Mercurial 1.1 released!

Matt Mackall — 2008-12-02T21:55:16

This is a major release with numerous new features. Full details here: http://www.selenic.com/mercurial/wiki/index.cgi/WhatsNew Available for download at: http://www.selenic.com/mercurial/release/mercurial-1.1.tar.gz

Converting from git loses branch names

Greg Ward — 2008-12-02T20:07:06

I am trying to convert a git repository with several named branches. "hg convert" seems to convert all the changesets and preserve the branch/merge history. But it loses the branch names: all of my git branches disappear and become unnamed Mercurial heads on branch "default". (Hope I've got my terminology right there!) To make it concrete: I first made a local clone of my git repo. In that clone: $ git rev-list --all | wc -l 96 $ git branch [...17 branches including "master"...] Thus I expect 96 changesets and 17 branches (including "default") in my Mercurial conversion. After "hg convert"ing that clone, I get $ hg log --template "{node}\n"| wc -l 97 # extra changeset for converting tags $ hg branches default 96:f0ce268bb09c See... no branches! Is this a known problem? Or perhaps a confusing feature? Google'ing didn't turn up much except someone else noticing the same problem converting a git repo of KDE to hg: http://mail.kde.org/pipermail/kde-scm-interest/2008-February/000160.html Thanks -- Greg

clone by tag, but tag doesn't appear in the clone

Paul J. Morris — 2008-12-02T04:47:59

Hi, I'm cloning a repository by a specific tag, yet in the cloned repository "hg tags" doesn't list the tag that the repository was cloned from. e.g d:\build> hg clone -r spec_1_0 c:\work d:\build\work> hg tags tip 55:c5537aefcdd9 How do I get the tag details into the cloned repository? thanks Paul Morris

ANN nearest extension

Gilles Moris — 2008-12-02T08:54:02

Hello, Here is the nearest extension, which is much a complete rewrite of the describe extension. Its goal is to find the nearest tag(s) from a given changeset, either backward or forward in the changesets history tree. The general usage is: hg nearest [-c] [-a] [revs] Code can be found here: http://freehg.org/u/morisgi/nearest/file/33ef01aff84f/nearest.py Here are the specificity of this extension: - the revs are searched ordered by date, so that 1.0.2 will appear first, even if there is a shorter path to go to the 1.0.1 tag - the default output format is "tag+dist", and "tag-dist" with the --contains option, where "dist" is the longest path to the tag. The longest path better matches the efforts between the revision, as merges and backouts in particular, can create artificial shortcuts. - customizable output format with % characters - better performance, so that a --limit option is not (hopefully) needed - the --contains flag searches the history forward to answer the question: which release contains this given revision ? This is my original use case. - a shell pattern match is used instead of regexp, because it's simpler and because the dot "." character usually used in tags is not interpreted. regexp can be added later. - branch selection dropped as tag matching is probably sufficient to fill the use case. Regards. Gilles.

filtering changelog messages - potention extension for hgweb?

Jan Capek — 2008-12-01T21:14:17

Hi, I have come across an hgweb usecase where I needed a filtering extension: Use case -------- - a mercurial repository is being used with trac for issue tracking. The repository has hooks to check that changelog messages contain valid bug ID's (similar to bugzilla extension) - the bug ID's that show up in changelog messages should be hyperlinks to the bug tracking system when using the hgweb interface Implementation -------------- To implement the above scenario a new filter extension has been created. This extension is configurable via hgrc ([urllnk] section). I am attaching the extension itself and an example patch of the gitweb style to show its use. Would this extension be of any interest to mercurial community, is it possible to post it on the selenic.com wiki? Thanks, Jan Index: git-web-with-logo/summary.tmpl =================================================================== --- git-web-with-logo.orig/summary.tmpl2008-11-27 13:20:32.000000000 +0100 +++ git-web-with-logo/summary.tmpl2008-11-27 16:28:39.000000000 +0100 < at >< at > -29,7 +29,7 < at >< at >

- +

description	#desc#
description	#desc\|urllnk#
owner	#owner\|obfuscate#
last change	#lastchange\|rfc822date#

Index: git-web-with-logo/changeset.tmpl =================================================================== --- git-web-with-logo.orig/changeset.tmpl2008-11-27 13:20:32.000000000 +0100 +++ git-web-with-logo/changeset.tmpl2008-11-27 16:28:39.000000000 +0100 < at >< at > -29,7 +29,7 < at >< at >

