[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [Jackit-devel] Re: RFC: EXP merge of OSX portability work



Hello!

[Wed, 24 Mar 2004] Jack O'Quin wrote:
> They should be, and they usually are, but it's not guaranteed.  With
> autoconf, users need not have the full developer tool chain installed.
> So, it's important that all the front-end generated files be up to
> date.  Debian deals with these issues in other ways (build-depends,
> etc.), AFAICT.

Debian can regenerate them during build. Normally you prefer not to
do that. If you do it without ugly safety nets (like keeping tarballs 
inside tarballs), you pollute the diff with those regenerations 
which makes them unreadable and huge.

> Please don't upload 0.195.x to experimental.  Let's wait until it has
> been merged into CVS HEAD.  Right now, I just want to prove that we
> won't totally break the build and packaging system when we do that.

Ok.

> The current Debian `jackd' package is built for these platforms...
> 
>   alpha arm hppa i386 ia64 m68k mips mipsel powerpc s390 sparc
> 
> I've got code (from libstdc++) for...
> 
>   alpha cris generic hppa i386 i486 ia64 m68k mips powerpc s390 sparc
> 
> The `mipsel' is configured as `mips' for linux platforms (only).
> Also, the <atomicity.h> in `generic' is still broken, so there may be
> problems with `arm'.  What is that, anyway?  Does libstdc++ support
> it?

I don't know about these. Maybe the buildlogs for libstdc++ will shed
light.

> I'll use the version scheme you recommend.  I didn't know about that
> --compare-versions trick.  It also looks like this is true, which
> should be useful...
> 
>   $ dpkg --compare-versions 0.95exp.16 '<<' 0.95.0 && echo true
>   true
> 
> But, since we're still working in a CVS branch, direct linear
> comparisons with HEAD version numbers are not totally meaningful.
> Both branches lack some features found in the other.  There is no
> "partial ordering" in the mathematical sense of that phrase.  

Comparison with branching and merging points is always possible.

> That is why I'd prefer to avoid publishing anything from EXP, even in
> experimental.  It will confuse people.

Ok.

        Robert.

-- 
Obviously I was either onto something, or on something.
             -- Larry Wall on the creation of Perl

Attachment: signature.asc
Description: Digital signature


Reply to: