[Kbabel] my previous post

[Nicolas Goutte]

On Friday 25 November 2005 12:44, Stanislav Visnovsky wrote:
> Dňa Piatok 25 November 2005 12:31 Nicolas Goutte napísal:
> > On Friday 25 November 2005 10:08, Yury Tarasievich wrote:
> > > --- Nicolas Goutte <nicolasg at snafu.de> wrote:
> > > ...
> > >
> > > > > Well, what I feel should be done here is exporting
> > > >
> > > > of
> > > >
> > > > > complete DB to smth. like plaintext/csv, then
> > > > > restoring to scratch TM DB. After all, it came to
> > >
> > > ...
> > >
> > > > That is what db_dumb and db_load are for, as far as
> > > > I understand. (Except that
> > > > the dumb format is not simple plain text or csv.
> > >
> > > Right, and that's what I'm missing in kbabel compound
> > > more and more -- the ability to deal with TM contents
> > > directly.
> >
> > Yes, I have just had the idea of creating a Python script to dump the
> > databases in an useful format. (However that would probably not work if
> > the database is in an inconsistent state.)
>
> Maybe this way we might be able to detect the inconsistency. The problem
> with the bug is that you can see crash as a symptom of a broken database,
> but it's completely unclear when the corruption really  happens.

I would not hope too much.

If the database is inconsistent for the db library, it will give problems 
whatever it is called by Python or by C/C++.

I suppose that only db_dump could help to try to find the inconsistency 
(assuming that the mix between the dump format and how the data is organized 
by KBabel makes it at least half-way readable.)

Have a nice day!

>
> Stano
>
> > Python has it has Berkeley support as far as I have understood and also
> > such a script can be stand alone, so that an old version should be able
> > to run, even if KDE was updated, as it is the case here.
> >
> > However as many ideas, the problem is to find time to realize them. But
> > perhaps somebody else is interested in creating such a program.
> >
> > As for KBabel itself, sure, improving KBabel for interchangeable formats
> > is probably a very good idea, but again it is a problem of who will do
> > what. (KBabel for KDE4 has already a long list of things to do, at least
> > I find so for the handful of developers that we are.)
> >
> > > -regards
> >
> > Have a nice day!
> >
> > (...)
> >
> > _______________________________________________
> > kbabel mailing list
> > kbabel at kde.org
> > https://mail.kde.org/mailman/listinfo/kbabel