[Harbour] Re: hbmk2 .po generation real life testing - speed issue

[Szakáts Viktor]

One thing I did is to do only one merge per build, as opposed
to one for each language, but that still leaves it quite slow,
especially in incremental mode.

Brgds,
Viktor

On 2009.05.06., at 19:12, Viktor Szakáts wrote:

> Hi Przemek,
>
> I'm testing .po generation in hbmk2. This project has more than 500
> source files with translatable texts, that equals to the amount  
> of .pots
> to be merged. There are 8 languages currently generated.
>
> What I noticed is that POTMerge() -> LoadPOTFiles() -> probably
> __i18n_potArrayJoin() gets slower and slower by processing .pot
> files which makes processing quite long, about 10 minutes on my
> machine.
>
> Do you think it's possible to speed that up? It may be an option to
> merge POTMerge() and AutoTrans() into one, as they are called next
> to each other in this scenario. ( That may also need reverse logic
> to retain existing translations. )
>
> You can test the same effect with hbi18n probably as above functions
> came from it. I can send you the .pot files if it helps.
>
> This is an issue even if we manage to solve incremental .po updates,
> as it happens for every fresh rebuild.
>
> Brgds,
> Viktor
>