This is an archive of the discontinued Mercurial Phabricator instance.

Differential D6768

split: handle partial commit of renames when doing split or record (issue5723)
ClosedPublic

Authored by spectral on Aug 27 2019, 3:45 PM.

Details

Reviewers
None
Group Reviewers
hg-reviewers
Commits
rHG3cf091843b4f: split: handle partial commit of renames when doing split or record (issue5723)
Summary

When using split or record, using either interface (text or curses), selecting
portions of the file to be committed/recorded did not work; the entire file was
treated as having been selected. This was because the logic for handling partial
application of the patches relies on knowing what files are "new with
modifications" and it doesn't treat "rename destination" as "new".

There was a complicating issue, however. We're relying on the patch header
specifying the copy from/to information, which works as long as the 'copy from'
file is there. In the case of renames, however, the 'rename from' file is *not*
there, so we need to add it back.

Diff Detail

Repository
rHG Mercurial
Lint
Lint Skipped
Unit
Unit Tests Skipped

Event Timeline

spectral created this revision.Aug 27 2019, 3:45 PM
spectral added a commit: rHG3cf091843b4f: split: handle partial commit of renames when doing split or record (issue5723).Sep 5 2019, 2:57 PM
This revision was not accepted when it landed; it landed in state Needs Review.
Closed by commit rHG3cf091843b4f: split: handle partial commit of renames when doing split or record (issue5723) (authored by spectral). · Explain Why
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

PathPackages
Mmercurial/cmdutil.py (22 lines)
Mmercurial/patch.py (2 lines)
Mtests/test-split.t (90 lines)

