Now that the translation memory is so useful to give you matches for the current string, it would be even more useful if the plugin itself pre-translates the whole translation.

So given a file not 100% translated, after selecting this option in the menu, the strings that have any 100% concordance match should be added as a translation and mark the string as fuzzy.

In case of more than one 100% concordance I would not bother to show a dialog or anything fancy, as the string will be marked as fuzzy the translator will at some point go to that string, check it and see if a better match should have been used.

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=697067

When the TBX glossaries is done should allow to load multiple TBX glossaries from different files and de/activate them with a simple UI.

TBX glossaries could be specialized and sometimes the translator need to unload glossaries that knows that won't be used.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625812

As per HIG, "Context" and "Translation Memory" menu entries should are toggle menu items (mutually exclusive).

PS could you also consider to simply use "Memory" as label for this action?

Originally reported by Luca Ferretti at https://bugzilla.gnome.org/show_bug.cgi?id=569079

TMX is a very used and "de facto" standar for interchange translation memories based on XML files.

Gtranslator should allow to import and export TMX files into/from TM database.
You can get the standard specification from http://www.lisa.org/Translation-Memory-e.34.0.html. OmegaT has implemented native support for this if you need some references.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625785

Since GTK 3.10 GTK_STOCK are deprecated and thus need to be replaced.

Originally reported by Muhammet Kara at https://bugzilla.gnome.org/show_bug.cgi?id=776327

In .po files there is often obsolete strings, add feature which let's to delete these obsolete strings. I translate .po file with gtranslator and later i have to delete all unnecessary strings with text editor, gtranslator could do this job for me...

Originally reported by at https://bugzilla.gnome.org/show_bug.cgi?id=662501

Since the original strings are not editable, the replace dialog should only select the translated strings.

Plus, sometimes it doesn't select/highligh the expressions properly: this is, when there are at least two (only tested with two) strings to replace in the same msgid or msgstr, the second one is usually not selected (use the Search button).

Originally reported by Jorge González at https://bugzilla.gnome.org/show_bug.cgi?id=571713

There is a tool called gtxml (it cames with pyg3t, https://launchpad.net/pyg3t) that allows you to check tags from the console.

Since typos and errors in tags are so common, and the can cause some compilation errors, it would be very interesting if gTranslator could check tags syntax using this tool, like it checks variables or line breaks.

Thanks :)

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=672080

When translating documentation strings get quite lengthy and the TM panel becomes quite empty as the matches are rare.

If would be really nice to allow selecting a part of the msgid or msgstr and do a TM lookup for just that selected text, thus a string like [1] which obviously doesn't have any match on the TM, the user could select part of it, say "The table has two fields" and ask gtranslator to make a lookup just for that string instead of the whole string.

btw, nice work on gtranslator 2.90.x it works really nice!!

[1] "This demo is a simple GTK application (with a single window) capable of inserting records into a database table as well as browsing all records of the table. The table has two fields: id, an integer, and name, a varchar. The first section (on the top) of the application allows you to insert a record into the table. The last section (bottom) allows you to see all the records of that table. Its content is refreshed every time a new record is inserted and on the application startup."

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=655676

Build log:
http://pkgsubmit.mageia.org/uploads/failure/cauldron/core/release/20150623111327.fwang.valstar.13264/log/gtranslator-2.91.7-2.mga6/build.0.20150623111334.log

CCLD libgtranslator-private.la
./.libs/libgtranslator.a(gtr-view.o): In function gtr_view_enable_spellcheck': /home/iurt/rpmbuild/BUILD/gtranslator-2.91.7/src/gtr-view.c:270: undefined reference to gtk_spell_checker_detach'
./.libs/libgtranslator.a(gtr-view.o): In function gtr_attach_gtkspell': /home/iurt/rpmbuild/BUILD/gtranslator-2.91.7/src/gtr-view.c:79: undefined reference to gtk_spell_checker_new'
/home/iurt/rpmbuild/BUILD/gtranslator-2.91.7/src/gtr-view.c:80: undefined reference to gtk_spell_checker_set_language' /home/iurt/rpmbuild/BUILD/gtranslator-2.91.7/src/gtr-view.c:94: undefined reference to gtk_spell_checker_attach'
collect2: error: ld returned 1 exit status
Makefile:737: recipe for target 'libgtranslator-private.la' failed

Patch:
http://svnweb.mageia.org/packages/cauldron/gtranslator/current/SOURCES/gtranslator-2.91.7-linkage.patch?view=markup&pathrev=840557

Originally reported by Olav Vitters at https://bugzilla.gnome.org/show_bug.cgi?id=751380

The interface to import files into TM is very simple but useful. This should allow to import files matching a regex pattern.
Before import the files should appear a dialog showing the files that will be imported.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625786

Do/Undo doesn't work properly: they lose the context when you go to another string, and they should remember the past actions.

