.. -*- coding: utf-8 -*-

Project apycot-moved
====================
:creation date: 2006/09/29

an Automated Pythonic Code Tester

This package is designed to run tests on a code repository on a daily basis. It
comes with a set of predefined test, essentially for python packages, and a set
of predefined reports to display execution results. However, it has been
designed to be higly extensible, so you could write your own test or report
using the Python language (i.e. this framework is NOT limited to test Python
code !).

Use it to daily run your tests and generate reports about code or document
quality.

Apycot now rely on Narval_: tests are defined as Narval's recipes (using
*actions* provided by apycot and its plugins) and are executed by the Narval
bot.

.. _Narval: http://www.cubicweb.org/project/narval

Version 2.0.3
-------------
:publication date: 2010/09/28
:expected date: n/a

Ticket #46866 various bug fixes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

* bad prev/next test execution (lead to erroneous notification) * inherited test
config are not properly started * dropped buggy urls * some other minor fixes

Version 2.0.2
-------------
:publication date: 2010/09/24
:expected date: n/a

Ticket #46156 various bug fixes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 5.000
:state: resolved

Version 2.0.1
-------------
:publication date: 2010/09/16
:expected date: n/a

Ticket #46107 fix 2.0!
~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 2.000
:state: resolved

lot of little stuff not working properly

Version 2.0.0
-------------
:publication date: 2010/09/13
:expected date: n/a

Ticket #18278 vcsfile should be kept optional
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: rejected

Ticket #22975 add link from Project to related ProjectEnvironment
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

Comments ::

  On 2010/05/11 08:04 - sthenault wrote :
  we already have this. What's missing according to you?

Ticket #19307 Change cardinality between ProjectEnvironment/TestConfig
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 2.500
:state: resolved

I have several TestConfig entities which are all the same. I see no reason to
not have the cardinality (\*\*)::

use_environment = SubjectRelation( 'ProjectEnvironment', cardinality='1*',
composite='object', description=_('project environment in which this test config
should be launched'), constraints=[RQLUniqueConstraint('S name N, Y
use_environment O, Y name N', 'Y')], inlined=True )

Comments ::

  On 2009/12/28 15:31 - Unknown author wrote :
  Question: why not merge TestConfig and TestGroupConfig ? Why are the differences
  except current cardinalities ?
  > On 2010/01/13 08:45 - sthenault wrote :
  > because that's two different things. TestConfig are meant to be 'executable'
  > while TestConfigGroup are not.
  > TestConfigGroup only provides a small part of a whole test configuration, and
  > I find it's simpler to have two
  > different kind of entities rather than having to guess if it's a group or not
  > latter.
  On 2010/01/13 08:42 - sthenault wrote :
  that could make things simpler... But we've to take care to some places in the
  code where this single card is expected.
  > On 2010/01/19 13:47 - sthenault wrote :
  > also currently TestConfig is used to hold test dependencies, so can't always
  > be shared. Ludal's idea to move such information into the source would solve
  > this.

Ticket #3007 jslint integration
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

to check javascript files

Ticket #6869 merge apycotbot / cubicweb-apycot repositories
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

Ticket #20004 ProjectEnvironment config is least important than TestConfigGroup one
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

We need to clearly set configuration priority

Ticket #18256 recipes for test execution
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 5.000
:state: resolved

one should be able to describe test by connecting blocks into a recipe à la
narval.

preprocessors and checks should be available as base building block.

sounds like I want to turn apycot bot into a narval...

Comments ::

  On 2009/09/29 10:06 - nchauvat wrote :
  Please go ahead and do it my friend :)

Ticket #20247 apycotbot server can't write data about execution.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 4.000
:state: resolved

This a serious issues as the bot can not report
early-arborted/killed/dead/failing-to-clean test

Ticket #20692 Have local "cache" for remote repositories for hg
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

There should be persistant clone of the source repository on the machine where
apycotbot is running so we do not clone over and over, which can be a huge
network and disk I/O consuming process.

This is really important if we want to be able to use apycot for several
mid-sized projects without having to buy an entire data center.

Version 1.10.0
--------------
:publication date: 2010/07/28
:expected date: n/a

Ticket #37174 store option's values and version information as standard messages...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

... instead of raw values to limit eid explosion and speed-up things

Ticket #37171 update for cw 3.9 api
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

Version 1.9.2
-------------
:publication date: 2010/06/11
:expected date: n/a

Ticket #33991 Avoid too much search engine indexing
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: resolved

It seems that the number of pages generated for each commit is pretty huge...
some of that info is not really usefull for a search engine. Find a way to avoid
them crawling all the pages ...

