I guess this also affects flake8-docstrings then.

The test could be done in i18n repository but it seems no tests are done there but the bot verify with +2 together with CR+2

Doesn’t trigger test anymore

In T183085#4648389, @Zoranzoki21 wrote:

Status of that?

I think a FutureWarning should be placed first instead of the hidden deprecation warning.

These translations are used for pywikibot's i18n only selected by the Site.lang language code which is sr only. Thus the sr-el translations cannot be accessed. We could redirect sr-el translations to sr and drop the old sr translation but we cannot have both wich the current implementation. And I also don’t see why we should have it.

Seem this dependeds on the used browser. With Safari typing „pyw“ I get Pybal first and Pywikibot second.

In such case I wouldn’t expect the item when retrieving create protected pages from api or seeing that flag as page property anyway.

API queries:

In T89589#4633483, @D3r1ck01 wrote:

@Xqt, anything left?

In T89589#4633212, @D3r1ck01 wrote:

@Xqt, do we still have instances of u'' or u"" in the code base to remove? Or has this been complete?

In T89589#4633212, @D3r1ck01 wrote:

@Xqt, do we still have instances of u'' or u"" in the code base to remove? Or has this been complete?

I am wondering why this page can still be retrieved as create protected whereas Page.protection() of that page gives an empty dict.

The page was restored after protecting:
https://en.wikipedia.org/w/index.php?title=Special:Log&page=Lubo%C5%A1+Motl

In T205860#4630148, @Dvorapa wrote:

(Could both tasks be duplicates?)

Seems pydot is avaliable for Python 3 since 1.2
https://pypi.org/project/pydot/1.2.0/

Can’t see T151110 in the task graph above.

Cannot follow this removing?

@Dvorapa: why did your remove that parent task?

Yes as described there this would mixup my local repository, break dependencies and merged patches wouldn’t flagged as merged anymore. On my gui merges can be hidden. I guess there is a git command for it.

I don’t get the sense of this task. All api query iterators yield Page objects. The api tier resides above comm level and is called by Site objects.

This is not a Pywikibot issue but in scope of the bot operator

@Dvorapa: could you link that path for reusing it?

As result of the settings changes one test in Travis needs one hour now vs. two hours before.

They are created by a meta class during executing test scripts. There are variant keys mostly a site code and one created method is for one key/site only. This is a common way to have many small tests instead of one big on and it is easier to investigate into issues when a test fails.

This issue was already solved with https://gerrit.wikimedia.org/r/117689

I changed the Travis settings a bit to enable more concurrent jobs and cancel jobs when a new patch was deployed; this will shorten tests time a lot. I also cancelled some jobs but they can be restarted if necessary. Unfortunately I cannot do such administrative things on apveyor.

I've no glue whether that configuration matrix in .travis.yml inside i18n repository ever runs.