Originally reported by Jorge González at https://bugzilla.gnome.org/show_bug.cgi?id=549924

Yes, I'm here again, I'm a pain in the neck. sorry about that :P (working as a translator, I use a number of different CAT tools; so, every time I see something that could be useful for gtranslator, I report it. you can punch me if you want me to shut up, though ^^)

A "Lock Tags" option would be very useful; specially when translating documentation, there are a lot of tags that we don't have to translate; they distract us from the actual content we are translating and, besides, we may accidentally delete part of the tag. Currently, tags have different colors; that's great, but adding a Lock tags option would be even better.

In addition, a Validate tags option could be added. I know we can do a msgfmt in the console, but having an option in gtranslator would be very convenient (i.e. when I started translating, I didn't know about msgfmt!).

Originally reported by Sílvia Miranda at https://bugzilla.gnome.org/show_bug.cgi?id=514508

This tool needs to accept keyboard input not just ASCII or deadkeys but Chinese/Japanese/Korean/Vetnamese/Thai/Arabic/... They usually use Input Method (IM) which translate simple keyboard input into selected codes with complicated huristics.

Most of these IM UI on X/Linux platform uses following keys:

CTRL-SPACE or SHIFT-SPACE:
IM On/ IM Off (Via IM or Direct)

ARROW, CTRL-ARROW or SHIFT-ARROW:
IM input range selection and moving conversion scope

So these keys must not be used as SHORTCUT!

Popular Input methods which interfare with this are:

• IBus (New GNOME)
• SCIM (Old GNOME)
• uim

All these use CTRL-SPACE as normal on/off switch key.

Gtranslator's choice of using these keys for its internal UI is very bad choice.

I think the best thing is make these shortcut key choice configurable.

If you really need some arrows, please stay within ALT-ARROWS. PGUP/PGDN, ALT-PGUP/PGDN.

Also, F6-F10 are frequently used as special conversion instruction.

May I suggest something like:

COPY ORIGINAL MESSAGE: CTRL-B or CTRL-N

Next/Prev fuzzy/untranslated (most used) PGUP/PGDN
Next/Prev fuzzy ALT-CTRL PGUP/PGDN
Next/Prev untranslated ALT-SHIFT PGUP/PGDN
Next/Prev ALT- PGUP/PGDN
First/Last CTRL-SHIFT PGUP/PGDN

This seems reasonable choice ....

http://bugs.debian.org/586116

Osamu

Originally reported by Osamu Aoki at https://bugzilla.gnome.org/show_bug.cgi?id=622197

in the TM-field isn't seeable any difference between … and ... it's a little problem, ĉar i can see the difference in the editor-field.

i use 2.90.5

Originally reported by Kristjan SCHMIDT at https://bugzilla.gnome.org/show_bug.cgi?id=663445

It would be very nice to have a plugin that would

• connect to arbitrary Transifex intance
• search and automatically download the component .po files
• allowing to lock the desired component while doing so
• allow for automatic submission of modified .po's when the user clicks "Save" as an option (checking if there were new commits after the .po was downloaded)

Originally reported by Tomasz Chrzczonowicz at https://bugzilla.gnome.org/show_bug.cgi?id=594441

The interface to import files into TM is very simple but useful. This should allow to import files matching a regex pattern.
Before import the files should appear a dialog showing the files that will be imported.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625786

TMX is a very used and "de facto" standar for interchange translation memories based on XML files.

Gtranslator should allow to import and export TMX files into/from TM database.
You can get the standard specification from http://www.lisa.org/Translation-Memory-e.34.0.html. OmegaT has implemented native support for this if you need some references.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625785

Currently gtranslator does not allow going back to previous "touched" message. It is useful to be able to go back and forward between touched messages.

