I translated IANA Time Zone Database too, because I tried googling for
my translation and the first results are about correct thing (even if
Google first suggests wrong spelling which (when clicked) suggests two
other wrong spellings…
[SKIP CI]
Fixes#783 (I hope). There is translatable string `$value` that I had
translated and I believe it's what causes that issue.
```diff
#: config.py:80
msgid "$value"
-msgstr "$arvo"
+msgstr "$value"
```
before
```
% travis lint .travis.yml
Warnings for .travis.yml:
[x] value for notifications section is empty, dropping
[x] in notifications section: unexpected key on_success, dropping
[x] in notifications section: unexpected key on_failure, dropping
```
now:
```
Hooray, .travis.yml looks valid :)
```
See also ProgVal/Supybot-plugins#217
Squashed commit of the following:
commit aa8823331ebccebdca83810545b167deb2e6303e
Author: Mikaela Suomalainen <mikaela.suomalainen@outlook.com>
Date: Tue Jul 15 20:47:29 2014 +0300
.travis.yml: remove duplicate x
commit ef3b363aa6b692c771733f1b4a9d352c9224f3d0
Author: Mikaela Suomalainen <mikaela.suomalainen@outlook.com>
Date: Sun Jul 13 22:32:10 2014 +0300
.travis.runonce.bash: add spaces
commit 004905729f954b5368510098fa2cac2a7181964a
Author: Mikaela Suomalainen <mikaela.suomalainen@outlook.com>
Date: Sun Jul 13 22:28:48 2014 +0300
.travis.runonce.bash: pipe msgcheck to true
commit d5bdeac5f7754c9443c983f374d41a2f9bc0d1b3
Author: Mikaela Suomalainen <mikaela.suomalainen@outlook.com>
Date: Sun Jul 13 11:04:40 2014 +0300
.travis.runonce.bash: remove RTFD
There is now hook for it at GitHub.
commit e1dab24f0f6f1c406b2284298a2e8ab3908bdc23
Author: Mikaela Suomalainen <mikaela.suomalainen@outlook.com>
Date: Sat Jul 12 23:34:14 2014 +0300
.travis.runonce.bash: care about exit status
This excludes msgcheck, because it's over-sensitive and always returns
exit value which logic I don't remember unless everything is correct
according to it.
Oh and the behaviour before this commit allows everything to fail
without having exit code which shows that it failed and we wouldn't
notice any failures in near future.
If I understand correctly, this makes the build fail when one build fails without having to wait for all builds to finish.
This might abort building for those which haven't finished yet, but we probably want to know that build fails as soon as one build fails.
