Matt Mackall <mpm@selenic.com> [Fri, 22 Jan 2016 12:08:20 -0600] rev 27945
log: speed up single file log with hidden revs (issue4747)
On repos with lots of heads, the filelog() code could spend several
minutes decompressing manifests. This change instead tries to
efficiently scan the changelog for candidates and decompress as few
manifests as possible. This is a regression introduced in 3.3 by the
linkrev adjustment code. Prior to that, filelog was nearly instant.
For the repo in the bug report, this improves time of a simple log
command from ~3 minutes to ~.5 seconds, a 360x speedup.
For the main Mercurial repo, a log of commands.py slows down from
1.14s to 1.45s, a 27% slowdown. This is still faster than the file()
revset, which takes 2.1 seconds.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 23 Jan 2016 23:32:49 -0500] rev 27944
largefiles: report the missing file count after a commit that does nothing
This is the same warning that is printed for normal files.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 23 Jan 2016 23:24:30 -0500] rev 27943
commit: factor the post commit status check into a cmdutil method
The largefiles extension needs to set lfstatus for this status call. Otherwise,
if a missing largefile is explicitly named, a confusing message is issued that
says the largefile wasn't found, followed by another that says nothing changed.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 23 Jan 2016 20:51:17 -0500] rev 27942
largefiles: fix an explicit largefile commit after a remove (issue4969)
The change in b68797f244e4 to handle a normal -> largefile switch was too
aggressive in preserving the original matcher names. If a largefile is
explicitly provided by the user, but only the standin exists in dirstate, then
only the standin can be committed.
There's still maybe an issue when the largefile is deleted outside of Mercurial:
$ rm large
$ hg ci -m "oops" large
large: The system cannot find the file specified
nothing changed
[1]
Martin von Zweigbergk <martinvonz@google.com> [Mon, 25 Jan 2016 15:48:35 -0800] rev 27941
context: back out sneaky code change in documentation change
In 81b391a45264 (context: clarify why we don't compare file contents
when nodeid differs, 2016-01-12), I also changed "node2 != _newnode"
into "self.rev() is not None". I don't remember why. They are similar,
but the former also catches the case where the file is clean in the
dirstate (so node2 is not _newnode), but different from the "other"
context. This resulted in unnecessary file content comparison a few
lines further down in the code. Let's just back out the code change.
Thanks to Durham Goode for spotting this.
Yuya Nishihara <yuya@tcha.org> [Thu, 23 Jul 2015 23:41:29 +0900] rev 27940
templater: abort if infinite recursion detected while compiling
In this case, a template is parsed recursively with no thunk for lazy
evaluation. This patch prevents recursion by putting a dummy of the same name
into a cache that will be referenced while parsing if there's a recursion.
changeset = {files % changeset}\n
~~~~~~~~~
= [(_runrecursivesymbol, 'changeset')]
Yuya Nishihara <yuya@tcha.org> [Wed, 22 Jul 2015 23:29:41 +0900] rev 27939
templater: abort if infinite recursion detected while evaluation (issue4758)
It would be nice if we could detect recursion at the parsing phase, but we
can't because a template can refer to a keyword of the same name. For example,
"rev = {rev}" is valid if rev is a keyword, and we don't know if rev is a
keyword or a template while parsing.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 22 Jan 2016 16:31:50 -0800] rev 27938
exchange: set 'treemanifest' param on pushed changegroups too
In 5c0fd878779c (treemanifests: set bundle2 part parameter indicating
treemanifest, 2016-01-08), I didn't realize I had to set the parameter
separately for getbundle and unbundle. Having the parameter there on
push allows us to push to an empty repo and have the requirements
updated correctly.
Laurent Charignon <lcharignon@fb.com> [Mon, 25 Jan 2016 11:37:02 -0800] rev 27937
crecord: fix help bar display issue (issue5063)
Before this patch, the help bar in crecord wouldn't be printed correctly when
the terminal window didn't have enough column to display it. This patch adds
logic to make sure that the help bar message is always displayed. We use an
ellipsis when it is not possible to display the complete message.
Laurent Charignon <lcharignon@fb.com> [Mon, 25 Jan 2016 10:13:03 -0800] rev 27936
crecord: fix typo in the help text
In the crecord help dialog, the toggle all option was wrongfully documented.
Instead of using 'a', one must use 'A' to toggle all the hunks. The crecord
header that is always displayed on the screen contains the right shortcut and
does not need to be changed.