Ticket #34338 don't see project environment in test exec report startup view
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.100
:state: resolved

making it hardly understandable...

Version 1.9.1
-------------
:publication date: 2010/06/07
:expected date: n/a

Ticket #22098 pyunit report NODATA without message about the faulty files
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

Comments ::

  On 2010/03/30 12:10 - sthenault wrote :
  http://www.cubicweb.org/testexecution/804940
  
  il est en 'pas de données', il devrait être en 'partial'
  
  si on regarde dans les logs on voit que c'est parce qu'il n'a pas réussi à
  parser des tests avec des trucs skippés.
  
  c'est sans doute que notre pyunit parser ne gère pas ça

Ticket #33896 bad test execution status
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

always <= failed...

Ticket #33040 migration failed for  1.8.3 to 1.9.0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

missing add_cube('file') -> can't execute command
add_relation_definition('TestExecution', 'log_file', 'File')

Version 1.9.0
-------------
:publication date: 2010/05/18
:expected date: n/a

Ticket #23317 cosmetid change : ensure column order and tab order match
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.300
:state: resolved

in the project environment view, executions tab, the data is displayed like
this:

================== ======= ========= ====================== nom branche lgp
check pyunit ================== ======= ========= ====================== config
test pegase Test erreur échec default échec succès stable échec en cours de
traitement ================== ======= ========= ======================

However in the execution view, the tabs are not ordered in the same fashion: on
my instance I get pyunit first ans lgp check second. While not strictly speaking
a bug, I find this "feature" disturbing each time I browse my project's report
pages.

Ticket #23027 apycotbot.pending_task() may return eid unrelated to the current instance
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

the instance ID should be taken as argument or included in the return value.

Ticket #19728 apycot bot state
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: resolved

The apycot bot state view must show all tests running, including tests running
on another instance that the current one.

For now, this view show only this instance tests and one can think the bot is
not doing anything whereas he is busy.

Ticket #21579 Automaticaly remove TestExecution older than D
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.300
:state: resolved

Ticket #20720 clarify the mail notification Subject header
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

The current mail sent by apycotbot has a terrible subject which makes it hard to
process:

[cwo] quick(email test env)@stable: pyunit status changed to success

All mails I've received so far start with [cwo] quick (XXXXX test env)

What I need to be able to see easily is the name of the project concerned by the
report, then the test env name, then the change.

I propose something like (to be discussed):

[cwo] projectname@branch testenv: old_status -> new_status (testname)

So the report above would have been labelled:

[cwo] cubicweb-email@stable email test env : failure -> success (pyunit)

I'd also advocate dropping the 'test env' part in test env names. These can be
added automatically in the TestEnvironment view if necessary.

Ticket #19798 Filter branch by branch in reports view
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

Ticket #22680 highlight the directory download
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.400
:state: resolved

If you keep the archive test directory:

Add a big Rectangle to donwload the directory or put a link in the value for the
label "Test directory archive name".

Currently the link is hidden in a submenu of the action box.

Ticket #21580 Improve report table
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: rejected

Ticket #19179 linking TestExecution to Revision pb
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.300
:state: resolved

one pb is that when a test is started, it may use some revision of dependency's
repositories that may not be already imported by vcsfile. So when it try to
create a link between the execution and the revision, it fails. In such case the
information is kept as a CheckResultInfo object, so we should probably use this
to create the link later when the changeset is imported.

Ticket #22974 simple action to create a local repository for a project environment
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

Ticket #20865 TestConfigGroup check_config values are not propertly displayed
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

The TestConfig primatry view have a nice table to show them but the
TestConfigGroupPrimary view show them unformated.

Ticket #23262 the two rss icons have the same title
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: rejected

at least one is wrong

Ticket #20252 'used by' panel of groups don't show names of tests
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

the 'used by' panel of the view of a group doesn't show names of tests

Comments ::

  On 2010/02/08 11:10 - anon wrote :
  The "used by" side box for TestConfigGroup Primary View seems to display the
  TextConfig Name "In context". But the "In context" for Test config is intended
  for "ProjectEnvironment Contex. This lead to TestConfig name not decorated with
  ProjectEnvironment name who can we are to distinct from each other.

