tests/test-context.py

commit: improve the files field of changelog for merges

Currently, the files list of merge commits repeats all the deletions
(either actual deletions, or files that got renamed) that happened
between base and p2 of the merge. If p2 is the main branch, the list
can easily be much bigger than the change being merged.

This results in various problems worth improving:
- changelog is bigger than necessary
- `hg log directory` lists many unrelated merge commits, and `hg log
  -v -r commit` frequently fills multiple screens worth of files
- it possibly slows down adjustlinkrev, by forcing it to read more
  manifests, and that function can certainly be a bottleneck
- the server side of pulls can waste a lot of time simply opening the
  filelogs for pointless files (the constant factors for opening even
  a tiny filelog is apparently pretty bad)

So stop listing such files as described in the code.  Impacted merge
commits and their descendants get a different hash than they would
have without this. This doesn't seem problematic, except for
convert. The previous commit helped with that in the hg->hg case (but
if you do svn->hg twice from scratch, hashes can still change).

The rest of the description is numbers. I don't have much to report,
because recreating the files list of existing repositories is not
easy:
- debugupgradeformat and bundle/unbundle don't recreate the list
- export/import tends to choke quickly applying patches or on
  description that contain diffs,
- merge commits from the convert extension don't have the right files
  list for reasons orthogonal to the current commit
- replaying the merge with hg update/hg merge/hg revert --all/hg
  commit can end up failing in hg revert
- I wasn't sure that using debugsetparents + debugrebuilddirstate
  would really build the right thing

I measured commit time before and after this change, in a case with no
files filtered out, several files filtered out (no difference) and 5k
files filtered out (+1% time).

Recreating the 100 more recent merges in a private repo, the
concatenated uncompressed files lists goes from 1.12MB to
0.52MB. Excluding 3 merges that are not representative, then the size
goes from 570k to 15k.
I converted part of mozilla-central, and observed file list shrinking
quite a bit too, starting at the very first merge, 733641d9feaf, going
from 550 files to 10 files (although they have relatively few merges,
so they probably wouldn't care).

Differential Revision: https://phab.mercurial-scm.org/D6613

from __future__ import absolute_import, print_function
import os
import stat
import sys
from mercurial.node import hex
from mercurial import (
    context,
    diffutil,
    encoding,
    hg,
    scmutil,
    ui as uimod,
)

print_ = print
def print(*args, **kwargs):
    """print() wrapper that flushes stdout buffers to avoid py3 buffer issues

    We could also just write directly to sys.stdout.buffer the way the
    ui object will, but this was easier for porting the test.
    """
    print_(*args, **kwargs)
    sys.stdout.flush()

def printb(data, end=b'\n'):
    out = getattr(sys.stdout, 'buffer', sys.stdout)
    out.write(data + end)
    out.flush()

ui = uimod.ui.load()

repo = hg.repository(ui, b'test1', create=1)
os.chdir('test1')

# create 'foo' with fixed time stamp
f = open('foo', 'wb')
f.write(b'foo\n')
f.close()
os.utime('foo', (1000, 1000))

# add+commit 'foo'
repo[None].add([b'foo'])
repo.commit(text=b'commit1', date=b"0 0")

d = repo[None][b'foo'].date()
if os.name == 'nt':
    d = d[:2]
print("workingfilectx.date = (%d, %d)" % d)

# test memctx with non-ASCII commit message

def filectxfn(repo, memctx, path):
    return context.memfilectx(repo, memctx, b"foo", b"")

ctx = context.memctx(repo, [b'tip', None],
                     encoding.tolocal(b"Gr\xc3\xbcezi!"),
                     [b"foo"], filectxfn)
ctx.commit()
for enc in "ASCII", "Latin-1", "UTF-8":
    encoding.encoding = enc
    printb(b"%-8s: %s" % (enc.encode('ascii'), repo[b"tip"].description()))

# test performing a status

def getfilectx(repo, memctx, f):
    fctx = memctx.p1()[f]
    data, flags = fctx.data(), fctx.flags()
    if f == b'foo':
        data += b'bar\n'
    return context.memfilectx(
        repo, memctx, f, data, b'l' in flags, b'x' in flags)

ctxa = repo[0]
ctxb = context.memctx(repo, [ctxa.node(), None], b"test diff", [b"foo"],
                      getfilectx, ctxa.user(), ctxa.date())

print(ctxb.status(ctxa))

# test performing a diff on a memctx
diffopts = diffutil.diffallopts(repo.ui, {b'git': True})
for d in ctxb.diff(ctxa, opts=diffopts):
    printb(d, end=b'')