Diff 16323

mercurial/cmdutil.py

_linebelow = "^HG: ------------------------ >8 ------------------------$" _linebelow = "^HG: ------------------------ >8 ------------------------$"
def ishunk(x): def ishunk(x):
hunkclasses = (crecordmod.uihunk, patch.recordhunk) hunkclasses = (crecordmod.uihunk, patch.recordhunk)
return isinstance(x, hunkclasses) return isinstance(x, hunkclasses)
def newandmodified(chunks, originalchunks): def newandmodified(chunks, originalchunks):
newlyaddedandmodifiedfiles = set() newlyaddedandmodifiedfiles = set()
alsorestore = set()
for chunk in chunks: for chunk in chunks:
if (ishunk(chunk) and chunk.header.isnewfile() and chunk not in if (ishunk(chunk) and chunk.header.isnewfile() and chunk not in
originalchunks): originalchunks):
newlyaddedandmodifiedfiles.add(chunk.header.filename()) newlyaddedandmodifiedfiles.add(chunk.header.filename())
return newlyaddedandmodifiedfiles alsorestore.update(set(chunk.header.files()) -
set([chunk.header.filename()]))
return newlyaddedandmodifiedfiles, alsorestore
def parsealiases(cmd): def parsealiases(cmd):
return cmd.split("|") return cmd.split("|")
def setupwrapcolorwrite(ui): def setupwrapcolorwrite(ui):
# wrap ui.write so diff output can be labeled/colorized # wrap ui.write so diff output can be labeled/colorized
def wrapwrite(orig, *args, **kw): def wrapwrite(orig, *args, **kw):
label = kw.pop(r'label', '') label = kw.pop(r'label', '')
try: try:
chunks, newopts = filterfn(ui, originalchunks, match) chunks, newopts = filterfn(ui, originalchunks, match)
except error.PatchError as err: except error.PatchError as err:
raise error.Abort(_('error parsing patch: %s') % err) raise error.Abort(_('error parsing patch: %s') % err)
opts.update(newopts) opts.update(newopts)
# We need to keep a backup of files that have been newly added and # We need to keep a backup of files that have been newly added and
# modified during the recording process because there is a previous # modified during the recording process because there is a previous
# version without the edit in the workdir # version without the edit in the workdir. We also will need to restore
newlyaddedandmodifiedfiles = newandmodified(chunks, originalchunks) # files that were the sources of renames so that the patch application
# works.
newlyaddedandmodifiedfiles, alsorestore = newandmodified(chunks,
originalchunks)
contenders = set() contenders = set()
for h in chunks: for h in chunks:
try: try:
contenders.update(set(h.files())) contenders.update(set(h.files()))
except AttributeError: except AttributeError:
pass pass
changed = status.modified + status.added + status.removed changed = status.modified + status.added + status.removed
fp.truncate(0) fp.truncate(0)
fp.write(reviewedpatch) fp.write(reviewedpatch)
fp.seek(0) fp.seek(0)
[os.unlink(repo.wjoin(c)) for c in newlyaddedandmodifiedfiles] [os.unlink(repo.wjoin(c)) for c in newlyaddedandmodifiedfiles]
# 3a. apply filtered patch to clean repo (clean) # 3a. apply filtered patch to clean repo (clean)
if backups: if backups:
# Equivalent to hg.revert # Equivalent to hg.revert
m = scmutil.matchfiles(repo, backups.keys()) m = scmutil.matchfiles(repo, set(backups.keys()) | alsorestore)
mergemod.update(repo, repo.dirstate.p1(), branchmerge=False, mergemod.update(repo, repo.dirstate.p1(), branchmerge=False,
force=True, matcher=m) force=True, matcher=m)
# 3b. (apply) # 3b. (apply)
if dopatch: if dopatch:
try: try:
ui.debug('applying patch\n') ui.debug('applying patch\n')
ui.debug(fp.getvalue()) ui.debug(fp.getvalue())
chunks, opts = recordfilter(repo.ui, originalchunks, match, chunks, opts = recordfilter(repo.ui, originalchunks, match,
operation=operation) operation=operation)
if operation == 'discard': if operation == 'discard':
chunks = patch.reversehunks(chunks) chunks = patch.reversehunks(chunks)
except error.PatchError as err: except error.PatchError as err:
raise error.Abort(_('error parsing patch: %s') % err) raise error.Abort(_('error parsing patch: %s') % err)
newlyaddedandmodifiedfiles = newandmodified(chunks, originalchunks) # FIXME: when doing an interactive revert of a copy, there's no way of
# performing a partial revert of the added file, the only option is
# "remove added file <name> (Yn)?", so we don't need to worry about the
# alsorestore value. Ideally we'd be able to partially revert
# copied/renamed files.
newlyaddedandmodifiedfiles, unusedalsorestore = newandmodified(
chunks, originalchunks)
if tobackup is None: if tobackup is None:
tobackup = set() tobackup = set()
# Apply changes # Apply changes
fp = stringio() fp = stringio()
# chunks are serialized per file, but files aren't sorted # chunks are serialized per file, but files aren't sorted
for f in sorted(set(c.header.filename() for c in chunks if ishunk(c))): for f in sorted(set(c.header.filename() for c in chunks if ishunk(c))):
prntstatusmsg('revert', f) prntstatusmsg('revert', f)
files = set() files = set()

mercurial/patch.py

class header(object): class header(object):
"""patch header """patch header
""" """
diffgit_re = re.compile('diff --git a/(.*) b/(.*)$') diffgit_re = re.compile('diff --git a/(.*) b/(.*)$')
diff_re = re.compile('diff -r .* (.*)$') diff_re = re.compile('diff -r .* (.*)$')
allhunks_re = re.compile('(?:index|deleted file) ') allhunks_re = re.compile('(?:index|deleted file) ')
pretty_re = re.compile('(?:new file|deleted file) ') pretty_re = re.compile('(?:new file|deleted file) ')
special_re = re.compile('(?:index|deleted|copy|rename|new mode) ') special_re = re.compile('(?:index|deleted|copy|rename|new mode) ')
newfile_re = re.compile('(?:new file|copy to)') newfile_re = re.compile('(?:new file|copy to|rename to)')
def __init__(self, header): def __init__(self, header):
self.header = header self.header = header
self.hunks = [] self.hunks = []
def binary(self): def binary(self):
return any(h.startswith('index ') for h in self.header) return any(h.startswith('index ') for h in self.header)