While you're only translating fuzzy or untranslated messages you end up passing a message while you know you need to turn back and fix/change something. However currently you can't do that, only neighbour messages can be iterated.

Originally reported by Baris Cicek at https://bugzilla.gnome.org/show_bug.cgi?id=552720

When translating documentation strings get quite lengthy and the TM panel becomes quite empty as the matches are rare.

If would be really nice to allow selecting a part of the msgid or msgstr and do a TM lookup for just that selected text, thus a string like [1] which obviously doesn't have any match on the TM, the user could select part of it, say "The table has two fields" and ask gtranslator to make a lookup just for that string instead of the whole string.

btw, nice work on gtranslator 2.90.x it works really nice!!

[1] "This demo is a simple GTK application (with a single window) capable of inserting records into a database table as well as browsing all records of the table. The table has two fields: id, an integer, and name, a varchar. The first section (on the top) of the application allows you to insert a record into the table. The last section (bottom) allows you to see all the records of that table. Its content is refreshed every time a new record is inserted and on the application startup."

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=655676

When the TBX glossaries is done should allow to load multiple TBX glossaries from different files and de/activate them with a simple UI.

TBX glossaries could be specialized and sometimes the translator need to unload glossaries that knows that won't be used.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625812

Lokalize highlight differences from source string in TM and actual string to translate. This is very useful to quick review the source string and evaluate to import from TM and fix the diferences or write the completed translation.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625788

There is a tool called gtxml (it cames with pyg3t, https://launchpad.net/pyg3t) that allows you to check tags from the console.

Since typos and errors in tags are so common, and the can cause some compilation errors, it would be very interesting if gTranslator could check tags syntax using this tool, like it checks variables or line breaks.

Thanks :)

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=672080

Lokalize highlight differences from source string in TM and actual string to translate. This is very useful to quick review the source string and evaluate to import from TM and fix the diferences or write the completed translation.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625788

Now that the translation memory is so useful to give you matches for the current string, it would be even more useful if the plugin itself pre-translates the whole translation.

So given a file not 100% translated, after selecting this option in the menu, the strings that have any 100% concordance match should be added as a translation and mark the string as fuzzy.

In case of more than one 100% concordance I would not bother to show a dialog or anything fancy, as the string will be marked as fuzzy the translator will at some point go to that string, check it and see if a better match should have been used.

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=697067

If the original strings contains a variable like %s, but user translates it as %d, GT doesn't show any warning message.

Also, if the linebreaks doesn't match in both original and translated strings, it doesn't warns about it.

Some commit migth have break this feature, since it was working in some previous versions.

Thanks!

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=682374

Seems that open-tran.eu is down and if you try to do a search with the plugin it crashes gtranslator.

It would be nice if instead of that, it just displays something telling the user that open-tran isn't reachable or the likes.

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=697006

The search dialog does not work fine. Sometimes it just moves to the next string no matter which is the search term.

Steps to reproduce the bug:

• Open a po file.
• Open the search window and introduce a quite long string (in my case: gbrainy.master.po and "Maximum number of stored games")
• Click Search and wait until the search has finished.
• It doesn't find the string (although it exists, first bug).
• Next time you hit Search in the same dialog, it just goes to the next string.

Originally reported by Jorge González at https://bugzilla.gnome.org/show_bug.cgi?id=610433

The plugin to see translations from alternative language hasn't a shortcut to copy the string suggested into our translation.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625789

Pology is a Python library and a set of command-line tools for processing PO files. The library aims to enable easy, fast and robust creation of scripts for tackling problems encountered in the "field", in everyday translation work. The tools perform various specialized operations, beyond that of other PO-handling software. Some of the tools are designed to (or have provisions to) treat a collection of PO files as an entity unto itself.

Pology allows to write per-language rules and batch review pofiles against those rules.

You can get pology from kde's svn repository, at the moment exists rules for a bunch of languages including Spanish, Galician, and others.

svn co svn://anonsvn.kde.org/home/kde/trunk/l10n-support/pology

Take notice that pology is only "the engine" that allow to execute rules.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=673355

Preferences entry menu is associated with gnome-shell menu. If user is running another desktop option like gnome fallback, unity or whatever, it's not possible to her to open the preferences options.

Originally reported by Pablo Sanxiao at https://bugzilla.gnome.org/show_bug.cgi?id=704870

While importing files to translation memory and after is completed should appear on dialog showing the progress or changes done at the data base.