Ticket #23271 include execution status in TestConfig execution listing
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.300
:state: resolved

Ticket #22471 installed apycotbot.ini use cw-app-id instead of cw-inst-id
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

Ticket #21619 revision used for test is not displayed for anon
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.300
:state: resolved

Comments ::

  On 2010/03/15 10:18 - pydavid wrote :
  The Revision number of environement linked to a repository is not displayed for
  anonymous user.
  On 2010/03/31 12:09 - anon wrote :
  I think this was solved "immediately" for http://apycot.hg-scm.org
  > On 2010/03/31 12:19 - pydavid wrote :
  > I quickfixed it for http://apycot.hg-scm.org. Clean solution is still to be
  > implement

Version 1.8.3
-------------

bug fix release
:publication date: 2010/03/31
:expected date: n/a

Ticket #21985 branch defined in dependencies not used
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Ticket #22649 html content of description and change views are escaped
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #22099 ldi_upload is broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.250
:state: resolved

Ticket #22114 Préprocessors warning are broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.250
:state: resolved

Version 1.8.2
-------------

bug fix version
:publication date: 2010/03/18
:expected date: n/a

Ticket #21979 lgp_check if broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Ticket #21978 pylint checker crash if project pylint RC ask for persitent result.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.250
:state: resolved

Ticket #21864 SyntaxError in migration file 1.4.0_Any.py
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.250
:state: resolved

missing ',' in "# fix needs_checkout relation"

Ticket #21977 dependancies are not installed
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Version 1.8.1
-------------
:publication date: 2010/03/11
:expected date: n/a

Ticket #21678 default options values ignored.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #21676 pkg_doc checker doesn't use the new checker API to call it's sub checkers (and therefor fails to call them)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #21677 preprocessors handler issue warnings about options defined at the right places
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Ticket #21691 Status change RSS report processing execution.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #21684 The branch notation should follow the mercurial convention
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

we use <repo>@<branch> but mercurial use <repo>#<branch>

Ticket #21681 The quiet option of python_setup preprocessors is not handled propertly
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

The type attribut of the option definition is not used by apycot.

Ticket #20250 Output of failing preprocessors are not visible in the log
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Version 1.8.0
-------------
:publication date: 2010/03/09
:expected date: n/a

Ticket #19306 TestConfig configuration not properly read by preprocessor
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

I have several Testconfig configuration lines as:

lgp_build_distrib=lenny lgp_build_verbose=yes

But theses one are not honored by the preprocessor. I need to add a duplicated
configuration into the ProjectEnvrionment entity.

Don't know if it's the expected behaviour. Just for record.

Ticket #9061 bad python_unittest status
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

if some tests are executed, and some other test files are discovered but no test
are found inside them, final status should be 'error' whatever the status of
executed tests.

Comments ::

  On 2010/01/27 10:04 - pydavid wrote :
  At the moment we have the following priority::
  
  success < partial < nodata < failure < error < killed < skipped < missing
  
  Does this ticket mean you would prefer::
  
  success < partial < failure < nodata < error < killed < skipped < missing
  
  or even further::
  
  success < partial < failure < error <nodata < killed < skipped < missing
  > On 2010/01/27 10:19 - Unknown author wrote :
  > IMO, We should follow xUnit definition. A recent discussion about that is
  > available in the archive of testing-in-python list.
  On 2010/01/28 15:42 - pydavid wrote :
  pyunit and pytest checker return a wrong status for empty file

Ticket #21057 direct link to a specific log message.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.250
:state: resolved

Ticket #20006 Enhance log coloring
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.250
:state: resolved

Ticket #20406 rss feed notification of status changes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.500
:state: resolved

Ticket #20293 when displaying messages, exclude debug
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

for example on http://www.cubicweb.org/testexecution/678486?tab=pyunit only the
failures could be displayed with a top-right menu to print a menu with the
option "filter", similarly to what is done for the tickets of a project in
cubicweb-forge.

Ticket #20249 preprocessor options are broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 2.000
:state: resolved

Comments ::

  On 2010/02/24 17:08 - anon wrote :
  Preprocessors only use the ProjectEnvironment configuration because it's also
  run on dependency.
  > On 2010/03/03 12:41 - pydavid wrote :
  > A warning will be added as a temporary solution.

