#581 New tag handling changes XML formatting for empty tags

[khagaroth]

The new tag handling causes some unnecessary reformatting of the XML, which makes it practically unusable for files tracked with any version control system due to excessive diffs. For example

      <varlistentry>
        <term><userinput>--no-rcfile</userinput></term>
        <listitem>
          <para>Zakáže načítání inicializačních souborů.</para>
        </listitem>
      </varlistentry>

      <varlistentry>
        <term><userinput>--rcfile <replaceable>soubor</replaceable></userinput></term>
        <listitem>
          <para>Načte inicializační příkazy ze souboru <replaceable>soubor</replaceable>.</para>
        </listitem>
      </varlistentry>

becomes

      <varlistentry>
        <term><userinput>--no-rcfile</userinput></term>
        <listitem>
          <para>Zakáže načítání inicializačních souborů.</para>
        </listitem>
      </varlistentry> <varlistentry> <term><userinput>--rcfile <replaceable>soubor</replaceable></userinput></term>
        <listitem>
          <para>Načte inicializační příkazy ze souboru <replaceable>soubor</replaceable>.</para>
        </listitem>
      </varlistentry>

I don't have this problem in 2.6.

[Alex Buloichik]

Which file filter do you use ?

[khagaroth]

DocBook

[Alex Buloichik]

2Didier: looks like "Remove leading and trailing whitespace in nonsegmented projects" affects on it. But I'm not sure how space handling should be changed. Should OmegaT just skip spaces-only segments ?

[Didier Briel]

2Didier: looks like "Remove leading and trailing whitespace in nonsegmented projects" affects on it.

Do you mean even if the option is checked? To the original poster: is this option checked, or not?

Because, normally, if the option is checked, nothing should have changed (as far as I remember).

But I'm not sure how space handling should be changed. Should OmegaT just skip spaces-only segments ?

The general idea (in OmegaT) is that non-translated/translatable parts should be as in source as far as possible.

Didier

[khagaroth]

The option is checked.

Actually, after a recheck, 2.6 behaves the same, the only difference is that it doesn't touch empty tags, so in the above example it only touches and reformats <term> and <para>, but I didn't notice that because it's only obvious if they are multiline like this:

 <para>
   some text
   and some more
 </para>

becomes

 <para>some text and some more</para>

Don't know how feasible it would be, but I would expect any tag that is on its own line (even if it has content) to be left untouched, but that's more of a feature request than a bug.

[Didier Briel]

2.6 behaves the same, the only difference is that it doesn't touch empty tags

So, the bug would be only that 3.0 touches the layout of empty tags?

Didier

[khagaroth]

Yes.

[Didier Briel]

Fixed in SVN (/trunk) for the XML filters.

It was a side effect of #582 Remove leading and trailing tags option not working.

The behaviour still has to be fixed in the HTML filter.

Didier

[Didier Briel]

Fixed in the HTML filter.

Didier

[Didier Briel]

Fixed in the released version 3.0.1.

Didier