The strip-po target modifies files in place. This sometimes creates
conflicts with other make targets when make is run in parallel mode.
* split strip-po into strip-po and strip-pot
* move strip-po[t] from dependency to explicit, serial execution
* declare dependencies on POT/POFILES
* don't run strip on clean
Fixes: https://pagure.io/freeipa/issue/8323
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
The po/LINGUAS file contains a list of all avilable translations.
According to the GNU gettext documentation it's is a whitespace
separated list. Our LINGUAS file used newline separated list with inline
comments. This conflicts with weblate automation.
Fixes: https://pagure.io/freeipa/issue/8159
See: https://www.gnu.org/software/gettext/manual/html_node/po_002fLINGUAS.html
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Several translated strings were splitted into smaller ones. The older
translation either is a duplicate of the new one or does not apply
anymore.
Reviewed-By: Christian Heimes <cheimes@redhat.com>
1. Build po/ipa.pot every time we update PO files (each build)
2. Drop any rebuilt PO changes if the only difference is in the
translation file's header in a timestamp or timestamp+bug report
link.
3. Only apply the logic for dropping the changes if we are operating on
a git tree checkout because there is no otherwise an easy way to
detect the changes.
4. Hook strip-po target to the cleanup target to allow dropping unneeded
translation changes automatically.
5. Finally, strip ipaclient/remote_plugins/* locations from the ipa.pot
template. This saves us around 23,000 lines from the ipa.pot file and
reduces visual clutter in the translation files.
This approach allows to avoid unneccesary commits because even when
there are no changes to translation files, po/ipa.pot header would be
updated with a new translation update timestamp.
Fixes: https://pagure.io/freeipa/issue/8159
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Weblate tool sends pull requests that update translations directly.
For this to work, we need to keep ipa.pot in the tree.
Fixes: https://pagure.io/freeipa/issue/8159
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Make it possible to run API, ACI, and potests under Python 3.
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
The lower version is needed while building on RHEL.
Also po/Rules-quot file is deleted and added to .gitignore.
https://fedorahosted.org/freeipa/ticket/6418
Reviewed-By: Petr Spacek <pspacek@redhat.com>
All the source files are in the very same repo so there is no point
in keeping the file in Git.
https://fedorahosted.org/freeipa/ticket/6418
Reviewed-By: Martin Basti <mbasti@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Lukas Slebodnik <lslebodn@redhat.com>
The Makefile test targets were lost when gettextize infrastructure was
introduced. Now it is re-added in its modernized form which counts with
generated .pot files.
ipatests/i18n.py is now explicitly setting character encoding in files
it generates. According to gettext manual chapter "Filling in the Header Entry"
the Content-Type header is language-specific so it does not make sense
to fill it in in .pot file.
https://fedorahosted.org/freeipa/ticket/6418
Reviewed-By: Martin Basti <mbasti@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Lukas Slebodnik <lslebodn@redhat.com>
configure is easiest option how to automatically generate POTFILES.in.
Attempts to add it to po/Makefile* have big potential to create cyclic
depedencies and cause other trouble.
Given how rare operation adding a source file is, I think it is sufficient
to document that configure needs to be run again after adding a source file
with translatable strings.
https://fedorahosted.org/freeipa/ticket/6418
Reviewed-By: Martin Basti <mbasti@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Lukas Slebodnik <lslebodn@redhat.com>