Version 1.7.3
-------------

version corrective du paquet source.
:publication date: 2010/02/19
:expected date: n/a

Ticket #20842 cubicweb-apycot package doesn't install
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

the wdoc conflict with cubicweb main wdoc

Version 1.7.2
-------------

bugfix release
:publication date: 2010/02/19
:expected date: n/a

Ticket #20696 Handling of loading pluging error are broken.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: open

Ticket #20698 make_uid is not used propertly
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

make_uid take an argument

Ticket #20697 start all command is broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

``apycot client start`` without target config is broken.

Ticket #20604 Unable to reference the project env checkout directory from TestConfig
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

Ticket #20776 no  apycotbot section in generated documentation (wdoc)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

I can find a valid link and no content is available in http://.../doc/main

Version 1.7.1
-------------
:publication date: 2010/02/16
:expected date: n/a

Ticket #20428 ':' in archive name confuse tar
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #20427 writer crash on empty log
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

Ticket #20416 email notification may failed when fetching additionnal information from vcsfile
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Ticket #20248 pluggin feature is broken
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

pluggin loading is broken because of some old reference to the writer.

Ticket #20417 TestExecution request may fail silently
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

They are never added to the Bot queue but doesn't report any error.

Comments ::

  On 2010/02/11 17:27 - pydavid wrote :
  This occurs on TestConfig without any checkers.

Version 1.7.0
-------------
:publication date: 2010/02/10
:expected date: n/a

Ticket #20298 remove 3.6 deprecation warnings
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

Version 1.6.0
-------------
:publication date: 2010/02/05
:expected date: n/a

Ticket #9400 integrate documentation provided by email to didrik on the python-projects ml
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

Ticket #18230 pp_buildep should follow lgp configuration by default
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

The default configuration doesn't honor /etc/lgp/lgprc. Set real optional fields
for 'distrib' and 'archi'

Ticket #20002 ProjectEnvironment config is never used for test execution
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.250
:state: resolved

check_config et check_environment of ProjectEnvironment is not inherited by
TestCondig.

Ticket #19308 TestConfig name cannot have space inside
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

seems to clash with shell exec..

Ticket #19193 daily tests of elo/cwo doesn't seems to be executed at all
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

find out why...

Ticket #18405 download/remove content of test tmp dir when it's kept
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 2.500
:state: resolved

* take care to security issue (should not provide a way to access to the bot
host fs) * when asking the bot to remove the tmp dir, remove the CheckInfo
telling its place and put TestExecution in "end" state. Then actions shouldn't
be accessible anymore

Comments ::

  On 2009/10/16 09:09 - Unknown author wrote :
  the recent ssl patch from syt to upstream maintainer could be used to enforce
  security here ?

Ticket #19765 Execution may have no branch
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.300
:state: resolved

Branch value is not enforced and TestExecution may have not branch data.
Moreover old TestExecution may lack this data too. TestExecution with branch
None cripple most report view.

Ticket #9398 get back documentation which still make sense from the previous apycot project
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

Ticket #19162 unreadable logs
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

currently it's hard to read preprocessors/logs file. The most notable pb is that
often the information that help to find why something failed is in a separated
DEBUG message above the ERROR message.

IMO:

* ERROR messages due to command failures should hold the debug information *
when a message do not apply to a particular path in the source, use the path
information to hold the command name * (moved to #20006) to colorize logs
according to severity * (moved to #20005)each checker should be responsible to
tell why it failed (and add this information to the failure notification mail)

Comments ::

  On 2009/12/03 14:43 - sthenault wrote :
  1.5.0 includes most of this but:
  
  * stylesheet should be completed
  * last item still to implemented
  * each checker messages handling should be reviewed

Version 1.5.0
-------------
:publication date: 2009/12/03
:expected date: n/a

Ticket #19161 nosy list support
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

should recommends nosylist cube and if available, support registration by
ProjectEnvironment and TestConfig

Ticket #18722 should restart apycotbot after upgrade
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: rejected

Ticket #5116 subprocess cleaning
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: deprecated

we should make sure all subprocess of a Test are terminated at it's end.

Ticket #7989 apycotbot process disappears
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: deprecated

some times the apycotbot process gets killed, unclear why. Also some test
process gets killed as well, living an apycot execution in the "processing"
state.

* This may be related to resource control, that should be tested (cpu time,
memory consumption...)

* Have a process watcher restarting the bot if necessary ? though it would be
much better to identify why it disappears and fix this that way

Comments ::

  On 2009/12/02 08:36 - sthenault wrote :
  this pb hasn't been hit since the new apycot reincarnation, deprecate it.

