make mq and tags hardlink safe The code didn't check for modes like "r+" or "rb+". Many thanks to agriffis for noticing it.

dirstate.walk: speed up calling match function

dirstate.walk: reduce sorting in step 3

dirstate.walk: inline imatch This lets us carefully avoid calling ignore and match where possible in the fast path.

dirstate.walk: more cleanups - group internal functions - simplify imatch - rename _join to join - use nf in step 3 - use .hg in results

dirstate.walk: fold in _supported - inline tests for regular or link files - simplify bad type reporting

dirstate.walk: build a dict rather than yield Since we're already building a seen dict, we might as well put our results in it to go slightly faster.

dirstate.walk: minor cleanups - add dirkind - simplify loop structure - avoid pconvert and join on directory entries (big speed bump) - remove redundant seen set in step 3

dirstate.walk: push sorting up

dirstate.walk: pull directory scanning into top-level loop

dirstate.walk: unify match.dir logic

dirstate.walk: simplify .hg scan bisect logic

dirstate.walk: track normalized directory names

dirstate: simplify normalize logic

dirstate.walk: change names for dc and known known -> seen dc -> dmap

dirstate.walk: fold findfiles into main walk loop

dirstate.walk: eliminate filter function - remove _filter and delete original - improve some filtering logic

dirstate.walk: eliminate src from yield - restructure 'step two' to stat unseen files - remove extra stat logic in status - remove redundant src

minor status fixups

Merge with crew-stable

test-branches: add a test for long branch name (issue 1230)

branches unparseable output fix (issue1230)

Bug:1201 hg convert on CVS working copy produces Traceback Fix identified by frank@kingswood-consulting.co.uk Changed usage fron os.environ["HOME"] to expanduser("~/.cvspass") as this is the only usage of this construct in mercurial sources.

Merge with main test-remove is still failing for status() does not return removed files in a sorted list. We can live with this for now, a fix is coming soon.

dirstate: remove superfluous normalize() call in walk()

localrepo: fix status() typos from 854b907527e5a and 97c12b1ed1e0

context: trigger missing username warning only when necessary

localrepo: do not modify a dictionary being iterated in status()

dirstate: fix _droppath() typo from 80605a8127e0

Add test for case folding issues

Sort removes first when applying updates (fixes issues 750 and 912) This change ensures that removes happen first in applyupdates(). This avoids issues where we try to make a case-only rename of a file on a case insensitive system. Without this patch, the add of the new name happens before the remove of the old one - which results in the file not existing, as the two names are effectively the same. With the patch, the old name gets removed then the new one gets added, which is always safe.

Added missing --config global option to zsh_completion. The global options --encoding and --encodingmode where also added. Without these options some functionality breaks. Also added basic completion for config items for the --config option.

Document password syntax in repositories URLs

Merge with crew-stable

mq: fix qrefresh losing copy information (issue 1134)

hgk: ctx.parents() problem introduced by 2d54e7c1e69d

Fix tests failures introduced by 7239e06e58e9

convert: add documentation for CVS source

merge another backout

Backed out changeset 4879468fa28f (incorrect Content-Length on Windows)

merge the backout head

Backed out changeset b9d6ab187523 (doesn't work on Python 2.3/2.4)

merge with main

merge with crew-stable

Check that git patches only touch files under root

Updating Copyright dates in Windows Installer

Force email package to be loaded in py2exe With Python 2.5, the email package is not fully loaded by py2exe, due to dynamic imports which are not found by modulefinder. This breaks the patchbomb extension. This patch forces the whole email package to be included so that the dynamic imports work as expected.

hgweb: use a single-element tuple to return from protocol.unbundle() Python 2.3/2.4 don't support yielding from inside try/finally blocks.

add an Accept header to the http client

tests: add a test for the hgweb graph

hgweb: return content iterator instead of using write() callable

hgweb: all protocol functions have become generators Using the write() callable supplied by the start_response() call is frowned upon by the WSGI spec, returning an iterable over the content chunks is the recommended way. Be aware, though: returning many small chunks will slow down responses, because the server has to flush each chunk separately.

templater: make a template a string-only iterator

streamclone yields chunks instead of accepting a file-like object

hgweb: protocol functions take repo instead of web This makes it much easier for implementers of hgweb alternatives to simply call into protocol functions after setting up a repo and a request.

hgweb: move capabilities calculation back into hgweb.protocol

hgweb: centralize permission checks for protocol commands Consistently enforces authorization checks set up in hgrc up front, so that the actual commands don't have to worry about them and implementers of hgweb alternatives can easily implement their own permission checks.

tests: add some tests for web.allowpull configurations

hgweb: protocol requests are processed immediately This makes separation of interface/webcommands and protocol easier and will make http repos slightly faster because there's less overhead.

merge with mpm