Index: git-web-with-logo/changelogentry.tmpl =================================================================== --- git-web-with-logo.orig/changelogentry.tmpl2008-11-27 13:20:32.000000000 +0100 +++ git-web-with-logo/changelogentry.tmpl2008-11-27 16:28:39.000000000 +0100 < at >< at > -8,7 +8,7 < at >< at > #author|obfuscate# [#date|rfc822date#] rev #rev#

Index: git-web-with-logo/fileannotate.tmpl =================================================================== --- git-web-with-logo.orig/fileannotate.tmpl2008-11-27 13:20:32.000000000 +0100 +++ git-web-with-logo/fileannotate.tmpl2008-11-27 16:28:39.000000000 +0100 < at >< at > -48,7 +48,7 < at >< at >

-{desc|strip|escape|addbreaks} +{desc|strip|escape|addbreaks|urllnk}

from mercurial import templatefilters import re class UrlLnkFilter(object): """This Filter allows replacing user tags with URL references User may define the 'base URL' and the 'regular expression' to match a user tag and the 'replacement expression'. None of these need to be defined as there are defaults values The filter can be used for the hgweb templates to generate links into a bug tracking system. The basic idea is to modify the templates like this: #desc|strip|escape|addbreaks# #desc|strip|escape|addbreaks|urllnk# By default the filter detects any text referencing tags prefixed with #,ticket,issue or bug followed by a number. This is replaced by original text. Example: Input text: ----------- Changelog entry references #3 Output text ----------- Changelog entry references #3 < at >var base_url - denotes the base URL to be prepended to generated link. Defaults to empty string. < at >var utag_regexp - regular expression to match the user tag. < at >var utag_replace - replacement expression """ def __init__(self, ui, repo): """Initializes regular expressions to match user tags The defaults can be overiden by the mercurial configuration < at >param self < at >param ui - user interface of the mercurial < at >param repo - current mercurial repository """ conf_section = 'urllnkfilter' # base url used for anchor self.base_url = ui.config(conf_section, 'base_url', '') # for better efficiency, precompile the regular expression txt_regexp = ui.config(conf_section, 'ticket_regexp', '(\s+)((?:#|(?:ticket|issue|bug)[: ]?)([0-9]+))') self.utag_regexp = re.compile(txt_regexp) # replacement string self.utag_replace = ui.config(conf_section, 'ticket_replace', '\g<1>\g<2>' % self.base_url) def __call__(self, text): """The object implements a callable interface to perform filtering. < at >param self """ return re.sub(self.utag_regexp, self.utag_replace, text) def reposetup(ui, repo): templatefilters.filters["urllnk"] = UrlLnkFilter(ui, repo)

Mercurial Eclipse 1.2 released

Bastian Doetsch — 2008-12-01T22:07:26

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, fyi, we've just unleashed Mercurial Eclipse 1.2.941. You can find the details at http://groups.google.com/group/mercurialeclipse/browse_thread/thread/afeb72d7be325848# Cheers, Bastian -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.8 (Darwin) iEYEARECAAYFAkk0YB4ACgkQBqJW4lDO0YJi3gCeJKTZpWvuyckbkMn7bochHDAz il0AnjcopvbhYn2BK4l4QpiyajAiYErW =lVWU -----END PGP SIGNATURE-----

Feature request: mark "dead ends"

Hans Meine — 2008-12-01T14:52:56

Hi again, in one of my repositories I have checked in some changeset X for feature A which I could never made to work. Later, I have restarted from X's parent and some dozens of revisions later I finished the same feature A in another way. Now I have this extra head with the unfinished changes which I don't need anymore. This extra head gets in my way when using "hgview" (it always appears at the very top, although it is very old). AFAICS, I have three options: 1) Use a dummy merge to make the unnamed branch inactive. That would look like I actually used code from my first try, which I did not. 2) Use strip/mq/convert etc. to remove this head from my repo. I have no intention to do this. 3) Leave it as it is, and try to fix hgview for example. I would like to have option 4) - mark the branch as inactive without merging it. Apparently, this is not yet possible with hg, right? What do you people think? (I bet most people are content with opion 1, but what do you think about 4?) Have a nice day, Hans