Ticket #18858 test status change notification enhancements
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

* change subject to `[<env name>] <test name>@<branch name>: ...`

* add vcs events between too executions

* each checker should be responsible to tell why he failed, then add this
information to mail (probably another ticket)

Version 1.4.3
-------------
:publication date: 2009/10/21
:expected date: n/a

Ticket #18404 allow to specify priority when starting task from ui
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

only for manager?

Version 1.4.0
-------------
:publication date: 2009/10/07
:expected date: n/a

Ticket #7795 add a status on TestExecution
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.300
:state: resolved

::

setup -> running tests -> cleaning -> end -> not removed

("not removed" when the test has been ran with the `--keep-test-dir` option)

Ticket #18263 -force option to apycotbot
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: resolved

to erase previous record in the pyro-ns insteal of failing

Ticket #18266 lgc.pytest checker
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

Ticket #18260 tabbed primary view for TestConfig
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

* configuration * execution (ordered table)

Ticket #5794 use jquery + ajaxtabs to make reports lighter
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: resolved

for each report about a project, put summary on front page and one tab per test.
load data only when needed.

Comments ::

  On 2008/10/29 19:23 - Unknown author wrote :
  IMO, a clean and simple solution should be to include an XSLT stylesheet in
  apycot output.

Ticket #7987 configuration primary view should display inherited information
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.400
:state: resolved

since some values may be "inherited" from used groups, we should see the
actually used configuration (displaying what's inherited differently from what's
defined on the project configuration)

Ticket #5265 install a default configuration into the debian package
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

The debian package should install a default configuration :

+--------------------------------+------------------------------------------------------------------------------------------------------+
| path | description |
+================================+======================================================================================================+
| /etc/apycot | base directory for apycot configuration |
+--------------------------------+------------------------------------------------------------------------------------------------------+
| /etc/apycot/projects-available | directory containing all available test or
deps (<name>.ini) |
+--------------------------------+------------------------------------------------------------------------------------------------------+
| /etc/apycot/projects-enabled | directory containing symlink to the available
item we want to test |
+--------------------------------+------------------------------------------------------------------------------------------------------+
| /etc/apycot/main.ini | Main configuration file running all test in
projects-enable |
+--------------------------------+------------------------------------------------------------------------------------------------------+
| /etc/apycot/main-data.ini | Configuration file defining the location of the
mode of the test |
+--------------------------------+------------------------------------------------------------------------------------------------------+
| /etc/apycot/main_reports.ini | Configuration file allowing generation of a
minimal reports from test runned with the main.ini file |
+--------------------------------+------------------------------------------------------------------------------------------------------+

Example of cron file using this configuration for daily mode shoudl also be
provide.

Comments ::

  On 2009/02/02 14:38 - Unknown author wrote :
  Maybe we could deprecated this ticket ? or do we want to provide a default
  content to CubicWeb ?
  > On 2009/02/09 13:17 - sthenault wrote :
  > nop, the debian package should install a default /etc/apycotbot.ini file.

Ticket #18298 missing command to apycotbot or apycotclient to stop / restart the bot
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: rejected

currently it has to be killed manually

Ticket #6859 use one tab per check
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

An ApycotExecution should have tabs to display its content. On the first page,
display start_time, end_time, execution information and checks summary, then put
the report of each check in its own tab.

Ticket #18259 vu affichant l'état du bot
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

* accessible ou non

si oui:

* tâches en cours * tâches en attente

Version 1.3.0
-------------
:publication date: 2009/09/22
:expected date: n/a

Ticket #10556 "add apycot configuration" action is available to anonymous users
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.100
:state: resolved

The problem can be seen here: http://www.cubicweb.org/project/cubicweb/3.4.7

Of course, the user can't really submit the form, but still, the action
shouldn't be displayed.

Ticket #10000 integration with forge : show in a tab rather than at the bottom of main tab
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: resolved

Ticket #18056 store execution logs into a blob instead of one entity per log items
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.300
:state: resolved

too heavy write cost...

Version 1.2.0
-------------
:publication date: 2009/06/24
:expected date: n/a

Ticket #8403 add action to run apycot test
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

It could be useful to be able to launch directly an apycot configuration test
suite from the web frontend.