# test safeness and correctness of "ctx.status()"
print('= checking context.status():')

# ancestor "wcctx ~ 2"
actx2 = repo[b'.']

repo.wwrite(b'bar-m', b'bar-m\n', b'')
repo.wwrite(b'bar-r', b'bar-r\n', b'')
repo[None].add([b'bar-m', b'bar-r'])
repo.commit(text=b'add bar-m, bar-r', date=b"0 0")

# ancestor "wcctx ~ 1"
actx1 = repo[b'.']

repo.wwrite(b'bar-m', b'bar-m bar-m\n', b'')
repo.wwrite(b'bar-a', b'bar-a\n', b'')
repo[None].add([b'bar-a'])
repo[None].forget([b'bar-r'])

# status at this point:
#   M bar-m
#   A bar-a
#   R bar-r
#   C foo

from mercurial import scmutil

print('== checking workingctx.status:')

wctx = repo[None]
print('wctx._status=%s' % (str(wctx._status)))

print('=== with "pattern match":')
print(actx1.status(other=wctx,
                   match=scmutil.matchfiles(repo, [b'bar-m', b'foo'])))
print('wctx._status=%s' % (str(wctx._status)))
print(actx2.status(other=wctx,
                   match=scmutil.matchfiles(repo, [b'bar-m', b'foo'])))
print('wctx._status=%s' % (str(wctx._status)))

print('=== with "always match" and "listclean=True":')
print(actx1.status(other=wctx, listclean=True))
print('wctx._status=%s' % (str(wctx._status)))
print(actx2.status(other=wctx, listclean=True))
print('wctx._status=%s' % (str(wctx._status)))

print("== checking workingcommitctx.status:")

wcctx = context.workingcommitctx(repo,
                                 scmutil.status([b'bar-m'],
                                                [b'bar-a'],
                                                [],
                                                [], [], [], []),
                                 text=b'', date=b'0 0')
print('wcctx._status=%s' % (str(wcctx._status)))

print('=== with "always match":')
print(actx1.status(other=wcctx))
print('wcctx._status=%s' % (str(wcctx._status)))
print(actx2.status(other=wcctx))
print('wcctx._status=%s' % (str(wcctx._status)))

print('=== with "always match" and "listclean=True":')
print(actx1.status(other=wcctx, listclean=True))
print('wcctx._status=%s' % (str(wcctx._status)))
print(actx2.status(other=wcctx, listclean=True))
print('wcctx._status=%s' % (str(wcctx._status)))

print('=== with "pattern match":')
print(actx1.status(other=wcctx,
                   match=scmutil.matchfiles(repo, [b'bar-m', b'foo'])))
print('wcctx._status=%s' % (str(wcctx._status)))
print(actx2.status(other=wcctx,
                   match=scmutil.matchfiles(repo, [b'bar-m', b'foo'])))
print('wcctx._status=%s' % (str(wcctx._status)))

print('=== with "pattern match" and "listclean=True":')
print(actx1.status(other=wcctx,
                   match=scmutil.matchfiles(repo, [b'bar-r', b'foo']),
                   listclean=True))
print('wcctx._status=%s' % (str(wcctx._status)))
print(actx2.status(other=wcctx,
                   match=scmutil.matchfiles(repo, [b'bar-r', b'foo']),
                   listclean=True))
print('wcctx._status=%s' % (str(wcctx._status)))

os.chdir('..')

# test manifestlog being changed
print('== commit with manifestlog invalidated')

repo = hg.repository(ui, b'test2', create=1)
os.chdir('test2')

# make some commits
for i in [b'1', b'2', b'3']:
    with open(i, 'wb') as f:
        f.write(i)
    status = scmutil.status([], [i], [], [], [], [], [])
    ctx = context.workingcommitctx(repo, status, text=i, user=b'[e-mail]',
                                   date=(0, 0))
    ctx.p1().manifest() # side effect: cache manifestctx
    n = repo.commitctx(ctx)
    printb(b'commit %s: %s' % (i, hex(n)))

    # touch 00manifest.i mtime so storecache could expire.
    # repo.__dict__['manifestlog'] is deleted by transaction releasefn.
    st = repo.svfs.stat(b'00manifest.i')
    repo.svfs.utime(b'00manifest.i',
                    (st[stat.ST_MTIME] + 1, st[stat.ST_MTIME] + 1))

    # read the file just committed
    try:
        if repo[n][i].data() != i:
            print('data mismatch')
    except Exception as ex:
        print('cannot read data: %r' % ex)

with repo.wlock(), repo.lock(), repo.transaction(b'test'):
    with open(b'4', 'wb') as f:
        f.write(b'4')
    repo.dirstate.normal(b'4')
    repo.commit(b'4')
    revsbefore = len(repo.changelog)
    repo.invalidate(clearfilecache=True)
    revsafter = len(repo.changelog)
    if revsbefore != revsafter:
        print('changeset lost by repo.invalidate()')