[PATCH] tests: allow cvs import to reorder its filename list

Frank Kingswood — 2008-12-01T14:24:18

# HG changeset patch # User Frank Kingswood kingswood-consulting.co.uk> # Date 1228138924 0 # Node ID b1c1b4f98b74bfa32991d33bdb4a47215cdeb842 # Parent 79d1bb737c16b343e33be7cd7e22dcc4e7bfab0b tests: allow cvs import to reorder its filename list also fix script to generate non-zero length test file b. diff -r 79d1bb737c16 -r b1c1b4f98b74 tests/test-convert-cvs-branch --- a/tests/test-convert-cvs-branchMon Dec 01 14:20:20 2008 +0100 +++ b/tests/test-convert-cvs-branchMon Dec 01 13:42:04 2008 +0000 < at >< at > -28,8 +28,9 < at >< at > mkdir src cd src -echo "1" > a > b -cvscall import -m "init" src v0 r0 +echo "1" > a +echo "1" > b +cvscall import -m "init" src v0 r0 | sort cd .. cvscall co src cd src diff -r 79d1bb737c16 -r b1c1b4f98b74 tests/test-convert-cvs-branch.out --- a/tests/test-convert-cvs-branch.outMon Dec 01 14:20:20 2008 +0100 +++ b/tests/test-convert-cvs-branch.outMon Dec 01 13:42:04 2008 +0000 < at >< at > -1,10 +1,10 < at >< at > % create cvs repository % Create a new project + + N src/a N src/b - No conflicts created by this import - cvs checkout: Updating src U src/a U src/b

Please, some "partial-silency" option for pull

Marcin Kasperski — 2008-12-01T11:45:30

It would be nice if hg pull could be configured to suppress output when there are no changes transferred. Rationale: I frequently use different "pull many repositories at once" scripts, from for dir in * do hg --cwd $dir pull done to more complicated "syncthemall" scripts. The output of such a command is difficult to read as truly useful information is lost among numerous pulling from (...) searching for changes no changes found messages. I'd prefer to see only info about directories where "something" happened. (at the same time hg -q pull supressess everything, also the useful info) Syntax suggestion: hg pull --silent Alternative syntax suggestion: It could be even better idea to introduce some global option of similar meaning (report "interesting" information but suppress notifications that nothing happened, nothing was found, no files were updated, nothing was added etc etc). This would establish a convention for different commands. It could be introduced as hg --quiet=partial command or sth similar. Again: it would be intended for use in scripts, cron jobs (like http://blog.mekk.waw.pl/archives/8-Mercurial-as-a-backup-tool-on-MoinMoin-example.html) etc etc

qdiff output with commit message and user info

Igor Bukanov — 2008-11-30T21:37:44

Hi, is there any way to include to hg qdiff output the commit message and the user name recorded during qrefresh? Alternatively can I force to include -p option to mq patches as stored under .hgrc/patches ? Regards, Igor

clone: source path can't contain '#'

TK Soh — 2008-11-30T02:50:46

A user of TortoiseHg has reported this, which I confirmed with recent crew version of Mercurial: $ hg -R crew# tip changeset: 6797:694223a29ad4 tag: tip user: byron< at >base2.cc date: Thu Jul 03 20:53:14 2008 -0700 summary: Added missing --config global option to zsh_completion. $ hg clone crew# c1 abort: repository crew not found!

email.charsets trouble

Giorgos Keramidas — 2008-11-29T23:03:57

Hi everyone, I recently had a chance to upgrade the Hg installation at hg.hellug.gr to a crew snapshot. One of the repositories hosted there is the common `translation tree' of the Greek FreeBSD documentation team at: http://hg.hellug.gr/freebsd/doc-el/ With older versions of Hg we would get commit email notifications whose included patches did not have a charset/encoding. The rest of the email was really encoded with email.charsets = iso-8859-7 though, so most mailers would correctly infer that the patch was iso-8859-7 too. With the recent Mercurial installation, patchbomb seems to be using utf-8 as the email charset, no matter what I try to use in the `.hg/hgrc' file of the relevant repository. Most of the patches we commit in the Greek documentation are encoded in iso-8859-7, so they show up as a long list of question marks in the emails sent by patchbomb, i.e.: % diff -r 913b74dbc96e -r 51e99cebb8bd el_GR.ISO8859-7/books/handbook/kernelconfig/chapter.sgml % --- a/el_GR.ISO8859-7/books/handbook/kernelconfig/chapter.sgml Sat Nov 29 16:11:31 2008 +0200 % +++ b/el_GR.ISO8859-7/books/handbook/kernelconfig/chapter.sgml Sat Nov 29 16:23:32 2008 +0200 % < at >< at > -253,9 +253,9 < at >< at > % ãéá áõôÞ ôç óõóêåõÞ áíáöÝñåé: % % Alternatively, to load the driver as a module at boot time, place the % - following line in &man.loader.conf.5: % +following line in &man.loader.conf.5: % % - if_ath_load="YES" % + if_ath_load="YES" % % ¼ðùò áíáöÝñåôáé óôéò ïäçãßåò, áí âÜëåôå ôçí ãñáììÞ % if_ath_load="YES" óôï áñ÷åßï This means that patch email is not really useful to reviewers. Is there any way to _force_ patchbomb to use a specific charset for the patch parts of the emails? I have tried setting: [email] charsets = iso-8859-7 but this always picks utf-8 for the outgoing emails.