Ticket #9059 "apycotclient start" without arguments should start tests for  every activated config
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.200
:state: resolved

Ticket #8157 change apycot email title
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.250
:state: resolved

change

[APyCoT] status change in myproject

to

[mywebsite] myproject now has X failures and Y errors

or

[mywebsite] myproject now has one more failure

Ticket #7826 "changes" command should consider branches
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

changes occuring in a branch which has no configuration should not be printed
out

Ticket #9057 crash sur "apycotbot -D"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.100
:state: resolved

parce que host vaut None par défaut. Avec "apycotbot -D --host=localhost" ça
marche

Ticket #7868 distinguate between project dependencies and test dependencies
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

* project dependencies should always be check-outed (only when install pp needed
actually)

* test dependencies should only be check-outed for the tested configuration, not
when used as a dependancy

this would avoid extra environment setup cost, since now we may check-out /
install useless projects.

Ticket #8794 integrate with vcsfile
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

so we can start tests correctly according to new changes / branches

Ticket #8284 vcs_tag seems to not be honored
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: deprecated

check the apycotconfig http://www.logilab.org/projectapycotconfig/logilab-common
(in this example we try to jump to the branch default)

Comments ::

  On 2009/03/05 13:16 - Unknown author wrote :
  the apycotclient command with -t option is running fine however.
  On 2009/04/22 07:00 - sthenault wrote :
  my experience is that it works at least if vcs_tag is a tag, not a branch

Ticket #4335 don't (re)test a project if it wasn't modified
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

don't (re)test a project if it wasn't modified since the last test and if it's
dependencies haven't changed either.

Comments ::

  On 2008/06/23 17:39 - nchauvat wrote :
  with architecture change [#5324], check jobs have to be posted to a queue and
  can be posted only when a project or one of its dependencies are modified.

Ticket #3720 quick tests on commit
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

it would be nice to have a command that can detect that something changed in a
mercurial repository (use tip id?) and run limited (hence fast) tests on the
project to make sure that most important things are not broken (skip pylint,
doc, packaging, etc. but not unit or functionnal tests)

Comments ::

  On 2008/06/23 17:35 - nchauvat wrote :
  with the architecture change [#5324] and a queue where external scripts can post
  checks to run, one could have a simple script that adds unittest_check in commit
  hooks. or maybe have two queues: a fast one where to post requests for quick
  tests and a slower one for complete checks. the jobs in the slow queue would be
  processed when the fast queue is empty.
  > On 2009/02/03 13:53 - Unknown author wrote :
  > The problem is that the invocation of the test chain is asynchronous and the
  > (apycot)client always exists with success code. Maybe we could imagine a new
  > option to the client to wait till a valid result.

Ticket #9054 setting up apycot from scratch way to complicated
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

to get things easier:

* get back documentation which still make sense from the previous apycot project
* integrate documentation provided by email to didrik on the python-projects ml
* debian package should provide default config (in a separated ticket) * apycot
cube should ensure pyro-server=yes into the instance configuration * apycot cube
should provide inline documentation about what the various configuration fields
mean and what value(s) can be set * apycot cube should create a bookmark to the
latest executions report * rename some checkers to a more intuitive name:

- python_lint -> pylint - python_pytest -> pytest - python_check -> pychecker -
debian_lint -> lintian - debian_piuparts -> piuparts - debian_dpkgdeb -> dpkgdeb

probably more could be done...

Version 1.1.0
-------------
:publication date: 2009/05/18
:expected date: n/a

Ticket #5796 use color for report summary
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: deprecated

when displaying the summary of the report with the links, use the same colors as
the ones for the overall summary (failures in red, etc.)

Ticket #5083 debian_piuparts checker is not working
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 1.000
:state: resolved

the debian_piuparts checker is broken (ask for a password) and have unparsable
output.



Ticket #7352 /etc/init.d/apycotbot status
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: rejected

# /etc/init.d/apycotbot status /etc/init.d/apycotbot: line 96: status_of_proc:
command not found

Comments ::

  On 2009/01/20 19:11 - Unknown author wrote :
  We should maybe use the start-stop-daemon utility.
  > On 2009/01/21 07:24 - sthenault wrote :
  > already done. Only the status_of_proc function must be missing

Ticket #9184 update for cw 3.2
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.400
:state: resolved

Ticket #8937 wrong permissions on apycot conf files
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: rejected

at the end of the apt-get upgrading apycot I get ::

[snip] Bad group for maybe chgrp UNKNOWN './apycotbot-cw-sources.ini' Bad group
for maybe chgrp UNKNOWN './apycotbot.ini'

on all the files in /etc/apycot/

the `ls -alh` on it gives me ::

[snip] drwxrwxr-x 5 root 2001 4.0K Nov 24 10:49 apycot -rw-r--r-- 1 1004 1004 76
Jan 14 17:42 apycotbot-cw-sources.ini -rw-r--r-- 1 1004 1004 78 Jan 15 16:13
apycotbot.ini

no user with uid 1004 or group with 2001...

