.. -*- coding: utf-8 -*- Project cubicweb-vcsfile ======================== :creation date: 2008/12/23 component to integrate version control systems data into the CubicWeb framework Summary ------- This cube stores the data found in a version content manager repository. It currently works with subversion or mercurial repository. This cube doesn't access to the data directly. It depends on the logilab.devtools.vcslib library. All the repository's metadata (eg revision, file names, commit message, date and autor, etc) are stored as entities, and thus queryable via RQL, while actual files content is kept in the repository and fetched from there on demand. README ------ .. vcsinclude:: README Version 2.4.2 ------------- :publication date: 2017/09/11 :expected date: n/a Ticket #15120115 Revision creation_date set at revision **local** date ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending but it's considered as UTC (creation_date is TZDatetime). The root cause is probably https://bz.mercurial-scm.org/show_bug.cgi?id=5356 Until this is fixed, we can hack something with tzlocal? Version 2.4.1 ------------- :publication date: 2016/08/24 :expected date: n/a Ticket #14921605 missing dependancy on mercurial evolve extension ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending though it's not properly packaged, so it's not that easy Version 2.4.0 ------------- :publication date: 2015/12/18 :expected date: n/a Ticket #9294631 set created_by on revisions created from the UI ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.250 :state: validation pending Repository.make_revision should set the created_by relation to the current user. Version 2.3.0 ------------- :publication date: 2015/12/08 :expected date: n/a Ticket #8984565 Add a tox config ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending Add a tox config Ticket #8984467 "entities" package not included in sdist ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending "entities" package not included in sdist Ticket #6354109 [sobjects] Add a diff service ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending This will be used in vcreview to display the diff between revisions of a given patch. This can also be used within vcsfile in place of export-rev service, since we don't really need an exported patch (with header etc.). Version 2.2.0 ------------- :publication date: 2015/10/15 :expected date: n/a Ticket #6352111 [entities] enable hg evolve extension ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending [entities] enable hg evolve extension We can't use the allprecursors() revset without it. Ticket #5962570 [hgrepo] don't pass a revision to hg clone ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending [hgrepo] don't pass a revision to hg clone It prevents hardlinking the repo's store to our cache, which makes the clone *way* slower on a big repo. Version 2.1.2 ------------- :publication date: 2015/07/24 :expected date: n/a Version 2.1.1 ------------- :publication date: 2015/07/20 :expected date: n/a Ticket #5706735 2.0 migration shouldn't be quite so aggressive ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending When dropping the VersionContent/DeletedVersionContent/VersionedFile entity types, we should make sure they're no longer in use by another cube. Version 2.1.0 ------------- :publication date: 2015/07/20 :expected date: n/a Ticket #5270275 VCSInclude directive should catch hglib errors ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending if the local repo is not uptodate (or not there) we shouldn't get a traceback into the log, an error in the rendered html is enough Ticket #5270355 VCSInclude noise in log ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending :: 2015-04-21 15:40:48 - (cubicweb.entity) WARNING: Could not get included file at revision "None", using latest revision instead that message is not useful: - i don't know which file/repo it's talking about - i don't control the content of every rest field on the site so it should be toned down or removed Ticket #5704344 VCSRepo.head may raise CommandError ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending for instance if you call it by specifying both a path and a branch while the path doesn't exist in the branch Ticket #5102930 creating Repository entity from datafeed fails ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending the hooks don't expect entity.cw_edited['local_cache'] == None, so we never set local_cache and things get upset. Ticket #5102904 truncate author field ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending When importing changesets, the author field may be large (more than 200 chars). There is a constraint that limits the author field up to 200 chars. This is a soft constraint. The author field shall be cut up to 200 chars (``cubicweb.utils.cut`` ?) instead of raising a validation error. Comments :: On 2015/03/19 11:10 - acampeas wrote : Why not entirely drop this size constraint ? > On 2015/03/19 11:34 - jcristau wrote : > It makes some sense to have a limit, IMO. But it shouldn't prevent importing > existing revisions. Version 2.0.3 ------------- :publication date: 2015/02/02 :expected date: n/a Ticket #4873159 hg crashes with big revset ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending Ticket #4873158 importing repos has become very slow ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending too many calls to hglib's log method Ticket #4873157 [migration] removing entire entity types takes forever ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending we shouldn't use drop_entity_type for something this massive Ticket #4873156 strip/renumber detection broken ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending broken again in 2.0 Ticket #4873154 typo in 2.0.0 migration script ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending Version 2.0.2 ------------- :publication date: 2015/01/13 :expected date: n/a Ticket #4845397 the vcsinclude directive is broken ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending when the repository is not explicitly set, the expected entity attribute is `repository`, not `vcsrepository`... Version 2.0.1 ------------- :publication date: 2015/01/12 :expected date: n/a Ticket #4819618 migration path for repos with path and no source_url ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending we should not just drop the path Version 2.0.0 ------------- drop svn support and use hglib everywhere :publication date: 2014/12/22 :expected date: n/a Ticket #154069 branches support for svn ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.000 :state: rejected Ticket #154143 renaming support for svn ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.000 :state: rejected Ticket #154136 tags support for svn ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.000 :state: rejected Ticket #4439400 Add a vcsinclude rest directive ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: task :state: validation pending That includes the content of a file from a Repository, like:: .. vcsinclude:: README :repository: :revision: Ticket #2794142 clarify local_cache handling wrt path value ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: validation pending As of today, a repo with a source_url will get a "local cache" clone, but a repo with a "path" (reachable through the filesystem) has none. The original use case for "path" => no local_cache may have been a misguided attempt at disk space optimisation. I believe the local_cache should always be there. Comments :: On 2013/04/05 12:52 - jcristau wrote : why? what would be the point of a local_cache of a repo that is directly accessible locally anyway? > On 2013/04/05 15:30 - acampeas wrote : > unsure > > the point is that we have some code that relies on having a local_cache > > now, what about ensuring local_cache = path in this case ? would this be > acceptable ? Ticket #2591192 drop subversion support ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: validation pending Comments :: On 2013/01/28 12:36 - sthenault wrote : we should ensure no customer rely on it first... > On 2013/01/28 12:43 - acampeas wrote : > AFAIK no Logilab customer uses it. Ticket #3915240 hide obsolete/hidden heads in description tab (repo primary view) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending Any repo that uses mutable history will likely have hundreds of hidden heads compared to actual visible heads (usually only a handful). Ticket #2897254 text index Revision.changeset ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.250 :state: validation pending So it becomes searchable. Ticket #2807382 transform repo.VCSRepository into a proper adapter ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.400 :state: validation pending That would help simplify many things and reduce the pain of some. Ticket #2464887 VersionedFile and Revision should be related to Repository through different rtypes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.000 :state: deprecated we currently read: .. sourcecode:: python class from_repository(RelationDefinition): subject = ('Revision', 'VersionedFile') object = 'Repository' Why would one want that ? I propose a new `revision` rtype. Comments :: On 2013/03/11 10:38 - pydavid wrote : What's are the concret issues with the current approach? > On 2013/03/11 10:57 - acampeas wrote : > This is useless polymorphism: too often are we forced to actually specify > wether we want Revisions or VersionedFiles, because, you know, it is probably > ver exceptional that we are interrested by both at the same time. > > To be even more specific: > > * it makes repo.related('from_repository') useless * it forces us to write > plain RQL with "Any X WHERE X from_repository R, X is Revision": just have a > look at the tests Version 1.17.0 -------------- Proper support 3.19 :publication date: 2014/11/13 :expected date: n/a Ticket #4294387 Compatibility with cw 3.19 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: task :load: 0.500 :state: validation pending Ticket #4294386 Compatibility with mercurial 3.1 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: task :load: 0.500 :state: validation pending Ticket #4294710 Fix primary view of files containing HTML code ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending The primary view tries to render the content of VersionnedFiles. For HTML files, the current behavior is to "render" its content directly inline. This cannot ever work correctly: body tag inside a body tag, CSS classes that may have nothing in common with CubicWeb's own CSS, etc. The better option would be to call pygments on the content to get nice syntax highlighting. Comments :: On 2014/08/29 16:03 - acampeas wrote : For the record, the "file" cubes is supposed to handle this properly. Ticket #4578973 strip/renumber detection broken ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: not validated Ticket #4294388 Support parsing og hg versions like 3.1-rc+foo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending Version 1.16.0 -------------- :publication date: 2014/07/08 :expected date: n/a Ticket #3933063 rebuilding of a hg repo does not properly handle revision 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: validation pending Ticket #3924727 Repository.heads_rset() should not return hidden changesets ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending Ticket #3845451 bridge.import_content crashes with mercurial 3.0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending Relevant traceback: :: self.refresh() File "/home/dlaxalde/src/simulagora/cubes/mercurial_server/testutils.py", line 84, in refresh bridge.import_content(session, raise_on_error=True) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/bridge.py", line 93, in import_content raise_on_error=raise_on_error) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/bridge.py", line 100, in import_vcsrepo_content repohdlr.import_content(vcsrepo, commitevery) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/repohg.py", line 271, in import_content del missing[:commitevery] # pop iterated value TypeError: 'orderedlazyset' object does not support item deletion no stdout no stderr Version 1.15.0 -------------- :publication date: 2014/05/13 :expected date: n/a Ticket #3114960 cw reload mecanism bug workaround removal hurts ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending :: 2013-09-09 12:13:22 - (cubicweb.repository) ERROR: Unhandled exception in LoopTask refresh_vcsrepos Traceback (most recent call last): File "/home/auc/confs/forges/cubicweb/server/utils.py", line 157, in auto_restart_func func(*args) File "/home/auc/confs/forges/cubes/vcsfile/hooks.py", line 127, in refresh_vcsrepos return _refresh_repos(self.repo) TypeError: 'NoneType' object is not callable I suspect someone did a "cleanup" of that code not too long ago. Ticket #3762947 debug reload in CW 3.19 breaks 'refresh repo' looping task ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending For some mind-boggling reason, CW 3.19 breaks the looping task after a reload (when in debug mode). Python's ``warn`` module - the one raising the exception at the heart of the bug - being partly written in C, it's quite difficult to precisely pinpoint the location of the bug. The solution (and hopefully the correct one after 3 or 4 failed attempts) is to call ``import`` in the looping task's callback to make sure that we're always calling the latest ``_refresh_repo`` function from ``hooks.py`` even after a reload. Ticket #3770419 "refresh repository" action is broken ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending Here's the traceback:: 2014-04-23 18:11:15 - (cubicweb.session) INFO: calling service refresh_repository 2014-04-23 18:11:15 - (cubicweb.web) ERROR: TypeError("call() got an unexpected keyword argument 'async'",) Traceback (most recent call last): File "/home/rcardona/Projects/grshell-industree/cubicweb/web/application.py", line 453, in core_handle result = controller.publish(rset=rset) File "/home/rcardona/Projects/grshell-industree/cubes/vcsfile/views/__init__.py", line 30, in publish eids=(vcsrepo.eid,)) File "/home/rcardona/Projects/grshell-industree/cubicweb/repoapi.py", line 120, in proxy return getattr(clt_cnx._cnx, name)(*args, **kwargs) File "/home/rcardona/Projects/grshell-industree/cubicweb/server/session.py", line 372, in wrapper return func(cnx, *args, **kwargs) File "/home/rcardona/Projects/grshell-industree/cubicweb/server/session.py", line 382, in check_open return func(cnx, *args, **kwargs) File "/home/rcardona/Projects/grshell-industree/cubicweb/server/session.py", line 1154, in call_service result = service.call(**kwargs) TypeError: call() got an unexpected keyword argument 'async' This is due to CubicWeb 3.19 removing the ``async`` parameter. Ticket #3579417 revision should not be updatable ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending It does not make much sense to allow a revision to be modified as it is in practice imported from the underlying vcs. Ticket #3662200 hg 2.9 compat ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending File "/home/jcr/src/cw/cubes/vcsfile/repohg.py", line 518, in precommit self.hgparent = p1 = self.repo.branchtags()[branch] AttributeError: 'evolvingrepo' object has no attribute 'branchtags' Version 1.14.0 -------------- :publication date: 2013/11/28 :expected date: n/a Ticket #2906673 provide a way to get the latest public changeset in a repo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: validation pending e.g. repo.branch_head('default', public_only=True) Sometimes we want to know which is the latest public changeset, ignoring drafts. Ticket #3331387 vcsfile-notifbroker is silent ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: validation pending would be better to have some logging when it gets events and (more importantly) when it crashes Version 1.13.3 -------------- :publication date: 2013/09/26 :expected date: n/a Ticket #3027232 15s sleep in HGRepository.pull_or_clone_cache is expensive ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending In particular in tests. A better solution to https://www.cubicweb.org/ticket/2818315 should be found. Ticket #3168250 don't disable integrity hooks when importing a repo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: validation pending Ticket #2827080 Repository.localcachepath should consider .path ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending Since we firmly assume the .local_cache and .path attributes are mutually exclusive, the .localcachepath property should just use either one. Ticket #3031696 TypeError upon invocation of svn.core.svn_stream_read from ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending E.g. running unittest_write expecting an integer for the buffer size Traceback (most recent call last): File "/home/dlaxalde/src/simulagora/cubes/vcsfile/bridge.py", line 258, in callback return vcsrepohdlr.file_content(dir, name, revnum) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/repo.py", line 97, in file_content return self._file_content(path, rev) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/reposvn.py", line 