copies: '
file_copy = ' {name} ({source})'
end_file_copies = '\n'
parent = 'parent: {rev}:{node|formatnode}\n'
manifest = 'manifest: {rev}:{node}\n'
branch = 'branch: {branch}\n'
tag = 'tag: {tag}\n'
extra = 'extra: {key}={value|stringescape}\n'