hgweb: don't send a body or illegal headers during 304 response Without this fix, mod_wsgi and spawning get in a wedged state after sending a 304 response. Not sending a body fixed that problem. The header change was discovered by using wsgiref.validate.validator to check for other errors.

merge with stable

test-url: skip test when ssl module is unavailable

revset: use 'requires' instead of 'wants' in error message

outgoing: unify common graphlog.outgoing and hg.outgoing code

bundlerepo: unify common code into a new getremotechanges The pattern where we fetch incoming remote changes and return them as a local bundlerepo seems common. It's nicer to have this code unified.

revset id(): fix error text to say "id wants..." instead of "rev wants..."

keyword: only use expensive fctx.cmp when needed Restrict expensive cmp to cases when: - comparing against working directory and - encode filters active or - path is configured for keyword expansion

context: narrow down filter special case in filectx.cmp() cmp via filelog when encode filters are present, but only when actually comparing to the working directory.

incoming: unify code for incoming and graphlog.incoming

incoming: rename variable Groundwork so that incoming and graphlog.incoming code look the same and get easier to unify.

import: only the first hg patch marker should be processed (issue2417) Proper use of the hgpatch state variable had been lost in the final edits of d7452292f9d3 - now it works more like intended.

backout: provide linear backout as a default (without --merge option) This changes backouts changeset to retain linear history, .e. it is committed as a child of the working directory parent, not the reverted changeset parent. The default behavior was previously to just commit a reverted change as a child of the backed out changeset - thus creating a new head. Most of the time, you would use the --merge option, as it does not make sense to keep this dangling head as is. The previous behavior could be obtained by using 'hg update --clean .' after a 'hg backout --merge'. The --merge option itself is not affected by this change. There is also still an autocommit of the backout if a merge is not needed, i.e. in case the backout is the parent of the working directory. Previously we had (pwd = parent of the working directory): pwd older backout auto merge backout --merge auto commit With the new linear approach: pwd older backout auto commit backout --merge auto commit auto: commit done by the backout command merge: backout also already committed but explicit merge and commit needed commit: user need to commit the update/merge

update: use revsingle to enable use of revsets as update targets (issue1993)

test-url: skip test when ssl module is unavailable

test-url: remove trailing whitespace

keyword: code cleanup - move preselection of expansion candidates for rollback and record into helper function - same overwrite order in rollback and record: 1. modified, 2. added - self.wlock() inside kwrepo class instead of repo.wlock()

tests: asciify output of test-encoding-align.t

tests: accept \-escaped test output It changes tsttest to accept expected outputs in python-style \-escapes. It aims to avoid trouble with outputs for non-ascii, color and progress tests.

merge with crew

parentdelta: compress.py extension is no longer necessary Using `hg --config format.parentdelta=True clone --pull src dst` is equivalent.

merge with stable

convert/darcs: support changelogs with bytes 0x7F-0xFF (issue2411) This is a followup to 4481f8a93c7a, which only fixed the conversion of patches with UTF-8 metadata. This patch allows a changelog to have any bytes with values 0x7F-0xFF. It parses the XML changelog as Latin-1 and uses converter_source.recode() to decode the data as UTF-8/Latin-1. Caveats: - Since the convert extension doesn't provide any way to specify the source encoding, users are still limited to UTF-8 and Latin-1. - etree will still complain if the changelog has bytes with values 0x00-0x19. XML only allows printable characters.

revset: add id() and rev() to allow explicitly referring to changes by hash or rev

revset: rename tagged() to tag() and allow it to take an optional tag name

bookmarks: add revset for referencing bookmarks

hgweb help: fix double help links

tests: add testcase for --config format.dotencode=false

fetch: fix and document exit codes (issue2356)

doc: add [format] dotencode description

filectx: use ctx.size comparisons to speed up ctx.cmp Comparing sizes is cheaper than comparing file contents, as it does not involve reading the file on disk or from the filelog. It is however not always possible: some extensions, or encode filters, change data when extracting it to the working directory.

localrepo: use propertycaches to access encode/decode filters

localrepo: load filter patterns outside of _filter

localrepo: have _loadfilter return the loaded filter patterns

merge with mpm

wireproto/http: drain the incoming bundle in case of errors

wireproto: introduce pusherr() to deal with "unsynced changes" error The behaviour between http and ssh still differ: - the "unsynced changes" is seen as a remote output in the http cases - but it is correctly seen as a push error for ssh

wireproto: redirect the output earlier

wireproto: return in finally was messing with the return inside the block

tests: use a glob for all of the host, might not be localhost

commands: add move alias for the rename command svn provides move, and it's also a standard tool on Windows.

util: make wrap() require a width argument This keeps hgweb's help engine from poking at file descriptors that don't exist.

showconfig: don't accept multiple sections and one config item Showconfig now behaves as documented and only accepts one section.name argument or a number of section names.

hgweb: another fix for the help termwidth bug

pager: don't run pager if nothing is written to stdout/stderr This decides when to run the pager based on the first call to ui.write() and ui.write_err(). This has the side effect of not the output of subprocesses that write output before hg does.

pager: add global --pager=<auto/boolean> option

color: add global option in extsetup() instead of globally This prevents side effects from occurring when importing the color extension.