97, in _file_content return _read_file_content(root, path) File "/home/dlaxalde/src/simulagora/cubes/vcsfile/reposvn.py", line 28, in _read_file_content data = core.svn_stream_read(stream, core.SVN_STREAM_CHUNK_SIZE) File "/usr/lib/pymodules/python2.7/libsvn/core.py", line 4801, in svn_stream_read return _core.svn_stream_read(*args) TypeError: expecting an integer for the buffer size Comments :: On 2013/07/22 08:59 - jcristau wrote : I would argue that's not our bug. If svn wants an int, it should define SVN_STREAM_CHUNK_SIZE as an int. On 2013/07/23 07:32 - dlaxalde wrote : Similar issues: - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=676176 - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=707669 Ticket #3034763 VersionContent's "repository content at this revision" action's URL is wrong ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: validation pending Its rql parameter contains the branch name without the needed surrounding quotes. Version 1.13.2 -------------- :publication date: 2013/06/13 :expected date: n/a Ticket #2924045 hgrc-path option is unreliable ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending it sets HGRCPATH in the looping task, which may be too late if an access to mercurial happens somewhere else. Ticket #2924046 Revision.export() crash if repo is not cached ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending File "/usr/share/cubicweb/cubes/vcsfile/entities.py", line 249, in export nodeid=self.changeset) File "/usr/lib/python2.6/dist-packages/cubicweb/dbapi.py", line 371, in call_service return self.cnx.call_service(regid, async, **kwargs) File "/usr/lib/python2.6/dist-packages/cubicweb/dbapi.py", line 543, in decorator return func(self, *args, **kwargs) File "/usr/lib/python2.6/dist-packages/cubicweb/dbapi.py", line 605, in call_service return self._repo.call_service(self.sessionid, regid, async, **kwargs) File "/usr/lib/python2.6/dist-packages/cubicweb/server/repository.py", line 886, in call_service return self._call_service_with_session(session, regid, async, **kwargs) File "/usr/lib/python2.6/dist-packages/cubicweb/server/repository.py", line 904, in _call_service_with_session return service.call(**kwargs) File "/usr/share/cubicweb/cubes/vcsfile/sobjects.py", line 28, in call hgrepo = hdrepo.hgrepo() File "/usr/share/cubicweb/cubes/vcsfile/repohg.py", line 446, in hgrepo repo = hg.repository(ui.ui(), self.path) File "/usr/lib/python2.6/dist-packages/mercurial/hg.py", line 111, in repository peer = _peerorrepo(ui, path, create) File "/usr/lib/python2.6/dist-packages/mercurial/hg.py", line 101, in _peerorrepo obj = _peerlookup(path).instance(ui, path, create) File "/usr/lib/python2.6/dist-packages/mercurial/localrepo.py", line 2591, in instance return localrepository(ui, util.urllocalpath(path), create) File "/usr/lib/python2.6/dist-packages/mercurial/localrepo.py", line 199, in __init__ raise error.RepoError(_("repository %s not found") % path) Version 1.13.1 -------------- :publication date: 2013/05/27 :expected date: n/a Ticket #2905618 typos in hooks.py ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending There are typos in VCSRepoNotifHook Version 1.13.0 -------------- :expected date: n/a Ticket #2897197 Add a notification machinery to make it possible to push mercurial repo updates instead of polling ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 1.000 :state: validation pending Ticket #2057810 More explicit error logging ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: validation pending I get a bunch of logs that say :: 2011-10-26 14:27:24 - (cubicweb.sources.hg) ERROR: 67457.url: It would be nice to include the URL in the error message so as to diagnose the problem Comments :: On 2013/03/11 10:36 - pydavid wrote : It would be nice if the ticket had a bit more details. > On 2013/05/23 10:17 - ddouard wrote : > I find it quite explicite as is. Ticket #2748283 remove optional "tags" cube dependency ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: validation pending It currently tags per directory component of each versioned file. Ticket #2894920 use revision short description as link title ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.250 :state: validation pending so there is an explicite tooltip Version 1.12.3 -------------- :expected date: n/a Ticket #2822693 diff2htmltransform depends on vcreview for its css ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending Ticket #2818315 hg changesets can be pulled without corresponding obsolete markers ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending hg pushes changesets and markers in separate transactions, so it's possible a pull would get the changesets without the markers. We should try and work around this race condition. Comments :: On 2013/07/19 10:50 - dlaxalde wrote : Isn't there a better solution than this 15s sleep? This is very expensive in tests (that of mercurial-server and deps in particular). Ticket #2818312 make it possible to disable the refresh-repos hook ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending we have setups where multiple instances share a local cache. only one of them should be in charge of updating the cache, which is not possible currently. Comments :: On 2013/04/12 11:45 - acampeas wrote : s/hook/looping task ? > On 2013/04/12 11:53 - jcristau wrote : > yes. that. Ticket #2735031 make sure revision changeset are not interpreted as anything else ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: validation pending for now, if one has an (weird) hg repo in which a branch is named as a revision changeset, the import mecanism affect the revision number of the nrahch head instead of the revision number of the changeset. Ticket #2532125 Make VCSFile not provide "write access" though the web by default ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: validation pending Current implementation of VCSFile includes controllers to upload revisions from the web. These should be disabled by default. Version 1.12.2 -------------- :expected date: n/a Ticket #2742334 add accessor on Repo entity to get the local hg repository path ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: validation pending Ticket #2758320 broken handling of non-ascii chars ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: validation pending seems like revision attributes (author, description) don't use the right charset Version 1.12.1 -------------- :expected date: n/a Ticket #2731144 add repository dc title in revision imported message ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: done This INFO level message about revision does not include any repository info. Ticket #2731063 broken handling of non-ascii chars ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: not validated seems like revision attributes (author, description) don't use the right charset Ticket #2731056 should detect new obsolescence from existing revisions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: done when an obsolescence relation is added between two existing revisions, vcsfile correctly detects that one is hidden, but doesn't add the obsolescence relation to the db. Comments :: On 2013/03/07 14:18 - pydavid wrote : It is very hard to get a correct result here. There is two tricky situation: * insertion a changeset in an obsolescence chain * new obsolescence marker applying a a public changesets. I'll do a best effort fix. Ticket #2731163 silently ignore rmtree failure if the directory does not exist at all ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: done Current implementation flood critical message when you delete a repo without cache yet. Ticket #2731025 revision primary view depends on vcreview ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: done the text/x-diff to text/annotated-html transform is in vcreview/docparser.py Version 1.12.0 -------------- :expected date: n/a Ticket #2723594 Improve import transactionality ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: done current import process is very inconvenient to hook in. sugguested changes are: - update phases and hidden status before importing new revision. - proper commit every behavior. - greater default value for import. This will allows operation to have a useful vision of the repository state at commit time. Ticket #2718635 tests depend on tag and folder cubes: make this optional ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: done for it is very inconvenient However I don't know how we can make the test/data/bootstrap_cubes thing more