Identifying messages in a Qt Linguist translation file (TS file)

Question

I hope this is the right forum to ask this question. If it is not, please redirect me to a more suitable forum.

I am trying to understand the format of a Qt Linguist TS file. I have modified a file and when I try to open it I get the error message "Duplicate messages found in ...". The file contains indeed one context with two messages that have

• the same source
• a location with the same file name but different line numbers
• different translation

Here is the source code of a minimal file that gives this error:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE TS>
<TS version="2.0" language="es">
<context>
    <name>MyContext</name>
    <message>
        <location filename="../../../../../MyFile.cpp" line="605"/>
        <source>Delete</source>
        <translatorcomment>Menu option</translatorcomment>
        <translation type="unfinished">&amp;Borrar</translation>
    </message>
    <message>
        <location filename="../../../../../MyFile.cpp" line="1572"/>
        <source>Delete</source>
        <translation>Eliminar</translation>
    </message>
</context>
</TS>

To my knowledge, each message is identified by context, source and location so this should not be a problem because the line numbers are different. Notice also that you cannot combine the two message elements into one, because they contain different translation texts.

So it is not clear to me what can cause the error message. I am using Qt version 4.6.2. I have looked in the following documents: Qt Linguist Manual: Translator, Qt Linguist Manual: Programmers, Linguist TS file format, but in none of them I have found the information I am looking for, namely how each message is identified and what could cause the error message in Qt Linguist.

If you do not know the answer to this question, a link to further, more detailed information on the TS format would also be helpful.

UPDATE

I have found out that by using the optional tag TS/context/message/comment it is possible to have different translation texts for the same context/message/source, e.g.

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE TS>
<TS version="2.0" language="es">
<context>
    <name>MyContext</name>
    <message>
        <location filename="../../../../../MyFile.cpp" line="605"/>
        <source>Delete</source>
        <comment>Comment 1</comment>
        <translatorcomment>Menu option</translatorcomment>
        <translation type="unfinished">&amp;Borrar</translation>
    </message>
    <message>
        <location filename="../../../../../MyFile.cpp" line="1572"/>
        <source>Delete</source>
        <comment>Comment 2</comment>
        <translation>Eliminar</translation>
    </message>
</context>
</TS>

The above file can be opened without errors.

This seems to solve my problem but it would be good to have some precise documentation about this format.

Answer 1

The error message reported by Qt linguist is meaningfull.

Try this modified version of your TS file:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE TS>

<TS version="2.0" language="es">
  <context>
    <name>MyContext</name>
        <message>
            <location filename="../../../../../MyFile.cpp" line="605"/>
            <source>Delete</source>
            <translatorcomment>Menu option</translatorcomment>
            <translation type="unfinished">&amp;Borrar</translation>
        </message>
    </context>
  <context>
    <name>MyOtherContext</name>
        <message>
            <location filename="../../../../../MyFile.cpp" line="1572"/>
            <source>Delete</source>
            <translation>Eliminar</translation>
        </message>
    </context>
</TS>

Duplicates are identified by context (i.e. MyContext vs. MyOtherContext) and source (Delete).

Edit:

I have scrutinized Qt Linguist's source code. As far as I can tell it never make use of ts.dtd for validation purpose, everything is hardcoded (See the TSReader class in ts.cpp).

The moral is: you can do every manipulation you want if and only if the resulting ts file can be loaded again by Qt Linguist and accepted by lupdate (to resync the application).

Answer 2

The complete lookup key for a translation is

Context + source string and, when present, disambiguation string

see http://qt-project.org/doc/qt-4.8/i18n-source-translation.html#disambiguation

The disambiguation string is stored in the comment field since this was also, in older Qt versions, the way to insert a developer comment (it is still recognized as such by Qt Linguist and so is the new comment (extracomment in the .ts file see http://qt-project.org/doc/qt-4.8/i18n-source-translation.html#translator-comments)).

So, as you noticed, you can have two strings with the same context and source string as long as they have different disambiguation strings (or one of them has one and the other does not).