tests/test-split.t

TODO: Fix this on Windows. See issue 2020 and 5883 TODO: Fix this on Windows. See issue 2020 and 5883
$ printf 'y\ny\ny\n' | hg split $ printf 'y\ny\ny\n' | hg split
abort: cannot split an empty revision abort: cannot split an empty revision
[255] [255]
#endif #endif
Test that splitting moves works properly (issue5723)
----------------------------------------------------
$ hg init $TESTTMP/issue5723-mv
$ cd $TESTTMP/issue5723-mv
$ printf '1\n2\n' > file
$ hg ci -qAm initial
$ hg mv file file2
$ printf 'a\nb\n1\n2\n3\n4\n' > file2
$ cat > $TESTTMP/messages <<EOF
> split1, keeping only the numbered lines
> --
> split2, keeping the lettered lines
> EOF
$ hg ci -m 'move and modify'
$ printf 'y\nn\na\na\n' | hg split
diff --git a/file b/file2
rename from file
rename to file2
2 hunks, 4 lines changed
examine changes to 'file' and 'file2'?
(enter ? for help) [Ynesfdaq?] y
@@ -0,0 +1,2 @@
+a
+b
record change 1/2 to 'file2'?
(enter ? for help) [Ynesfdaq?] n
@@ -2,0 +5,2 @@ 2
+3
+4
record change 2/2 to 'file2'?
(enter ? for help) [Ynesfdaq?] a
EDITOR: HG: Splitting 8c42fa635116. Write commit message for the first split changeset.
EDITOR: move and modify
EDITOR:
EDITOR:
EDITOR: HG: Enter commit message. Lines beginning with 'HG:' are removed.
EDITOR: HG: Leave message empty to abort commit.
EDITOR: HG: --
EDITOR: HG: user: test
EDITOR: HG: branch 'default'
EDITOR: HG: added file2
EDITOR: HG: removed file
created new head
diff --git a/file2 b/file2
1 hunks, 2 lines changed
examine changes to 'file2'?
(enter ? for help) [Ynesfdaq?] a
EDITOR: HG: Splitting 8c42fa635116. So far it has been split into:
EDITOR: HG: - 478be2a70c27: split1, keeping only the numbered lines
EDITOR: HG: Write commit message for the next split changeset.
EDITOR: move and modify
EDITOR:
EDITOR:
EDITOR: HG: Enter commit message. Lines beginning with 'HG:' are removed.
EDITOR: HG: Leave message empty to abort commit.
EDITOR: HG: --
EDITOR: HG: user: test
EDITOR: HG: branch 'default'
EDITOR: HG: changed file2
saved backup bundle to $TESTTMP/issue5723-mv/.hg/strip-backup/8c42fa635116-a38044d4-split.hg (obsstore-off !)
$ hg log -T '{desc}: {files%"{file} "}\n'
split2, keeping the lettered lines: file2
split1, keeping only the numbered lines: file file2
initial: file
$ cat file2
a
b
1
2
3
4
$ hg cat -r ".^" file2
1
2
3
4
$ hg cat -r . file2
a
b
1
2
3
4
Test that splitting copies works properly (issue5723) Test that splitting copies works properly (issue5723)
---------------------------------------------------- ----------------------------------------------------
$ hg init $TESTTMP/issue5723-cp $ hg init $TESTTMP/issue5723-cp
$ cd $TESTTMP/issue5723-cp $ cd $TESTTMP/issue5723-cp
$ printf '1\n2\n' > file $ printf '1\n2\n' > file
$ hg ci -qAm initial $ hg ci -qAm initial
$ hg cp file file2 $ hg cp file file2