Richard Dale is missing and generating QtC

[Germain Garand]

Le Vendredi 26 Juillet 2002 19:11, Adam Treat a écrit :
> Hello All,
>
> I am having a hell of a time trying to regenerate QtC.  I've discovered
> that Richard kept a patch file for manual edits that had to be carried over
> from release to release, but this patch file is missing along with Richard
> ;-)
>
> So, I've gone back to checking out kdebindings right after his last major
> update to QtC to make sure kalyptus hasn't changed, but when I regenerate
> the bindings some methods are missplaced... ie, all of the QFooEvent
> handlers are reordered and I can't make logical sense of why.  There are
> only three things I can think of that would have reordered them: I'm using
> different Qt headers than Richard used, Richard was carrying around these
> reorders in his patch file (I mean something could have changed in kalyptus
> long ago that changed the order of these methods and he's been carrying the
> old order in his patch) or there is some subtle nuance to generating these
> that I don't see.  The kdebindings I checked out are from 11/21/02, the day
> after he last regenerated them.  I'm using the headers from the libqt3-dev
> package in debian.  I wonder what headers he used...  I don't think I'm
> missing anything about the way in which qtc is generated.
>
> The goal here is to regenerate the bindings _exactly_ as Richard had done
> for the last major release and then obtain a patch file against qtc in
> current cvs, that way I'll have a patch of _only_ the manual edits which I
> can apply to future generations...  Any ideas would be most appreciated!
>

Hello Adam,

I think Richard used also some greping formulas on top of kalyptus...
I recall having seen a description of them once, but I can't find the mail...

Quoting him :

>> do you have scripts to e.g. generate Makefile.ams. You
>> often "regenerate the bindings", but I can't find the scripts to do that in
>> CVS - do you have those out of CVS?
>
>No at the moment I do it by hand, I've meant to automate it a bit more for 
>some time. I think the source code without any manual edits should go into a 
>different cvs branch called AUTOGENERATED_BINDINGS, HEAD_AUTOGENERATED or 
>similar. Then the manual edits are the diffs between this branch and the 
>current one.