Re: [Simulavr-devel] SimulAVR Compilation Problem

[Bill]

Hmm...if you could point me to any documentation of being standard 
practice for GNU projects, that would be enough to make me reconsider. I 
honestly haven't researched it at all...

Thanks Paul. I knew it was confusing me and Klaus, so I jsut ditched the 
whole thing.

Do you have any experience with autotools? I've found them very 
frustrating to work with...the simplest things are not easy to figure 
out....either that's just the way it is or I'm going about it wrong. I'd 
really like a competent second pair of eyes on my build process....  I 
know, I'm fishing now ;-)

Also, I have a thick skin... I really do try to not be offended when 
intentions are good. (and try to not care when they are not ;-))

Paul Schlie wrote:

> > From: Bill <address@hidden>
> > Basically, it is(was) a real problem to keep the generated files up to
> > date. They are generated, and as a general principal, generated files
> > don't belong under revision control. It's my opinion that in general the
> > "tarballs" are intended for end-users, not CVS. Making CVS provided
> > files easy to use is a great goal, but I know that we fall short of it.
> >    
>
> Although I understand that maintaining correct configure/etc. updates in
> CVS adds some burden, however note that it's standard practice for all
> GNU projects that I'm aware of; and enables the maintainers to utilize
> whatever versions of autoxxx scripts and tools they choose, without
> unnecessarily burdening others with tool version idiosyncrasies (just as
> lex/yacc generate files are typically distributed for the same reasons)
>
> (please don't perceive the comment as a criticism, it's merely an
> observation that may actually reduce the longer-term burden on all)
>
> [although not typically necessary, why not simply run and update generated
> files after any patch update check-in, and automate it via a script?]
>
>
>
>
>