> After an evening of chasing "ghosts", I think that I have found another
> incorrect type declaration in of the C include files, more precisely in
> the one that defines the ApplicationLaunchBlock:
>
> In alb.goh, the ALB_dataFile[] component ist declared as
>
> char ALB_dataFile[PATH_BUFFER_SIZE];
>
> while alb.def declares it as being
>
> ALB_dataFile FileLongName
>
> (which seems to be consistent with the way it is interpreted by the Geos
> system). As a consequence, the offsets of all the fields after that are
> misplaced in the C version of the file, which means that any initializations
> of the structure will be totally messed up...
>
> ciao marcus
Thanks for reporting this. This bug exists in the OmniGo SDK, but has
since been fixed: alb.goh now reads:
FileLongName ALB_dataFile
to match alb.def.
> my ceterum censeo (one of them ;-)): C/GOC and DEF/UIH files should be
> created automatically from a single source as far possible to make sure
> that the various interfaces are always kept in sync. See IBM's SOM for
> an idea how this could be done...
Yes, your idea has definite merit. I'll bring it up with Engineering,
to see if such a plan could be integrated into their current processes.
Regards,
Paul.
PS Can you define "ceterum censeo"? My Webster's 10th doesn't list it,
even in its "Foreign Words and Phrases" section, and to call my Latin
"rusty" would be hyperbole.