After the import should be showed one dialog indicating a summary that contains:

• how many files were imported
• how many strings were imported
• little resume of the state of the data base:
  • total entries at database
  • strings before import

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625782

The application should save (in gconf/gsettings) the language being used in the text entries, so that the spell checker's language doesn't need to be set on every invocation.

However this relies on getting the info from gtkspell, and see gtkspell feature request #3015525.

Originally reported by Yaron Sheffer at https://bugzilla.gnome.org/show_bug.cgi?id=634168

Created attachment 224219
test .po file

gtranslator don't show symbol "\a". in attached .po file there is string "\a test \a \n \r \t" gtranslator understands "\n", "\r" and "\t" but don't understand "\a" simply shows... unicode number of it(as in screenshot). with all fonts same result.

Originally reported by at https://bugzilla.gnome.org/show_bug.cgi?id=683953

Seems that open-tran.eu is down and if you try to do a search with the plugin it crashes gtranslator.

It would be nice if instead of that, it just displays something telling the user that open-tran isn't reachable or the likes.

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=697006

If I translated an po before, it may be like this in it header

YunQiang Su [email protected], 2010, 2011.

Then if I translate it now, it will be

YunQiang Su [email protected], 2010, 2011

,2012.

The format of po file is broken. :(

Originally reported by at https://bugzilla.gnome.org/show_bug.cgi?id=673655

Pology is a set of scripts that include a script with the ability to perform bulk checks in translation files and to report each error in the translations.

It can tell Lokalize to open the files with entries that didn't passed any of the checks, but to only open the conflicting entries, and to show for each of these entries the error messages reported by the filters didn't pass. For example if a .po file with 1500 entries only has 83 entries which didn't passed any of the checks, Lokalize only shows these 83 entries, and for each of these 83 the messages reported are shown in the "Unit metadata" pane. If one entry doesn't pass three of the rules (or filters) it shows the three messages in the pane.

It is not only useful for the usual checks (punctuation, french guillemets,...) but also is useful for check that some english words have been translated to the target language using some specific words. For example "window" can be translated to galician like "ventá", "fiestra" or "xanela" since the three are perfectly valid, but in order to improve the consistency in the galician translation only "xanela" is recommended, so the Pology "check rules" sieve has a rule for this particular case.

I want to translate offline and then bulk check a bunch of files with Pology and get them opened in a CAT editor, in the conflicting entries with the error message reported in order to easy and fast correct the mistakes.

Can you make Gtranslator provide a public API with this functionality in order to integrate Pology with Gtranslator?

Pology can be fetched using: svn co svn://anonsvn.kde.org/home/kde/trunk/l10n-support/pology
The documentation is in the "doc-html" directory.

Originally reported by Leandro Regueiro at https://bugzilla.gnome.org/show_bug.cgi?id=640445

Since the original strings are not editable, the replace dialog should only select the translated strings.

Plus, sometimes it doesn't select/highligh the expressions properly: this is, when there are at least two (only tested with two) strings to replace in the same msgid or msgstr, the second one is usually not selected (use the Search button).

Originally reported by Jorge González at https://bugzilla.gnome.org/show_bug.cgi?id=571713

git branch of nacho

the "toggle fuzzy" icon of the corresponding button in the toolbar is missing

there is only text on the button

Originally reported by Christian Kirbach at https://bugzilla.gnome.org/show_bug.cgi?id=549634

When parser adds the current year to header comments, sometimes it generates a new line, starting with ",", and this breaks PO file structure.

This behaviour seems to be due to an 80 characters per line limitation. I've searched in the source code, but didn't find where this limit is set...

The file whre this comment is added to PO file is gtr-header.c, lines 682-690, but I think the problem is not there, but might help to investigate.

Could somebody please help with this issue? Please note that severity of this bug is high, because if PO file has a wrong structure, module's build might fail.

Thanks in advance

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=771765

Currently, the msgid and msgstr fields use a hardcoded default value unless a custom font is set. They should instead use the specified system font if available. gedit does this and its code could be used as an example.

Originally reported by Seán de Búrca at https://bugzilla.gnome.org/show_bug.cgi?id=550770

Shift+Ctrl+Page_Down and the corresponding toolbar button which finds the next uncertain or untranslated message, does not work.