Comments ::

  On 2009/04/22 07:21 - sthenault wrote :
  isn't this a system administration bug, not a packaging bug ?
  > On 2009/04/27 06:38 - alutz wrote :
  > indeed, sorry I realized this was not an apycot bug, but forgot to remove this
  > ticket...

Version 1.0.7
-------------
:publication date: 2009/02/18
:expected date: n/a

Ticket #7365 using uid does'nt work
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

for instance with uid=nobody, the daemon does not start. We should probably
distribute the debian package with a config file and predefined permission
settings for the pid, log files (etc...) so apycotbot doesn't run as root

Comments ::

  On 2009/02/09 14:15 - Unknown author wrote :
  I will try with new user apycotbot and add new parameters to start-stop-daemon.
  On 2009/02/16 17:36 - Unknown author wrote :
  --pid-file parameter doesn't seem to be honored as well...
  > On 2009/02/17 08:09 - Unknown author wrote :
  > not honored or doesn't work as you expected?
  On 2009/02/17 08:08 - Unknown author wrote :
  where is uid=nobody passed?
  
  Have you tried the ``--chuid <login|uid>`` option to start-stop-daemon?
  
  Do you know which operation fails?
  > On 2009/02/17 09:24 - Unknown author wrote :
  > That's already pushed to tip. The new version of apycotbot-init is now running
  > as apycotbot user (and we use --chuid).

Version 1.0.6
-------------
:publication date: 2009/01/22
:expected date: n/a

Ticket #7175 different revision for a same apycot config should be allowed
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.100
:state: resolved

These second job should be added properly:

apycotclient start PROJECT -t stable apycotclient start PROJECT

But it is refused with the log 'already running test PROJECT'

Version 1.0.1
-------------
:publication date: 2009/01/15
:expected date: n/a

Ticket #7125 don't launch the same test twice in a very short time
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.100
:state: resolved

this may occurs if a test is launched as reverse dependency and explicitly at
(almost) the same time : when the second one occurs the first is already
launched to it's queued anyway.

Ticket #7119 fix buggy 1.0.1
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: resolved

make it works

Ticket #7121 missing command to stop the apycotbot daemon + init.d script
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

Version 1.0.0
-------------

architecture change
:publication date: 2009/01/14
:expected date: n/a

Ticket #6481 change pp_buildep for lgp integration
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 2.000
:state: resolved

Remove buildpackage python dependancy for a separated lgp process.

Ticket #5269 Improve error handling in apycot-runtest
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: deprecated

Errors occurring into runtests.py leads to missing test into the final reports.
Making them hard to detect unless the apycot-runtests trace are read. Such
errors should be reported with the other test result and appear in reports.



Ticket #5274 make sure temporary file are cleaned
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.100
:state: resolved

Every Test should create they temporary file and dirs into a common dir created
by runtests and remove when all tests are runned. It will avoid temporay file of
crashed test to remain forever in /tmp.



Ticket #2809 proper signal handling
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: rejected

at least in runtests

Ticket #6871 read/write from/to multiple cubicweb instances
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 2.000
:state: resolved

so config/test results are located in the same instance as the associated
project (eg intra/elo)

Ticket #5601 runtest.dependencies may hide a previous exception
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.200
:state: deprecated

this function use a "try/except ConfigError" which may hide the actual
configuration pb, giving a hard time to find what the actual problem is.

Ticket #6035 suppress linda dependency
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: rejected

Read http://bugs.debian.org/469039

Ticket #6923 restore notification using cw hooks
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.300
:state: resolved

when an execution fail while the previous one succeed.

Ticket #6921 use a specific user group
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

which will be only allowed to add apycot execution related entities

Ticket #2807 use spawn + wait instead of threads
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: rejected

rewrite runtests to stop using threads

Comments ::

  On 2008/10/29 19:25 - Unknown author wrote :
  spawn or subprocess ? Do we support window$ in apycot ?
  > On 2008/11/04 06:23 - Unknown author wrote :
  > we don't want to exclude windows of the platforms supported by apycot.
  > 
  > spawn or subprocess should be fine.

