Jeremy Collins <[EMAIL PROTECTED]> writes:
> I have already started to solve this by restructuring the ftp site.
> There should now be source packages for *ALL* pre-reqs. There are
> also RPMS for RedHat 5.2 based systems. Yes, this includes a
> version of Lesstif with Motif 2.0 compatibility compiled in, and
> also an XmHTML RPM built using this RPM. The current cvs Motif code
> compiled using this combination... so hopefully it should work for
> everyone else. There is also a Guile RPM. No SWIG RPM yet.
All of this should go a long way toward addressing Linas' first point.
> 1) Drop the Motif/Qt code. Maybe split it out into another source tree
> for anyone
> that feels the need to hack on it. Also say bye-bye to the XmHTML
> problems!
>
> 2) Drop the SWIG/Perl code. Rob has the guile/g-wrap code working
> good. Atleast he
> said so. This would also drop eperl as a pre-req.
>
> 3) Reimplement the build system. I believe Matt is working on this
> already.
I actually think that this is a really good plan with two
modifications:
1) freeze a "current snapshot" for people like me who are actually
using the motif GnuCash for real stuff (the Gnome version isn't
usable without some more critical work on the register). Do this
either via CVS tagging/branching (ideal) or via a tar file
(perhaps both).
2) Don't kill off the SWIG/perl stuff completely. Remove it from
the main source tree (physically or conceptually), and make sure
that you don't need it for anything we consider "standard"
(reports, etc). I.e. "make" doesn't need it, but keep it around
either as a separate CVS module, or as a subdir with it's own
"make" step so that a perl module becomes an option. Then
problems in there won't hold up the main app, and you won't have
to have the pre-reqs if you don't want them. This is a good way
to divorce these bits from the build process without abandoning
them entirely right now. If they wither and die after than in
favor of guile, then we can deal with that then, if not, then no
harm done.
Dropping all the UI's but gnome would really allow/force us to focus
on hammering the Gnome front end into shape.
> My reasoning for this is simple. Dropping the Motif code will resolve
> 90% of the current build problems as Lesstif and XmHTML are gone.
> Dropping the SWIG/Perl code will significantly reduce the pre-reqs as
> SWIG, and Eperl will be gone.
I agree.
If everyone else essentially agrees with this, and if Linas will
handle the branching, I'll generate a patch that implements these
changes immediately. (That or I'll work with Matt to generate such a
patch -- he and I can figure out which would work better. Depending
on how far along he is, it might be easier to build on his existing
work, or it might be better to just modify the existing tree and let
him merge with that.).
We'll still need to decide how to handle 2. Do we want a separate CVS
module? I tend to prefer that. It emphasizes the separation you
should find between a library and a wrapper, but it's not a big deal
to me either way.
--
Rob Browning <[EMAIL PROTECTED]> PGP=E80E0D04F521A094 532B97F5D64E3930
----- %< -------------------------------------------- >% ------
The GnuCash / X-Accountant Mailing List
To unsubscribe, send mail to [EMAIL PROTECTED] and
put "unsubscribe gnucash-devel [EMAIL PROTECTED]" in the body
