When can a new release be expected?

Asked by Max Rees on 2018-05-21

The latest tagged revision is qpdfview-0.4.17beta1 from Nov 2016, 53 revisions ago.

Question information

English Edit question
qpdfview Edit question
No assignee Edit question
Last query:
Last reply:
Adam Reichold (adamreichold) said : #1

Hello Max,

your question is certainly valid and 0.4.17 was basically missed and it does not look any better for 0.4.18. The main problem is for me to find the time and motivation to do the tedious stuff like importing translations, providing tarballs and updating packaging scripts. So an honest answer would probably be as soon as sometimes actually steps up to help with these things. I am sorry that I do not have a better or at least more definitive answer, but there are always so many things to do..

Best regards,

Razi Alavizadeh (srazi) said : #2

Hi Adam,

If you want, maybe I can help at least with translations, but I need to know what converters did you use to convert "po" files to "ts" and "HTML" files?

Indeed, I tested "lconvert" and "po2html" but result was not good, maybe I used wrong arguments?

Best Regards,

Adam Reichold (adamreichold) said : #3

Hello Razi,

I have pushed the scripts I use for the conversion to a public repository here [1]. However, do note that there is still quite a lot of manual work needed after that since there are ambiguities when converting the PO to TS files that need to be resolved manually using Qt's Linguist tool.

Best regards,

[1] https://code.launchpad.net/~adamreichold/qpdfview/scripts

Razi Alavizadeh (srazi) said : #4

Hello Adam,

Thanks for sharing scripts.
Please see rev.2079 [1]. It's result of the following steps:
1- Call lupdate: (rev.2078)
2- Call convert_translations.sh: (a modified version)
3- Call lupdate: To remove big diffs related to changing line numbers

I think result is good. I'm not sure what do you mean by "ambiguities" that need manual works. If you mean somethings like what I did in rev.2080 [2] then they are issues with po files. The po files are not updated and also have some missing/bad translations.

Best Regards,

[1] https://bazaar.launchpad.net/~srazi/qpdfview/update-translations2/revision/2079
[2] https://bazaar.launchpad.net/~srazi/qpdfview/update-translations2/revision/2080

Adam Reichold (adamreichold) said : #5

Hello Razi,

the procedure you describe is correct (in addition you should call lupdate -no-obsolete in the last step), but the problem is that the translation contexts were not properly preserved and hence that second call to lupdate drops translation except where it can match them via the "same text heuristic". While this is often good enough, several languages will have dropped translations due to this, i.e. translations which are present in Launchpad and the PO files but missing in the resulting TS files after running lupdate again.

However, I just pushed a small Python script into the scripts repository to fix these up and testing this with a few translations seems to work quite well. But I would really appreciate if you could give it another try? Maybe we can finally try to make this an almost automatic process without recreating the above dropped translations manually using Linguist.

I would also be interested in any changes to the shell scripts that would help other - I mean specifically you - to work with these scripts since they have been single-purpose for my personal use until now.

I also just updated the translation templates, i.e. they were 11 texts in the source code which were not even present on Launchpad. The PO files themselves are automatically generated by Launchpad and should only be edited via its web UI or the translation upload functionality. Hence they could not be correct without me updating the templates...

Best regards,

Razi Alavizadeh (srazi) said : #6


Oops, I should apologize that I didn't send the modified convert_translations.sh in first place. Since, I also fixed the same issue. And indeed I made a CI to do conversion.

If I have been understood the work of python script correctly, I did its job with these sed commands (a little hakish :D):

sed -i "s/^\"X-Generator:..*/\"X-Qt-Contexts: true\\\\n\"/" ${source_dir}/${source_file}

sed -i -r "/^msgctxt \"[^Qq][^:]+\"/ s/^msgctxt \"([^Qq][^:|\"]+)|\"$/msgctxt \"qpdfview::\1/" ${source_dir}/${source_file}

To see the modified version of "convert_translation.sh" that also does some other cleanups see [1].
And [2] is link to github repository that uses travis to do automatic conversion. If you have github account then I can give you write access or you can fork it if you want. "convert_translation1.sh" uses your python script.

Best Regards,

PS: The sed command:
sed -i "s/<translation>/<translation type=\"unfinished\">/" ${dest_dir}/${dest_file}
is just to have a clean diff against current translations, but I suggest that we should remove it.

[1] https://github.com/srazi/qpdfview-translations-import/blob/master/scripts/convert_translation.sh
[2] https://github.com/srazi/qpdfview-translations-import

Adam Reichold (adamreichold) said : #7

Hello Razi,

I really like the idea of having Travis do these chores for us! I would also be happy for [1] to replace my scripts Bazaar repository. My GitHub account is "adamreichold".

Concerning the Travis manifest: Are you sure you require sudo? And ccache seems unrelated as well since we are not building any C/C++ code?

Concerning the script: It also fixes up the namespaces, but that is only a side issue. The main problem is that lconvert puts all messages into a single context named "" with the original context in the comment. The Python script fully rearranges the TS file so that each message is put into a separate context according to its comment. I really think we have to do this to not rely on the same text heuristic.

But I am not sure what "X-Qt-Contexts: true" does? Does it - after fixing up the namespaces using sed - have the same effect as the Python script? It does seem so? If that is the case, we can certainly drop the Python script.

Best regards, Adam.

[1] https://github.com/srazi/qpdfview-translations-import

Adam Reichold (adamreichold) said : #8

By the way, I do not think you want to include the commented-out translations from the PO as they are obsolete (as considered by Launchpad w.r.t. the latest uploaded template). Also I agree that we do not need to remove the "unfinished" marker...

Razi Alavizadeh (srazi) said : #9

Hello Adam,

> Concerning the Travis manifest: Are you sure you require sudo? And ccache seems unrelated as well since we are not building any C/C++ code?

Which sudo did you mean? I just used sudo when installing needed packages (travis workers are not login-ed into root user). Yes, ccahce is not needed. As I started with another travis script I forgot to clean up all things.

> But I am not sure what "X-Qt-Contexts: true" does? Does it - after fixing up the namespaces using sed - have the same effect as the Python script? It does seem so? If that is the case, we can certainly drop the Python script.

Yes. By adding "X-Qt-Contexts: true" to po files, lconvert uses msgctxt as context name (I inspired this from KDE world that also uses po files for all applications and I never heard that they have same issue with converting translations)

P.S.: I added you as collaborator.

Best Regards,

Razi Alavizadeh (srazi) said : #10

As you said the sudo was not required. I removed it and also, I cleaned up and refactored the travis script.

Adam Reichold (adamreichold) said : #11

Hello again,

so with Razi's help, we set up a fully automated version of importing the translations and I just did the first import. Hopefully I will the time to tag 0.4.18beta1 sometimes this week so we can do a two week translation freeze, wait for further bug reports and then release 0.4.18 to get those fixes and improvements out before having another look a what is then actually left in the bug tracker...


Can you help with this problem?

Provide an answer of your own, or ask Max Rees for more information if necessary.

To post a message you must log in.