Version 0.99.0
--------------

aka apycotbot / cubicweb-apycot 0.1.0
:publication date: 2008/12/23
:expected date: n/a

Ticket #5324 change architecture
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 10.000
:state: resolved

change architecture to have one daemon running tests from a FIFO queue and have
scripts/commands add tests to the queue.

Ticket #2799 RDBMS storage backend instead of xml
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 3.000
:state: resolved

idea: directly store result in jpl

Version 0.12.3
--------------

Small bug fix.


:publication date: 2008/09/26
:expected date: n/a

Ticket #5981 pkg_doc report no data when all seems fine
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:state: resolved

The package doc checker status is "no data" for project which don't seem to have
doc issue.



Version 0.12.2
--------------

this version aim compatibility with lxml 2.1


:publication date: 2008/07/29
:expected date: n/a

Ticket #5660 lxml 2.1 incompatibility
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

lxml now raise ValueError instead of AssertionError. uncaught exception leads to
crash in the xmlwriter.



Ticket #5279 Unit test checker are too greedy
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

The python_unittest checker are parser *.*,*F*,*S* and *E* on the standard
output without any care of the context leading to incomplete trace.

Version 0.12.1
--------------

small bug fix removing a mis raised warning and incompatible minor options.
:publication date: 2008/06/13
:expected date: n/a

Version 0.12.0
--------------
:publication date: 2008/06/12
:expected date: n/a

Ticket #5127 add ignore python version option in python_unittest checker
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.500
:state: resolved

allow apycot to skip some version of python when running test.



Ticket #5129 force checker result to lower status
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

It may be good to cap the best status a checker can have for a project. This may
avoid success status on test caused by partial configuration.



Ticket #5110 Glob support for includes
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: resolved

add glob pattern support in the filename used by the "includes" field on apycot
configuration file.

Ticket #5112 rewrite runtests's test process creation
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.500
:state: resolved

we should use the 2.4 subprocess module to have more controll on the subprocess
created by apycot-runtests



Ticket #5123 separator overflow with timestamp.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

the 80 char length separator (--- or ~~~ or ...) are too long when timestamp or
prefix are activated.



Ticket #5055 stop using our own data model, use etree directly
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 3.250
:state: resolved

don't think we'ld loose much things and it would enhance speed and memory
consumption



Ticket #2781 subprocesses control
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 1.000
:state: resolved

Today if some tests are eating all the memory or never finish, apycot doesn't do
anything against that.

We need to have a timeout and RAM usage control on spawned subprocesses.

Ticket #5121 timestamp option for runtests
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: resolved

add a timestamps option to runtests to enable/disable timestamp in runtests
and/or in runtest.



Ticket #5114 Unit support for number
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

Some configuration field required time in second or capacity in byte. Being able
to specify the unit (eg "1.5MB" or "43m 33s") may ease configuration

Ticket #5042 add link to source repository in reports
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: resolved

in the reports, the left-most column is often a filename that could link to the
file in a browseable repository. sometimes there is even a line number.



Comments ::

  On 2008/05/21 14:48 - sthenault wrote :
  should be already in, though it's may be missing for mercurial

Ticket #5053 ease configuration
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.000
:state: resolved

should be easier to configurate, at least: * possibility to share part of the
test configuration * ease addition of new project to test



Version 0.11.0
--------------

monitor repo and run quick tests on changes.
:publication date: 2008/05/21
:expected date: n/a

Ticket #2793 only display pylint errors in reports 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.250
:state: resolved

ça pourrait être une option qui dit qu'on veut n'afficher que les erreurs du
signalées par pylint et ensuite afficher le tableau récapitulatif qu'on obtient
sous forme texte (msgXXXX: 4 occurences, etc.)

Ticket #2783 piuparts integration
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: enhancement
:load: 0.250
:state: resolved

to test debian packages

Ticket #4915 port to core xml python packages
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

The actual code uses old python xml packages. It should use recent modules.

In debian and ubuntu, the deb package python-xml will be removed_, and their
current version (sid & hardy) are "disabled" by default.

.. _removed::
https://launchpad.net/ubuntu/hardy/+source/python-xml/0.8.4-10ubuntu1

Ticket #4437 python_pkg looks for deprecated runtests.py
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

as far as I know, test/runtests.py is now deprecated, but apycot's python_pkg
tester still looks for it. isn't pytest the recommended solution ?

Ticket #2779 config and reports examples outdated
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
:type: bug
:load: 0.000
:state: resolved

Version 0.10.3
--------------
:publication date: 2006/09/25
:expected date: n/a