flexible. Comments :: On 2013/02/21 08:49 - sthenault wrote : I'm not sure to understand the real pain here, while I clearly see problems: no more tested features / variation accross different installation. vcsfile has optional features depending on tag/folder cubes, which are then in recommends. Forcing their install in test so those features are actually tested sounds better to me than not testing them (which **will** be the case by making them optional). > On 2013/02/21 09:40 - acampeas wrote : > I use a virtualenv like system and I don't like to be forced to have tag and > folder in them just to please a very marginal feature. > > And, yes, the feature is very marginal/invisible/undocumented. I did never > notice anything about it in our vcsfile using apps. Ticket #2723593 view revision diff on Revision view ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: done For now we just have touched file information which is mostly useless. Version 1.11.0 -------------- :publication date: 2013/03/05 :expected date: n/a Ticket #2502647 Add cubicweb-ctl command to start repository update manually ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved same as ticket #1635341 but with a cubicweb-ctl command rather than an Action button Comments :: On 2013/02/01 07:56 - sthenault wrote : have you checked the vcscheck command first? Ticket #2718619 do not clone local cache just at Repository creation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: resolved leave it to the looping task (which may fail in a non-blocking way) Ticket #2459803 enhance schema and import mecanism to support obsolete extension ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 2.000 :state: resolved Changes necessary to the ui should go in another story Comments :: On 2012/08/23 12:01 - acampeas wrote : I see: * phase support (per cset) * predecessor/ancestor relation Anything else ? > On 2012/09/14 14:58 - acampeas wrote : > do not forget to handle killing a patch (in mercurial this is done with an > obsolescence relation to the null revision) Ticket #2706946 hg: problems with latest import_content implementation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.250 :state: resolved The unicity contraint on revision prevent from doing 'SET R revision NULL' beforere numbering (on Postgresql), and the knownheads should recomputed just before computing 'incoming' revisions. Ticket #2681160 mercurial 2.5 support ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: resolved Ticket #1635341 provide a mechanism to start repository update manually ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved This mechanism might then be used by mercurial hook to get a faster update. Ticket #2508157 upgrade to API of CW 3.15 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: resolved Version 1.10.0 -------------- :expected date: n/a Ticket #2586293 do **not** import hg repositories by ranging on revision numbers ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: done use a "hg incoming" like method instead; current implementation is notoriously broken. Version 1.9.4 ------------- :expected date: n/a Ticket #2294506 Be less verbose by default ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: done do not log at info level each ``hg pull`` Comments :: On 2012/08/23 11:53 - acampeas wrote : why is it a problem ? shouldn't by default production instance run with a > INFO log level ? Ticket #2189200 VersionContent should store HG rev ID, not revision numbers ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: done Keeping revision number is the assurance that stored data will soon be inconsistent... Version 1.9.3 ------------- :publication date: 2012/06/28 :expected date: n/a Ticket #2410675 crash when strip detected on repository that don't import revision content ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved due to bad rql queries Ticket #2150538 crash while patch content is being loaded ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved If you push to a patch repository then switch to the web UI, you can get a "NoneType object has no method getvalue" when the changeset was seen by the system and created in CW, but the content not yet imported. This looks like a failure, but needs a proper message warning the user that the patch was seen, but its content is not available yet. Ticket #2410674 missing category on hooks ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved one may want to e.g. create a repository ignoring errors if the source url isn't yet accessible Version 1.9.2 ------------- :publication date: 2011/12/22 :expected date: n/a Ticket #2120344 missing "Repository granted_permission CWPermission" relation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.9.1 ------------- :publication date: 2011/12/08 :expected date: n/a Ticket #2088025 crash while importing repositories content ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.9.0 ------------- :publication date: 2011/11/09 :expected date: n/a Ticket #2083010 add an option to clone repositories from an external source ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: resolved Ticket #2058151 Implement local_cache for "foreign" Repositories ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: resolved the local_cache functionnality is only coded for Repositories in the "system" source. It would be usefull to have a cache for all multisource (or Datafeed) Repositories... but we have to make sure not to import all revisions and patches... Ticket #2065862 TypeError: __init__() takes exactly 5 arguments (2 given) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved 2011-11-02 11:04:05 - (cubicweb.sources.hg) ERROR: error while updating local cache of repository 2038922 Traceback (most recent call last): File "/usr/share/cubicweb/cubes/vcsfile/hooks.py", line 145, in refresh_local_repo_caches repohdlr.pull_or_clone_cache(vcsrepo.source_url) File "/usr/share/cubicweb/cubes/vcsfile/repohg.py", line 97, in pull_or_clone_cache raise bridge.VCSException(str(ex)) TypeError: __init__() takes exactly 5 arguments (2 given) Ticket #2080256 typo in check-revision-commit-every option ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved site_cubicweb.py (and thus all-in-one.conf) uses a wrong name. The code checks for the right one, so the config doesn't actually change anything. Ticket #2026628 wrong english translation for parent_revision_object ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved Ticket #1993000 Upgrade to CW 3.14 API ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved Needed (at least): * linkto api update * add localperms cube dependency Ticket #2058152 when strip is detected, query used to drop extra revisions is buggy ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved as it doesn't consider the repository... Version 1.8.7 ------------- :publication date: 2011/09/26 :expected date: n/a Ticket #1958218 don't log traceback on url error or unauthorized when cloning/updatig an hg repo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved this clutters the log for no benefit Ticket #1777404 fix VersionedFile dc_long_title (+ improve its i18n) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved It was previously using a non-world-readable attribute and its i18n string was incorrect english. See proposed changeset d766989873cd of https://bitbucket.org/fcayre/vcsfile Comments :: On 2011/07/11 16:52 - sthenault wrote : I've currently no access to this repository (permission denied). If not resolved quickly, I'll publish a release with only access to non-word readable attribute fixed. > On 2011/09/26 14:31 - fcayre-old wrote : > should be fixed now, sorry Ticket #1958220 task doing repo pull should watch the repository being shutting down ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved as its thread may: * don't receive the KeyboardInterrupt or shutdown signal * loop a while without accessing the session (until some repository is actually not up-to-date) (access to the session should abort the task) Version 1.8.6 ------------- :publication date: 2011/09/08 :expected date: n/a Ticket #1902356 NameError: global name 'CACHEROOT' is not defined ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved VCSFileInitHook raises NameError instead of ValueError when it doesn't have write permissions to the repo cache. Ticket #1916263 source_url is used as both checkout and browse url ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved The url to give to hg clone / svn checkout is not always browseable, but the source_url attribute for