Integrity problems on CIFS mounted file system

Seamor — 2008-11-28T17:06:11

Running the following script on the local file system works fine. But running it on a CIFS mounted file system causes integrity errors. #!/bin/bash hg init original cd original echo '' > NotRenamed.txt hg add hg commit -m 'NotRenamed' cd .. hg clone original clone cd clone cp NotRenamed.txt Renamed.txt hg remove NotRenamed.txt hg add hg commit -m 'Renamed' hg push hg verify cd ../original hg update hg verify Here is my fstab. If anybody has any suggestions I would be most thankful. //seabsrv01/dev$ /mnt/dev cifs uid=morgan,gid=morgan,credentials=/home/morgan/.smbcredentials,iocharset=utf8,dir_mode=0777,file_mode=0777 0 0

LOC of different version control systems

Gabor Grothendieck — 2008-11-28T18:06:27

According to: http://www.ohloh.net/projects/compare?metric=Codebase&project_0=Mercurial&project_1=Git&project_2=Bazaar+Version+Control Git, Bazaar and Mercurial have 350k, 150k and 50k lines of souce code respectively. Initially I made the above comparison between just Git and Mercurial and figured it might be accounted for by the fact that Git is in C and Mercurial is in Python but then I added Bazaar and it is also written in Python. Other possibilities might be how extensions are counted or just plain error in measurement. Does anyone understand the differences between these systems sufficiently to explain these seemingly large differences?

Merge commits do only show files changed during the merge as changed (i.e. those with conflicts or similar). Bug or feature?

2008-11-28T12:37:56

First a few steps to reproduce the problem. Then the question is at the end: ~/tmp$ hg init hgtest ~/tmp$ cd hgtest/ ~/tmp/hgtest$ echo file1 > file1.txt ~/tmp/hgtest$ echo file2 > file2.txt ~/tmp/hgtest$ hg add . adding file1.txt adding file2.txt ~/tmp/hgtest$ hg ci -m "Initial checkin" ~/tmp/hgtest$ ls >> file1.txt ~/tmp/hgtest$ hg ci -m "Change in file1" ~/tmp/hgtest$ hg update 0 1 files updated, 0 files merged, 0 files removed, 0 files unresolved ~/tmp/hgtest$ ls >> file2.txt ~/tmp/hgtest$ hg ci -m "Change in file2, create another head" created new head ~/tmp/hgtest$ hg merge -r1 1 files updated, 0 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit) ~/tmp/hgtest$ hg ci -m "Merge" ~/tmp/hgtest$ hg log -v -r3 changeset: 3:936aaa1c9975 tag: tip parent: 2:1dd8a64b18b0 parent: 1:741fd3bcb38d user: Jens Lund nordea.com> date: Fri Nov 28 12:31:52 2008 +0100 description: Merge What puzzles we is that the merge commit (3) show no files as changed even though file1.txt respective file2.txt is changed compared to one or the other parent. Only files changed during the merge will be shown. This I might understand in this context from the command line, but it has the side effect that in "hg view/hgk" or "hgtk log" in tortoisehg no differences are shown compared to the parents, and that I see as highly confusing because clearly "hg diff -r2 -r3" will show changes. Is the a bug or a feature in hg? In "hgk/hg view/hgtk log"? Best regards, Jens

