[Gambas-user] gbc3-gb_error.o compiling rev #5670

Tobias Boege taboege at ...626...
Thu May 23 14:23:27 CEST 2013


On Thu, 23 May 2013, Beno?t Minisini wrote:
> Le 22/05/2013 23:58, Beno?t Minisini a ?crit :
> > Le 22/05/2013 22:27, Tobias Boege a ?crit :
> >>
> >> the funny thing was that it needed three commits to fix Emanuele's
> >> problem
> >> which is two too much for this kind of error.
> >>
> >> Actually, I'm not really concered about localisation (one of the most
> >> important lines in my .bashrc is LC_ALL=C :-)) and now my ignorance
> >> bit me.
> >>
> >> Basically, I added a simple line to the configure scripts so that the
> >> interpreter, when compiled from a subversion repository, prints the
> >> revision
> >> it was compiled from, like:
> >>
> >> $ gbx3 --version
> >> 3.4.90 #5672
> >>
> >> This number is extracted by a shell script line that makes assumptions on
> >> the language settings of the system - which is apparently
> >> non-portable. And
> >> that's the whole story.
> >>
> >> Regards,
> >> Tobi
> >>
> >
> > Ha ha. I have reworked completely the management of the TRUNK_VERSION
> > constant, fixing the localization problem and so on, but I forgot to
> > commit before leaving at noon!
> >
> > Now I have to check if what I have done is compatible with your
> > changes... :-/
> >
> 
> Sorry, I prefer my way of defining TRUNK_VERSION: you should not modify 
> the compilation flag for defining a constant. You'd rather use the 
> 'config.h' mechanism (AC_DEFINE & AC_DEFINE_UNQUOTED).
> 

But when is this file regenerated? Only by configure, correct? This would
mean that I have to issue a ./configure everytime I want the new correct
version number from gbx3 which is less desirable, in my opinion, than
changing a compiler flag.

Regards,
Tobi




More information about the User mailing list