repositories is used for both in vcsfile. Ticket #1916050 svn repos not imported ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :state: resolved the repo import hook is currently restricted to mercurial repositories, so no revisions/versioncontent/... are created from svn repos. Version 1.8.5 ------------- :publication date: 2011/07/12 :expected date: n/a Ticket #1809469 shouldn't systematically pretend the repo has anonymous access ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Ticket #1809464 VersionContent.data is None in after_add_hook ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved because of a bug in VCStorage Version 1.8.4 ------------- :publication date: 2011/06/27 :expected date: n/a Ticket #1780272 at_revision must not be set in a late operation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved so bare operation may rely on it Ticket #1780282 name error when renamed/copied file not found during import ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.8.3 ------------- :publication date: 2011/06/21 :expected date: n/a Ticket #1749091 buggy set_at_revision w/ merge nodes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved Ticket #1749093 c-c vcscheck should delete erroneous revision/files/contents ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved so w e ensure once ran, everything is ok Ticket #1749092 c-c vcscheck should reset parent_revision and at_revision ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved so we ensure we don't pull erroneous stuff Ticket #1749094 fix vcsfile initialization ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved * don't modify configuration directly, use an attribute instead * proper init on server_maintenance Version 1.8.2 ------------- :publication date: 2011/05/16 :expected date: n/a Ticket #1749085 [entities] fix `branches()` for empty repository ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.8.0 ------------- :publication date: 2011/05/11 :expected date: n/a Ticket #650339 [integrity] nothing to check we don't have two VersionedFile identifying the same file ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved should be done using RQLUniqueConstraint (and tested) Ticket #479752 provide a tool to check cw/hg consistency ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved * report unknown revision (striped) * hg / vcsfile data consistency -> propose to autofix (by default) Ticket #542285 views: versioncontent summary likely not escaped ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.000 :state: deprecated see (with ff): https://www.logilab.net/cwo/versioncontent/504366 Ticket #467055 sounds like it loops indefinitly when importing some library ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.000 :state: deprecated between rev 80 and 90 on a client lib. It seems to have some nasty changeset around there which are probably the cause of this Version 1.7.2 ------------- :publication date: 2011/04/07 :expected date: n/a Ticket #1609905 default breadcrumbs for versioned files ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.010 :state: resolved Ticket #1609888 fix VersionedFile.head implementation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved should return something if head is not in default branch (aka 'tip') Ticket #1609965 kill old next_version method hiding the correct one ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.020 :state: resolved Version 1.7.1 ------------- :publication date: 2011/03/07 :expected date: n/a Ticket #1524661 only show repository's files tab if it's configured to import revision content ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.100 :state: resolved Ticket #1524581 fix previous/next version implementation on VersionContent ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved current impl miss some branch switches Version 1.7.0 ------------- :publication date: 2011/02/11 :expected date: n/a Ticket #1472778 repository with no local_cache but path set aren't imported anymore ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Ticket #1472736 specify wether revision's content should be imported per repository ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.300 :state: resolved not with a global option Ticket #1472801 ui tweaks ~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: resolved Version 1.6.6 ------------- :publication date: 2011/01/25 :expected date: n/a Ticket #1446864 upgrade to cw 3.10 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.000 :state: resolved - kill deprecation warnings (CubicWeb + unittest2) - minor test update Version 1.6.5 ------------- :publication date: 2010/10/22 :expected date: n/a Ticket #1347953 don't clone / pull up if we don't import revision content and repository-import = no ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved since we don't need the local cache in such cases. Also, this avoid error on our extranet when trying to set source_url to an url that is not accessible from the DMZ (only from the intranet where a CI dedicated instance run) Version 1.6.4 ------------- :publication date: 2010/10/14 :expected date: n/a Ticket #1296786 don't access to path accidentaly ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved may cause security issue / missing data Ticket #1296688 need update for cw 3.10 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: resolved since vcsfile use low-level tricks on the repository side... Ticket #1296691 nicer incontext/outofcontext views ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: resolved Version 1.6.3 ------------- :publication date: 2010/09/28 :expected date: n/a Ticket #1256566 repo primary tab should inherit from PrimaryTab, not PrimaryView ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved so right boxes will be properly displayed Ticket #1256563 set_pool in the wrong place makes import fail ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.6.2 ------------- :publication date: 2010/09/24 :expected date: n/a Ticket #1251928 multi-source: should not attempt to pull/import if repository is from an external source ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved Ticket #1251906 work around memory leak in mercurial ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved mercurial up to 1.6.3 has a memory leak in mercurial.url.httphandler caused by a cycle and a __del__ method. Submitted patch removing the __del__ method to mercurial. Monkey patch in the cube required for now. Comments :: On 2010/09/24 06:57 - Unknown author wrote : note: the fix has been pushed in mercurial crew stable (ca5fd84d62c6), so it should appear soon in an official release. The change in vcsfile should be save to use even in a new mercurial release. Version 1.6.1 ------------- :publication date: 2010/09/16 :expected date: n/a Ticket #1251789 local_cache should be store relative to instance data directory ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved when a relative path is specified in the config. Else that may cause pb on setup such as the one we have with on instance for the website, and another on another host, with a different id. Ticket #1251788 anonymous can't access to a file's content ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.300 :state: resolved due to security on repository.path Version 1.6.0 ------------- :publication date: 2010/09/13 :expected date: n/a Ticket #1251278 distinguish source url (public) and local path (private) and handle a local cache for distributed repos ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 2.000 :state: resolved Version 1.5.0 ------------- :publication date: 2010/07/16 :expected date: n/a Ticket #1188870 update for cw 3.9 api ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved Version 1.4.1 ------------- :publication date: 2010/06/07 :expected date: n/a Ticket #479359 repository revisions and files tabs should be paginated ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved else its way to costly Version 1.4.0 ------------- fix some migration pbs. :publication date: 2010/05/18 :expected date: n/a Ticket #969137 config option to deactivate automatic import ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.300 :state: resolved Ticket #891522 critical on access to some repo cache ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved because import_content initializes the cache & it is called only 5 mins after server startup -- in the interim: empty cache Ticket #891521 migrations pbs ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved need to remove some sync_schema_props_perms Version 1.3.1 ------------- bug fix release :publication date: 2010/04/08 :expected date: n/a Ticket #853019 bug in Repository.clear_all_cache ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Version 1.3.0 ------------- :publication date: 2010/03/26 :expected date: n/a Ticket #803590 update for cw 3.7.2 storage/sql gen api ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.600 :state: resolved and make vcsfile backend independant! Version 1.2.0 ------------- :publication date: 2010/02/26 :expected date: n/a Ticket #704893 Have an option to disable fulltext indexing of code imported ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.200 :state: resolved To avoid massive database activity on import etc. etc. Ticket #704873 import only revision information without full content ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved sometimes we don't need to import all the information from hg, it would be nice to just have the revisions without the full code Comments :: On 2010/02/17 10:46 - acampeas wrote : the code is not actually imported (it is "virtually" imported :) > On 2010/02/18 11:44 - alutz wrote : > well... it doesn't virtually overload our servers ;) Ticket #724802 kill the way too costly ancestor_revision relation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved move it too cubes needing it Version 1.1.1 ------------- Bug fix release :publication date: 2010/02/17 :expected date: n/a Ticket #696077 search is broken in forge instance ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved :: une erreur est survenue remove {'AA': 'String', 'AC': 'String', 'AB': 'String', 'E': 'VersionedFile', 'AD': 'Datetime', 'X': 'Repository'} from solutions since anon has no read access to Repository http://www.cubicweb.org/view?rql=sphinx&__fromsearchbox=1&subvid=tsearch Ticket #708819 vcsfiles is unable to scale ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved Importing Thousand of revision bloat the database. Version 1.1.0 ------------- :publication date: 2010/02/10 :expected date: n/a Ticket #676671 remove 3.6 deprecation warnings ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved Version 1.0.13 -------------- :publication date: 2010/02/02 :expected date: n/a Ticket #656172 api cleanup ~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved there is some methods on VersionedFile that doesn't smell good, remove them Ticket #656175 fix VC.head to properly consider branch ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Version 1.0.12 -------------- :publication date: 2009/12/23 :expected date: n/a Ticket #615262 only set revision's creation date as actual revision date ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved other entities (VersionContent, VersionedFile...) should have standard creation date (eg time of insertion in the repo) Version 1.0.11 -------------- :publication date: 2009/10/07 :expected date: n/a Ticket #473071 some view fixes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved form, primary views... Ticket #473068 use sql to insert ancestor_revision/at_revision relation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved since the number of those relations to insert when inserting a new revision is growing with the number of revision, we want speed here even if there not "hookable" anymore Version 1.0.10 -------------- bugfix :publication date: 2009/09/23 :expected date: n/a Ticket #465366 deleting from a rev in two branches leads to crash ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved trace: :: File "/home/auc/src/fcw/cubes/vcsfile/hooks.py", line 110, in commit_event transaction.commit() File "/home/auc/src/fcw/cubes/vcsfile/repohg.py", line 214, in commit raise QueryError('nothing changed') QueryError: nothing changed Version 1.0.7 ------------- :publication date: 2009/09/02 :expected date: n/a Ticket #353023 error while setting ancestor revision for merge nodes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Ticket #353021 should commit every X revisions while importing content to avoid pb ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved such as memory consumption or error after a lot of imported revision, in which case everything has to be done again Ticket #347144 pb w/ repository.path used in url & apache redirection ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved :: AllowEncodedSlashes On in apache configuration avoid systematic 404, but it still decode the %2F which is then seen by mod_proxy. The apache ml gave me the following ideas: # 'option 'B' de RewriteRule, mais dans ce cas il m'encode aussi tout les '/' et '?' présent dans l'url d'origine # utiliser ProxyPass à la place de nos rewrite rule, qui prend une option "nocanon". C'est prometteur mais dès qu'il y a des paramètres, je me les ramasse en double, eg si j'accède à view?vid=index, twisted récupère view?vid=index?vid=index (ça sent le bug, ou alors ProxyPass est pas du tout fait pour nous, ce qui n'est pas impossible non plus) Version 1.0.5 ------------- :publication date: 2009/08/18 :expected date: n/a Ticket #345243 can't use (yet) in place "and" or "or" w/ selectors ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Ticket #345247 we need to handle a ancestor_revision relation to be able to express some question using a single rql query ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.500 :state: resolved such as those necessary for docaster export Version 1.0.4 ------------- :publication date: 2009/08/06 :expected date: n/a Ticket #344973 missing import in migration script ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.000 :state: resolved crash if sqlite3 isn't installed because `sys` is not imported Ticket #344976 should set umask before adding a revision to hg ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved since in daemon mode, umask is set to 077 and we want 022 when creating new file in hg store. Ticket #344969 when creating revision using rql queries, the revision is reimported at synchronization time ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Version 1.0.3 ------------- :publication date: 2009/08/06 :expected date: n/a Ticket #344943 error while importing content should not block startup ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.100 :state: resolved Ticket #344939 syntax error in .po file ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.000 :state: resolved Ticket #344946 write support broken with hg 1.3.1 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Version 1.0.2 ------------- :publication date: 2009/07/28 :expected date: n/a Ticket #344806 misc fixes ~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.400 :state: resolved * packaging * fixed migration * crashing hook Version 1.0.0 ------------- do something wrt performances :publication date: 2009/07/22 :expected date: n/a Ticket #343688 general slowness ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 2.000 :state: resolved it seems like the slowness is proportional to the qty of versionedfile being accessed in the case I've observed Ticket #344238 avoid usage of sqlite database ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 1.000 :state: resolved we could move at least all table but VersionContent to the system database to avoid most queries being run in multi-sources mode. Also we could probably implements file data fetch in plpythonu, but getting necessary info using a query in the registered procedure, which would remove the need for the sqlite db. This requires important rewrite of the source though, which rely on some of the external entities properties. Version 0.9.1 ------------- :publication date: 2009/06/29 :expected date: n/a Ticket #323860 Bug in hg revision numbering ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved The current view of a hg repo view in CW is like:: type: mercurial encoding: utf-8 latest known revision: 3 path: /home/src/prive/sysinfo-doc contains: revision 2, revision 1, services/administration_fastt500.rst, revision -1, applications/crm.rst in which we can see that: - it shows revision -1 of the hg repo (which should not be visible), - revision 0 is not visible, - latest known revision is 3, whereas this repo was at revision 2 when this view has been generated. Comments :: On 2009/06/16 12:46 - sthenault wrote : can't reproduce by importing this repo into a cw instance. > On 2009/06/16 13:40 - anon wrote : > 'cause the -1 pb. has been fixed i'd say (it's 0 now) > > but this does not explain everything said there Ticket #344243 hg: revisions no modifying any files are not imported ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.400 :state: resolved also, when filtering on subpath, all revisions should be imported anyway else some methods won't works correctly (since we're missing some parents). Affects svn as well. Ticket #344240 ui needs uicfg tweaks ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved Version 0.9.0 ------------- :publication date: 2009/06/22 :expected date: n/a Ticket #154072 branches support for hg ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 2.000 :state: resolved Comments :: On 2009/04/24 18:02 - sthenault wrote : read support done Ticket #154141 renaming support for hg ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 1.000 :state: resolved Ticket #1369 problem under (light) load ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 0.200 :state: resolved trace: :: 2008-12-23 18:04:53 - (cubicweb.web) ERROR: ProgrammingError('SQLite objects created in a thread can only be used in that same thread.The object was created in thread id 1123035472 and this is thread id 1102387536',) Traceback (most recent call last): File "/home/auc/rep/cubicweb/cubicweb/web/application.py", line 320, in main_publish result = controller.publish(rset=rset) File "/home/auc/rep/cubicweb/cubicweb/web/views/basecontrollers.py", line 67, in publish return self.vreg.main_template(self.req, template, rset=rset) File "/home/auc/rep/cubicweb/cubicweb/cwvreg.py", line 165, in main_template res = self.render('templates', oid, req, **context) File "/home/auc/rep/cubicweb/cubicweb/cwvreg.py", line 158, in render return selected.dispatch(**context) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 143, in dispatch view_func(**context) File "/home/auc/rep/cubicweb/cubicweb/web/views/basetemplates.py", line 139, in call view.dispatch(w=self.w) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 143, in dispatch view_func(**context) File "/home/auc/rep/cubicweb/cubes/docaster/views/startup.py", line 66, in call self.wview('table', rset, headers=(_('document'), _('revision'), _('state'))) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 220, in wview self.view(__vid, rset, __fallback_vid, w=self.w, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 212, in view return view.dispatch(**kwargs) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 143, in dispatch view_func(**context) File "/home/auc/rep/cubicweb/cubicweb/web/views/tableview.py", line 153, in call table.render(self.w) File "/home/auc/rep/cubicweb/cubicweb/web/htmlwidgets.py", line 32, in render self._render() File "/home/auc/rep/cubicweb/cubicweb/web/htmlwidgets.py", line 275, in _render for column, sortvalue in self.itercols(rowindex): File "/home/auc/rep/cubicweb/cubicweb/web/htmlwidgets.py", line 289, in itercols yield column, self.model.sortvalue(rowindex, column.rset_sortcol) File "/home/auc/rep/cubicweb/cubicweb/common/uilib.py", line 523, in newfunc ret = function(*args, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/common/uilib.py", line 507, in newfunc ret = function(*args, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/common/uilib.py", line 497, in newfunc ret = function(*args, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/web/views/tableview.py", line 243, in sortvalue return entity.sortvalue() File "/home/auc/rep/cubicweb/cubicweb/entities/__init__.py", line 232, in sortvalue return self.dc_title().lower() File "/home/auc/rep/cubicweb/cubes/docaster/entities/document.py", line 246, in dc_title title = self.name if dc_title is None else '[%s] %s' % (self.keyword, dc_title) File "/home/auc/rep/cubicweb/cubicweb/common/entity.py", line 1057, in __get__ return eobj.get_value(self._attrname) File "/home/auc/rep/cubicweb/cubicweb/common/entity.py", line 714, in get_value rset = execute(rql, {'x': self.eid}, 'x') File "/home/auc/rep/cubicweb/cubicweb/dbapi.py", line 251, in execute return self.cursor.execute(*args, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/dbapi.py", line 578, in execute parameters, eid_key, build_descr) File "/home/auc/rep/cubicweb/cubicweb/server/repository.py", line 548, in execute build_descr) File "/home/auc/rep/cubicweb/cubicweb/server/querier.py", line 647, in execute results = plan.execute() File "/home/auc/rep/cubicweb/cubicweb/server/querier.py", line 160, in execute result = step.execute() File "/home/auc/rep/cubicweb/cubicweb/server/ssplanner.py", line 347, in execute inputmap) File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 186, in syntax_tree_search cursor = session.pool[self.uri] File "/home/auc/rep/cubicweb/cubicweb/server/pool.py", line 84, in __getitem__ cursor = self.source_cnxs[uri][1].cursor() File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 47, in cursor return self.cnx.cursor() File "/home/auc/rep/cubicweb/logilab/common/db.py", line 443, in cursor return self._cnx.cursor(PySqlite2Cursor) ProgrammingError: SQLite objects created in a thread can only be used in that same thread.The object was created in thread id 1123035472 and this is thread id 1102387536 Comments :: On 2009/01/07 17:50 - acampeas wrote : or :: 2009-01-07 18:49:35 - (cubicweb.web) ERROR: RuntimeError("svn source is busy, can't acquire connection lock",) Traceback (most recent call last): File "/home/auc/rep/cubicweb/cubicweb/web/application.py", line 320, in main_publish result = controller.publish(rset=rset) File "/home/auc/rep/cubicweb/cubicweb/web/views/basecontrollers.py", line 67, in publish return self.vreg.main_template(self.req, template, rset=rset) File "/home/auc/rep/cubicweb/cubicweb/cwvreg.py", line 165, in main_template res = self.render('templates', oid, req, **context) File "/home/auc/rep/cubicweb/cubicweb/cwvreg.py", line 158, in render return selected.dispatch(**context) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 143, in dispatch view_func(**context) File "/home/auc/rep/cubicweb/cubicweb/web/views/basetemplates.py", line 139, in call view.dispatch(w=self.w) File "/home/auc/rep/cubicweb/cubicweb/common/view.py", line 143, in dispatch view_func(**context) File "/home/auc/rep/cubicweb/cubes/docaster/views/startup.py", line 61, in call rset = self.req.execute(rql % 'LIMIT 5', {'dtt': dtype}) File "/home/auc/rep/cubicweb/cubicweb/dbapi.py", line 251, in execute return self.cursor.execute(*args, **kwargs) File "/home/auc/rep/cubicweb/cubicweb/dbapi.py", line 578, in execute parameters, eid_key, build_descr) File "/home/auc/rep/cubicweb/cubicweb/server/repository.py", line 548, in execute build_descr) File "/home/auc/rep/cubicweb/cubicweb/server/querier.py", line 647, in execute results = plan.execute() File "/home/auc/rep/cubicweb/cubicweb/server/querier.py", line 160, in execute result = step.execute() File "/home/auc/rep/cubicweb/cubicweb/server/mssteps.py", line 87, in execute self.inputmap) File "/home/auc/rep/cubicweb/cubicweb/server/sources/__init__.py", line 194, in flying_insert res = self.syntax_tree_search(session, union, args, varmap=varmap) File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 186, in syntax_tree_search cursor = session.pool[self.uri] File "/home/auc/rep/cubicweb/cubicweb/server/pool.py", line 84, in __getitem__ cursor = self.source_cnxs[uri][1].cursor() File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 47, in cursor return self.cnx.cursor() File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 34, in cnx timeout_acquire(self.source._cnxlock, 5) File "/home/auc/rep/cubicweb/cubicweb/server/sources/extlite.py", line 24, in timeout_acquire raise RuntimeError("svn source is busy, can't acquire connection lock") RuntimeError: svn source is busy, can't acquire connection lock On 2009/03/06 12:43 - anon wrote : now, the shit hits the fan, at a customer site : :: 2009-03-06 12:55:36 - (cubicweb.sources.vcsfile) CRITICAL: rollback error Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/cubicweb/server/pool.py", line 49, in rollback cnx.rollback() File "/usr/share/cubicweb/cubes/vcsfile/svnsource.py", line 46, in rollback ConnectionWrapper.rollback(self) File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 44, in rollback self._cnx.rollback() ProgrammingError: SQLite objects created in a thread can only be used in that same thread.The object was created in thread id -1290785904 and this is thread id -1274000496 2009-03-06 12:55:38 - (cubicweb.repository) ERROR: unexpected error Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/cubicweb/server/repository.py", line 549, in execute build_descr) File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 648, in execute results = plan.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 161, in execute result = step.