CodeWarrior + Multi-lingual app support

[John.Holloway]

Is this yet another "feature" of codewarrior. - v3.1 ??

We have a perfectly correctly formed mmp file. We have specified 5 languages + have all the language text.

Built from the command line it all builds correctly. We can test as expected in the emulator. (Also builds + tested correctly on target - gcce)

Import mmp into Codewarrior + rebuild. It appears this decides that the target files for each of the different languages are the same file, specifically the .01 file - according to the <rsc targetfile="filename"> </rsc> xml generated in the .resources file.

When built from the ide it decides to stamp all over the .r01 (English) generated file with the last one in the list - in our case Italian. Looking at file creation dates it appears to do this after its compiled all the C code + before it generates the .exe

Is there a known fix for CW for this behaviour or is this yet another 'feature' of the ide we just have to live with ?

[mgroeber9110]

I can at least confirm this problem - it still occurs with the S60 3rd MR SDK, and appears to be related to the generation of the WINSCW.resource XML file for the CW resource compiler.

It looks as if all languages are listed under each target file, even though the name of the file is always taken from the first language in the list.

Our project currently builds 17 languages, and from the time that resource compilation takes (several minutes at full CPU load on a 1.6GHz machine!) I suspect that the full 17^2 variants are actually built and written over and over again into the same file.

As you write, it seems that in the end the localization for the *last* language in the LANG list ends up being written to the file whose language number comes *first* in the list. In our case, this means that the emulator opens the *.r01 file (English), but in fact gets Greek (54) resources.

The command line tools apparently deal with language lists just fine.

I am just going to use something like this for the time being...

#if defined(WINS)
# define LANG_LIST 01
#else
# define LANG_LIST 01 02 03 04 05 06 07 08 09 10 13 14 16 18 27 37 54
#endif

LANG LANG_LIST

[rburgst2]

my recommendation is to not use the mmp resource compilation at all and run epocrc.pl from the commandline. However, another solution would be to use 2 mmps one for CW and one for the command line. You can place #include into the mmp so you just need to code one mmp without resources (or only one) and then reuse (#include) the common stuff from a different file.

hope that helps