Wed, 21 Oct 2015 17:21:18 +0200 [postcreate] properly set is_instance_of for bootstrap entities
Julien Cristau <julien.cristau@logilab.fr> [Wed, 21 Oct 2015 17:21:18 +0200] rev 10803
[postcreate] properly set is_instance_of for bootstrap entities Entities created before CWEType itself need the is/is_instance_of relations created later. We were actually only inserting "is". Closes #7656606.
Wed, 28 Oct 2015 17:39:30 +0000 [migration] re-try without print_function when compiling a script fails
Julien Cristau <julien.cristau@logilab.fr> [Wed, 28 Oct 2015 17:39:30 +0000] rev 10802
[migration] re-try without print_function when compiling a script fails As part of the python3 porting we added the __future__.print_function import to cubicweb.migration. That started affecting all migration scripts it runs, which may not be prepared. So if compiling a script results in SyntaxError, try again without the print_function flag. Closes #7935030
Wed, 23 Sep 2015 15:32:17 +0200 [test] do not overwrite user's CW_CUBES_PATH in unittest_migration
Sylvain Thénault <sylvain.thenault@logilab.fr> [Wed, 23 Sep 2015 15:32:17 +0200] rev 10801
[test] do not overwrite user's CW_CUBES_PATH in unittest_migration Currently when you rely on your CW_CUBES_PATH environment variable to locate cubes necessary for cubicweb's tests, the migration test crashes because it overrides the env var.
Tue, 03 Nov 2015 16:00:34 +0100 [server/test] dict.itervalues() → iter(dict.values())
Julien Cristau <julien.cristau@logilab.fr> [Tue, 03 Nov 2015 16:00:34 +0100] rev 10800
[server/test] dict.itervalues() → iter(dict.values())
Wed, 16 Sep 2015 16:04:17 +0200 make sync_schema_props_perms(<computed rtype>) work as expected
Sylvain Thénault <sylvain.thenault@logilab.fr> [Wed, 16 Sep 2015 16:04:17 +0200] rev 10799
make sync_schema_props_perms(<computed rtype>) work as expected It currently ends up with an ExecutionError while we want to synchronize permissions. This makes 3.21.1 migration of any application using computed relation crashing. Also, remove pre 3.6 compat code in hooks that prevent living schema update on computed relation's permission changes.
Thu, 08 Oct 2015 13:51:54 +0200 [web/http_headers] use classmethod decorator
Julien Cristau <julien.cristau@logilab.fr> [Thu, 08 Oct 2015 13:51:54 +0200] rev 10798
[web/http_headers] use classmethod decorator
Thu, 08 Oct 2015 13:45:00 +0200 [schema] use classmethod decorator
Julien Cristau <julien.cristau@logilab.fr> [Thu, 08 Oct 2015 13:45:00 +0200] rev 10797
[schema] use classmethod decorator
Fri, 16 Oct 2015 16:44:36 +0200 [devtools] delete postgresql socket directory after killing the server
Julien Cristau <julien.cristau@logilab.fr> [Fri, 16 Oct 2015 16:44:36 +0200] rev 10796
[devtools] delete postgresql socket directory after killing the server ... or when failing to start it.
Mon, 27 Jul 2015 10:13:25 +0200 [skel] stop deleting cubes egg-info
Julien Cristau <julien.cristau@logilab.fr> [Mon, 27 Jul 2015 10:13:25 +0200] rev 10795
[skel] stop deleting cubes egg-info It doesn't hurt and provides some metadata.
Tue, 06 Oct 2015 15:06:45 +0200 [dataimport] Update doc to reflect new API
Rémi Cardona <remi.cardona@logilab.fr> [Tue, 06 Oct 2015 15:06:45 +0200] rev 10794
[dataimport] Update doc to reflect new API relate() looks deprecated.
(0) -10000 -3000 -1000 -300 -100 -10 +10 +100 +300 +1000 tip