Merging with Emacs and ediff-mode

Giorgos Keramidas — 2008-11-28T04:24:28

Hi everyone, I've been using two script-wrappers for merging for a while now: hg-merge This one tries Mercurial's pre-merge, then diff3 and falls back to merging in an ${EDITOR} buffer with conflict markers. ediff-merge This one disables pre-merge, it tries diff3 but if that fails too it starts with a clean copy of $local and fires up Emacs in ediff-mode to do a 3-way merge. The .hgrc options that enable these are: [merge-tools] ediff.priority = 0 ediff.premerge = False ediff.binary = False ediff.gui = True ediff.executable = /home/keramida/bin/ediff-merge ediff.args = $local $base $other script.priority = 1 script.premerge = True script.binary = False script.gui = False script.executable = /home/keramida/bin/hg-merge script.args = $local $base $other and I have in my ~/bin directory the attached scripts. Do these look ok as merge helpers for UNIX workstations? Should I add these to the Wiki?

Setting Up HG To Work With Winmerge - Windows Cmd Line Use

The Dude — 2008-11-27T07:37:24

Hello, I'm using Mercurial on Windows using the command line, and would like to ask how to set up merge to use the WinMerge utility. Specifically, the online documentation shows this for kdiff3: In Mercurial 1.0 you can add the following to .hgrc: [merge-tools] kdiff3.args = $base $local $other -o $output and I'd like to ask what .hgrc configuration is necessary for this. Many thanks! Bye, TD

Bug report (hg qpush)

Peter Weilbacher — 2008-11-27T07:29:54

Hi, I got the message below when running hg qpush gfx_debugging_stuff_2008-11-25.diff on a Mozilla repository. After this, the queue state looks like this $ hg qa 0 U gfx_debugging_stuff_2008-11-25.diff 1 A cairo-os2_fix_as_upstream.diff 2 A bug453705_domquickstubs.diff 3 A bug451278_nscom_workaround.diff The files that the patch is supposed to touch were not changed. I worked around this problem now by doing hg qpop -a before trying the qpush again... The messages don't look like a system specific problem, but this is with Mercurial running on OS/2. Cheers, Peter. ** report bug details to http://www.selenic.com/mercurial/bts ** or mercurial< at >selenic.com ** Mercurial Distributed SCM (version a490a0c59b94fef613ac87813e12b992421d0948) Traceback (most recent call last): File "f:/Python/Scripts/hg", line 22, in ? mercurial.dispatch.run() File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 20, in run sys.exit(dispatch(sys.argv[1:])) File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 29, in dispatch return _runcatch(u, args) File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 45, in _runcatch return _dispatch(ui, args) File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 364, in _dispatch ret = _runcommand(ui, options, cmd, d) File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 417, in _runcommand return checkargs() File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 373, in checkargs return cmdfunc() File "f:/Python/lib/site-packages/mercurial/dispatch.py", line 356, in d = lambda: func(ui, repo, *args, **cmdoptions) File "f:/Python/lib/site-packages/hgext/mq.py", line 1946, in push mergeq=mergeq) File "f:/Python/lib/site-packages/hgext/mq.py", line 813, in push end = self.series.index(patch, start) + 1 ValueError: list.index(x): x not in list

mercurial + nginx

allnightlong — 2008-11-26T23:40:27

rebasing only a single changeset

Benjamin Smedberg — 2008-11-26T15:27:30

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Mozilla is going to cut a long-lived release branch shortly. We will be cherry-picking individual changes from the trunk to the release branch. I originally thought that the rebase extension would be perfect for this task: hg rebase -s revtocherrypick -t releasebranch But I forgot that rebase will not rebase the single revision I'm interested in: it will only rebase entire branches. The transplant extension could be used for this, but doesn't do 3-way merging (and the rebase extension is in general better designed). How hard would it be to add "rebase a single revision" to the rebase extension? - --BDS -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFJLWriSSwGp5sTYNkRApc8AKDFSkO6OdFLtNVljy7iOQOYRY9xjgCeL8y0 o4zeG35ugDZneposLEYhTz8= =JbZh -----END PGP SIGNATURE-----

Call for testing: mercurial 1.1

Benoit Boissinot — 2008-11-26T11:29:26

Since the 1.1 release is approaching, it would be nice to have some additional testing, preferably using crew repo: http://hg.intevation.org/mercurial/crew Thanks! Benoit

gmane.comp.version-control.mercurial.general