Merge with stable

static-http: disable lazy parsing This only hits if you're crazy enough to use static-http on a repository with revlogs larger than 1M. Don't do it.

expand paths to local repository or bundle in appropriate classes This avoids problem with unexpanded paths when it's not possible to expand it at higher level (for example, if file:~/path/ is supplied as path in schemes).

push: document return values between various repo methods. This starts at localrepository.push() and seeps down to push_addchangegroup(), push_unbundle(), prepush(), addchangegroup(), and leaks out to sshrepository.unbundle(), sshrepository.addchangegroup(), and httprepository.unbundle(). Seems to cover everything you ever wanted to know about pushing but were afraid to ask.

cmdutil: Warn when trying to copy/rename --after to a nonexistant file. Currently running 'hg rename --after foo.txt bar.typo' is a silent no-op. This patch adds a warning. It also updates the copy and rename tests. No actual functionality is changed. fixes issue 1822

slightly improve memctx api

patchbomb: Reply-To support From RFC 5322: an optional reply-to field MAY also be included, which contains the field name "Reply-To" and a comma-separated list of one or more addresses. [...] When the "Reply-To:" field is present, it indicates the address(es) to which the author of the message suggests that replies be sent. In the absence of the "Reply-To:" field, replies SHOULD by default be sent to the mailbox(es) specified in the "From:" field unless otherwise specified by the person composing the reply. Reply-To addresses may be specified either via command line with --reply-to or via the 'email' or 'patchbomb' sections of the config file.

merge: tool.check = prompt will force an interactive merge check tool.check = prompt can be used when the exit status of a merge tool is unreliable but an explicit user signoff on the merge result is acceptable.

merge: introduce tool.check parameter tool.check is a list of check options, and can be used in place of tool.checkchanged and tool.checkconflicts: Equivalences: tool.checkchanged = yes tool.checkconflicts = no tool.check = changed tool.checkchanged = no tool.checkconflicts = yes tool.check = conflicts tool.checkchanged = yes tool.checkconflicts = yes tool.check = changed, conflicts Add _toollist() wrapper for ui.configlist() to implement this consistently. checkchanged and checkconflicts are still supported, but check is preferred for implementing new check options.

give worst-case 'merge' merge-tool lowest priority Another tool had -10 already. Since 'merge' is clearly a worst-case tool (internal), lowering to -100 ensures there's plenty of room for slightly better cases.

merge: tool.premerge=keep will leave premerge markers in $local

merge with mpm

merge with stable

context: fallback function should not return None

test-subrepo-svn: fix strange svn behaviour on windows I got a full path when pusing the external like: Sending C:/foo/bar/test-subrepo-svn/sub/t/s/alpha instead of the relative path

tests: force color ansi mode on Windows

acl: improve undefined group error handling

test-acl: avoid $HGTMP on windows sed does not like colons and there are other path normalization issues.

acl: grp module is not available on windows

convert/svn: report path discovery progress When converting non-local repositories, scanning changed paths before retrieving data can be almost as slow as retrieving the data itself, thanks to HTTP calls overhead.

convert: display files data retrieval progress

convert: add progress support

convert: merge sources getmode() into getfile()

convert/svn: remove useless sort

convert/svn: list files explicitely, stop checking their type We do not care about directories when looking for recursively added or removed items, and the redundant _checkpath() call is expensive with remote repositories.

acl: clarify acl.sources, fix typo

convert/svn: remove unused variable

Merge with crew-stable

convert/svn: handle files/links replaced by dirs (issue2166)

convert/svn: do not retrieve removed files It gives us a way to not retrieve content of entries we know are no longer files. And it is faster when converting remote repositories.

convert/svn: correctly encode deleted entry paths This recode call was removed in ede2247e61aa, because it looked the encode(decode()) construct was a no-op. In fact, the first decode() call was wrong, and entries still have to be encoded before being passed to the sink.

convert/svn: remove broken but unused copy filtering code For some reason, if a copy source is deleted in the same revision it is referenced, it is filtered out. This is silly, because this happens all the time with move operations. Fortunately, the filtering code is buggy and ends being a no-op 99% of the time, since it does not delete the right key. Just remove all this nonsense.