Originally reported by at https://bugzilla.gnome.org/show_bug.cgi?id=732410

Since GTK 3.10 GTK_STOCK are deprecated and thus need to be replaced.

Originally reported by Muhammet Kara at https://bugzilla.gnome.org/show_bug.cgi?id=776327

Building fails in data directory for this reason:

• AppData requires to use INTLTOOL_XML_RULE to merge translations to xml
• gsettings does not merge translations to xml, so it should use INTLTOOL_XML_NOMERGE_RULE

this makes clean build to fail.

To resolve this issue, I suggest to move appdata somewhere else, I moved it to appdata subdir under data in Gnote.

Originally reported by Aurimas Černius at https://bugzilla.gnome.org/show_bug.cgi?id=720929

The plugin to see translations from alternative language hasn't a shortcut to copy the string suggested into our translation.

Originally reported by Fran Diéguez at https://bugzilla.gnome.org/show_bug.cgi?id=625789

When parser adds the current year to header comments, sometimes it generates a new line, starting with ",", and this breaks PO file structure.

This behaviour seems to be due to an 80 characters per line limitation. I've searched in the source code, but didn't find where this limit is set...

The file whre this comment is added to PO file is gtr-header.c, lines 682-690, but I think the problem is not there, but might help to investigate.

Could somebody please help with this issue? Please note that severity of this bug is high, because if PO file has a wrong structure, module's build might fail.

Thanks in advance

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=771765

As requested by Ignacio (http://bugzilla.gnome.org/show_bug.cgi?id=461021), I will explain what TBX is.

TBX (TermBase eXchange) is an open-XMl standard format for terminological data; has been submitted to ISO for possible adoption as a standard for terminology databases.

Its main advantage is the fact that it can be used as an interchange format; it is for termbases what TMX is for translation memories. Information stored in a termbase cannot be used in another termbase with a different data model; TBX is a data model standard, so it allows us to use termbases created with different applications. TBX termbases are basically xml documents - no mistery :)

Example of a TBX entry (taken from www.lisa.org):

You can see an example of a TBX file here: http://www.lisa.org/standards/tbx/tbxISO_final.html#exampleTBX

Why is it useful for gtranslator?

If it is implemented in gtranslator, it will allow translators to get termbases from a variety of sources and use them with the application. For example, if a GNOME translation team creates a termbase (with terms, definitions, context...), translators will be able to use that termbase within gtranslator.

I hope you find it useful!

For further information, and to see the format specifications, visit the TBX site: http://www.lisa.org/standards/tbx/
http://www.lisa.org/standards/tbx/tbxISO_final.html

Originally reported by Sílvia Miranda at https://bugzilla.gnome.org/show_bug.cgi?id=513662

Created attachment 212535
Partial patch. Just tell me if this is the correct format. If it is, I'll modify .h files according to it

We should unify indentation in .h files.

Originally reported by Daniel Mustieles at https://bugzilla.gnome.org/show_bug.cgi?id=674560

It would be nice (maybe, which probably will not be true, I can give a try to it) to have a plugin to check translations on MyMemory.net website.

They have an API http://mymemory.translated.net/doc/spec.php limited to 100 queries per day, or given an email address on each request, up to 1000.

Now that open-tran.eu is down, it would be cool to have a fallback for it (as usually open-tran will probably give you better terminology).

Originally reported by Gil Forcada at https://bugzilla.gnome.org/show_bug.cgi?id=697538

Using the "--previous" option from msgmerge, the generates PO shows something like this:

#: ../src/brasero-src-image.c:244 ../src/brasero-burn-options.c:308
#, fuzzy
#| msgid "Please, select another image."
msgid "Please select another image."
msgstr "Scegliere un'altra immagine."

Files generated for http://l10n.gnome.org are currently using this feature.

Could be a really good idea support this in Gtranslator 2.0, but honestly I've no stong or really good idea about the proper place to show the previous message.

One solution could be put an icon near "Original Text" label and show the previous message in a tooltip that appears when you mouse-over this icon.

Please note could be good remove the "previous" message from PO file on unfuzzy(+save), in order to keep clean the file (if you unfuzzy, you no longer need the reference to previous message, I suppose).

Originally reported by Luca Ferretti at https://bugzilla.gnome.org/show_bug.cgi?id=570902