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/ssplanner.py", line 347, in execute inputmap) File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 186, in syntax_tree_search cursor = session.pool[self.uri] File "/usr/lib/python2.5/site-packages/cubicweb/server/pool.py", line 84, in __getitem__ cursor = self.source_cnxs[uri][1].cursor() File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 47, in cursor return self.cnx.cursor() File "/usr/lib/python2.5/site-packages/logilab/common/db.py", line 443, in cursor return self._cnx.cursor(PySqlite2Cursor) ProgrammingError: SQLite objects created in a thread can only be used in that same thread.The object was created in thread id -1321206896 and this is thread id -1437209712 2009-03-06 12:55:38 - (cubicweb.web) ERROR: ProgrammingError('SQLite objects created in a thread can only be used in that same thread.The object was created in thread id -1321206896 and this is thread id -1437209712',) Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/cubicweb/web/application.py", line 321, in main_publish result = controller.publish(rset=rset) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/basecontrollers.py", line 66, in publish return self.vreg.main_template(self.req, template, rset=rset) File "/usr/lib/python2.5/site-packages/cubicweb/cwvreg.py", line 165, in main_template res = self.render('templates', oid, req, **context) File "/usr/lib/python2.5/site-packages/cubicweb/cwvreg.py", line 158, in render return selected.dispatch(**context) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/basetemplates.py", line 135, in call self.template_header(content_type, view) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/basetemplates.py", line 162, in template_header self.template_body_header(view) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/basetemplates.py", line 191, in template_body_header self.nav_column(view, 'left') File "/usr/lib/python2.5/site-packages/cubicweb/web/views/basetemplates.py", line 222, in nav_column box.dispatch(w=self.w, view=view) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/share/cubicweb/cubes/docaster/views/boxes.py", line 85, in call self.view('pendingproofreading', rset, w=w) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 211, in view return view.dispatch(**kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/share/cubicweb/cubes/docaster/views/worklist.py", line 108, in call self.show_user_pending_proofreading(self.req.user) File "/usr/share/cubicweb/cubes/docaster/views/worklist.py", line 72, in show_user_pending_proofreading self.wview('table', rset) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 216, in wview self.view(__vid, rset, __fallback_vid, w=self.w, **kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 211, in view return view.dispatch(**kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/tableview.py", line 158, in call table.render(self.w) File "/usr/lib/python2.5/site-packages/cubicweb/web/htmlwidgets.py", line 32, in render self._render() File "/usr/lib/python2.5/site-packages/cubicweb/web/htmlwidgets.py", line 281, in _render self.model.render(cellvid, rowindex, colindex, w=self.w) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/tableview.py", line 225, in render self.view('cell', self.rset, row=row, col=col, cellvid=cellvid, w=w) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 211, in view return view.dispatch(**kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/lib/python2.5/site-packages/cubicweb/web/views/tableview.py", line 269, in cell_call e.view(cellvid or 'outofcontext', w=self.w) File "/usr/lib/python2.5/site-packages/cubicweb/common/entity.py", line 407, in view row=self.row, col=self.col, **kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/cwvreg.py", line 158, in render return selected.dispatch(**context) File "/usr/lib/python2.5/site-packages/cubicweb/common/view.py", line 144, in dispatch view_func(**context) File "/usr/share/cubicweb/cubes/docaster/views/secondary.py", line 36, in cell_call desc = text_cut(entity.dc_description(), 300) File "/usr/share/cubicweb/cubes/docaster/entities/document.py", line 244, in dc_description return self.metadata.description or u'' File "/usr/share/cubicweb/cubes/docaster/entities/document.py", line 232, in metadata return self.head.content_metadata[0] File "/usr/share/cubicweb/cubes/vcsfile/entities.py", line 140, in head rset = self.req.execute(rql, {'x': self.eid}) File "/usr/lib/python2.5/site-packages/cubicweb/dbapi.py", line 251, in execute return self.cursor.execute(*args, **kwargs) File "/usr/lib/python2.5/site-packages/cubicweb/dbapi.py", line 578, in execute parameters, eid_key, build_descr) File "/usr/lib/python2.5/site-packages/cubicweb/server/repository.py", line 549, in execute build_descr) File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 648, in execute results = plan.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 161, in execute result = step.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/ssplanner.py", line 347, in execute inputmap) File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 186, in syntax_tree_search cursor = session.pool[self.uri] File "/usr/lib/python2.5/site-packages/cubicweb/server/pool.py", line 84, in __getitem__ cursor = self.source_cnxs[uri][1].cursor() File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 47, in cursor return self.cnx.cursor() File "/usr/lib/python2.5/site-packages/logilab/common/db.py", line 443, in cursor return self._cnx.cursor(PySqlite2Cursor) ProgrammingError: SQLite objects created in a thread can only be used in that same thread.The object was created in thread id -1321206896 and this is thread id -1437209712 On 2009/03/09 09:40 - anon wrote : You are using an sqlite database in production ? Why aren't you using postgres ? > On 2009/03/09 22:19 - acampeas wrote : > sqlite is an implementation detail of the multisource mecanism for vcsfile > > postgres is used for everything that is not part of the vcsfile schema Version 0.8.0 ------------- :publication date: 2009/05/18 :expected date: n/a Ticket #342938 update for cw 3.2 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 0.500 :state: resolved Ticket #154074 write support for hg ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: enhancement :load: 2.000 :state: resolved Version 0.7.4 ------------- :publication date: 2009/03/31 :expected date: n/a Ticket #320285 crash on migration ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :load: 1.000 :state: resolved many of these: :: Traceback (most recent call last): File "/usr/share/cubicweb/cubes/vcsfile/svnsource.py", line 112, in version_data repohdlr = source.repo_eid_path[int(repoeid)] KeyError: 687 2009-03-31 13:53:15 - (cubicweb.entity) ERROR: can't add value of data to text index for entity 10424 Traceback (most recent call last): File "/usr/lib/python2.5/site-packages/cubicweb/common/entity.py", line 1037, in get_words value = self.printable_value(rschema, format='text/plain') File "/usr/lib/python2.5/site-packages/cubicweb/common/entity.py", line 500, in printable_value value = getattr(self, attr) File "/usr/lib/python2.5/site-packages/cubicweb/common/entity.py", line 1069, in __get__ return eobj.get_value(self._attrname) File "/usr/lib/python2.5/site-packages/cubicweb/common/entity.py", line 720, in get_value rset = execute(rql, {'x': self.eid}, 'x') File "/usr/lib/python2.5/site-packages/cubicweb/server/session.py", line 285, in unsafe_execute propagate) File "/usr/lib/python2.5/site-packages/cubicweb/server/session.py", line 299, in execute rset = self._execute(self, rql, kwargs, eid_key, build_descr) File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 648, in execute results = plan.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/querier.py", line 161, in execute result = step.execute() File "/usr/lib/python2.5/site-packages/cubicweb/server/ssplanner.py", line 347, in execute inputmap) File "/usr/lib/python2.5/site-packages/cubicweb/server/sources/extlite.py", line 187, in syntax_tree_search cursor.execute(sql, args) File "/usr/lib/python2.5/site-packages/logilab/common/db.py", line 431, in execute self.__class__.__bases__[0].execute(self, final_sql , kwargs) OperationalError: user-defined function raised exception Version 0.7.3 ------------- :publication date: 2009/02/09 :expected date: n/a Ticket #1520 title of versioned files wrongly tells file is 'DELETED' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :type: bug :state: resolved albeit it is not when this happens, a hard refresh in ff and it is displayed correctly