* deduplicate addons paths, on runbot the tests are run twice and
indeed when listing test modules some are listed twice, maybe due to
duplicated addons paths?
Test modules:
openerp.tests.test_mail_examples
openerp.tests.test_uninstall
openerp.addons.web.tests.test_js
openerp.addons.website.tests.test_requests
openerp.addons.project.tests.test_project_base
openerp.addons.web.tests.test_js
openerp.addons.website.tests.test_requests
openerp.addons.project.tests.test_project_base
* flush stdout after having printed test modules list to try and
ensure it appears before the tests result (the TextTestRunner
outputs to stderr and this gets flushed a lot as there's a huge
amount of data going into it)
bzr revid: xmo@openerp.com-20131106134032-m85o6osa7ezgizt9
* make --fast-suite, --sanity-checks and --slow-suite exclusive at
argument parsing time
* split module directly in the argument parser
bzr revid: xmo@openerp.com-20131105135534-kzwph02kwxl77ivf
This is done automatically on a read(), but t-field uses a browse,
which does not do access-rights bypassing. We've already browsed the
root record so we should have the rights to the reading, and we just
want the textual representation for the child m2o.
bzr revid: xmo@openerp.com-20131025124720-i10eywz70o9c9dpc
check: verify the permissions even when no ids are passed (skipped permission checking for create)
create: verify has the write access on the related model (instead of create, was not checked anyway)
function field: execute the write in fnct_inv as superuser (was impossible to have creation without write access)
bzr revid: mat@openerp.com-20131030084408-t857gl7d4lkbrj5p
check: verify the permissions even when no ids are passed (skipped permission checking for create)
create: verify has the write access on the related model (instead of create, was not checked anyway)
function field: execute the write in fnct_inv as superuser (was impossible to have creation without write access)
bzr revid: mat@openerp.com-20131029171420-x87wu7ph8ej7mtro
This error probably stems from the useless complexity of
the SQL view declaration and the double GROUP BY levels.
The patch rewrites the view query with a single GROUP BY
level and proper aggregation levels, but the core part
of the patch is to replace the outer `group by product_qty`
with a `sum(product_qty)`.
Some columns were also mentioned twice in the same GROUP
BY clause, for some reason.
bzr revid: odo@openerp.com-20131025103626-7l78kdjjr7c2wesb
It's completely broken in case of optional parameters
e.g. (ir.translation)._get_source, as a call with the optional
parameter won't be matched by a clear without, and the other way
around. This becomes even more problematic in the website branch as
_get_source now has *two* optional parameters (source and res_id).
After discussion with odo and discovery that in multiprocess only the
current node will use any granularity (other nodes not only clear all
of the current method cache, but all caches of all models), simplify
cache clearing, ignore parameters and just blow the current method's
cache entirely.
bzr revid: xmo@openerp.com-20131024132956-4tl3prum8za47igy
Fixed length computation of text in html nodes: multiples successive whitespaces are considered as one whitespaces; better truncate position when adding a read more link; now always protect words (placed after the first word that exceeds the shorten position); pre nodes are preserved about whitespaces; when the read more link should go into a quote, it instead goes at the end of the first parent node not being quoted instead of at a wrong position.
Misc :
- removed an unnecessary loop by merging two root.iter() that do not interfere
- changed occurrences of getiterator to iter, because getiterator is deprecated
- nodes are cleaned after processing, removing all attributes used to tag nodes
- fixed a bug about cleaning span containing only whitespaces not always cleaned
Added tests for shorten position.
bzr revid: tde@openerp.com-20131024084411-pytwt9g1gdmoebzc
Xavier Morel