hgweb: more ui fiddling

hgweb: fix hgweb_mod as well as hgwebdir_mod

hgwebdir_mod: use fake termwidth to avoid mod_wsgi's wrath

termwidth: move to ui.ui from util

help: improve description of update --check

store: encode first period or space in filenames (issue1713) - Mac OS X has problems with filenames starting with '._' (e.g. '.FOO' -> '._f_o_o' is now encoded as '~2e_f_o_o') - Explorer of Windows Vista and Windows 7 strip leading spaces of path elements of filenames when copying trees Above problems are avoided by encoding the first space (as '~20') or period (as '~2e') of all path elements. This introduces a new entry 'dotencode' in .hg/requires, that is, a new repository filename layout (inside .hg/store). Newly created repositories require 'dotencode' by default. Specifying [format] dotencode = False in a config file will use the old format instead. Prior Mercurial versions will abort with the message abort: requirement 'dotencode' not supported! when trying to access a local repository that requires 'dotencode'. New 'dotencode' repositories can be converted to the previous repository format with hg --config format.dotencode=0 clone --pull repoA repoB

demandimport: fix an obscure corner-case. Python's __import__() function has 'level' as the fourth argument, not the third. The code path in question probably never worked. (This was seen trying to run Mercurial in PyPy. Fixing this made it die somewhere else...)

keyword: enforce subn method via boolean switch There are only 2 patterns to choose, and so far only 1 case where kwtemplater.re_kw.subn is applied on data read from the working directory: when recording added files. With this change the code reflects more closely the boolean character of the switch and underlines the special case.

keyword: fix weeding of expansion candidates when recording Rearrange tests to check this, i.e. that there are changes in other files, not only the recorded one.

copies: don't detect copies as "divergent renames" (For the purposes of this patch copy is defined as a rename where the source continues to exist.)

strip: add --keep flag to avoid modifying wc during strip Fixes issue1564.

update: use higher level wording for "crosses branches" error When using "hg update" to update to a revision on another branch, if the user has uncommitted changes in the working directory, hg aborts with the following message: abort: crosses branches (use 'hg merge' to merge or use 'hg update -C' to discard changes) If the user isn't trying to update to tip and they follow the command examples verbatim, they would end up updating to the wrong revision. This patch removes the command examples in favor of just telling the user to either merge or use --clean: abort: crosses branches (merge branches or use --clean to discard changes) hg also aborts if the user tries to use "hg update" to get to tip (without specifying a revision) and tip is on another branch: abort: crosses branches (use 'hg merge' or use 'hg update -c') This message is changed in the same fashion: abort: crosses branches (merge branches or use --check to force update)

hgweb: add help link to templates missed in ead4e21f49f1

test-convert-svn-encoding.t: ignore that subversion %-encodes $TESTTMP Use a glob instead of expecting $TESTTMP.

run-tests.py: remove support for .bat files

run-tests.py: do not install hg when the tests do no exist

setup.py: write libdir as a python string literal

patch: fails immediately upon malformed hunk patch(1) does silently ignore malformed hunks but this is not something we want to copy.

patch: always raise PatchError with a message, simplify handling

patch: simplify externalpatch() arguments

util: remove needbinary(), no longer used for external patching

patch: remove internal patcher fallback and NoHunk error By now the internal patcher is probably more reliable than anything we might find on the command line.

patch: remove unused patchmeta.lineno

patch: remove unused flags from readgitpatch()

test-subrepo-svn.t: ignore that subversion %-encodes $TESTTMP Use a glob instead of expecting $TESTTMP.

revsets: fix stray * in help topic

web: add a help view for getting hg help output

fix indentation error

merge: handle no file parent in backwards merge (issue2364)

merge with crew

ui.paths: expand paths directly in fixconfig (issue2373) var and home expansion should be done first.

setup/hg: always load Mercurial from where it was installed. This provides two new features: - Mercurial may be installed into a non-standard location without having to set PYTHONPATH. - Multiple installations can use Mercurial from different locations.

revsets: add a sample query to the help for getting active branches

revsets: add descriptions for sample queries in the help

mq: silence spurious output. When using a versioned patch repository, you would get a spurious warning when deleting and adding the same patch. Before: $ hg qdelete --keep 3.diff $ hg qimport --existing 3.diff adding 3.diff to series file 3.diff already tracked! After: $ hg qdelete --keep 3.diff $ hg qimport --existing 3.diff adding 3.diff to series file

revsets: clarify in the help that head() returns named branch heads

tests: add testcase for 208fc9ad6a48 to test-alias.t The testsuite lacks a testcase for the bug introduced in 208fc9ad6a48. This patch amends 301d7626e0ff (which fixed 208fc9ad6a48) by adding a testcase for that bug. With 208fc9ad6a48, test-alias.t (as modified by this patch) fails with "hg tglog: invalid arguments".

mq: handle deleting the same patch twice in one command (issue2427)

progress: make sure stderr has isatty before calling (issue2191)

windows: handle spaces in path to Python (issue2074)

minirst: pull admonition titles out formatblock function

minirst: refactor/simplify findblocks

inotify: raise correct error if server is already started in a deep repository When path is too long to be an Unix socket address, we create a socket in a temporary directory and link from the long path to the shorter one. But checks in server code at startup were insufficient in this case, and used to raise an unclear "tried linking .hg/inotify.sock to a temporary socket but .hg/inotify.sock already exists"

setup: user-friendly error message if Python headers are missing

test-keyword: ignore subject in notify hook mails Long tmpdir names caused truncation of subject anyway, and that made $TESTTMP replacement fail.

test-keyword: fix test glob to ACL/SELinux flag

test-notify: fix fix for line continuation in long mail header lines Continued lines can also start with space. Note that this hack also incorrectly patches the diff output.

import: don't strip '#' lines from patch descriptions (issue 2417) Previously no '# ' lines came through the parser. Now only the first '# ' lines are processed, from '# HG changeset patch' and to the first line not starting with '# '.

test-notify: stabilize output test-notify is made a bit more stable by stripping "all" folders away instead of just 3.