On Mon, 8 Jan 2001, Loic Dachary wrote:

> > How come? There are .spec files included that should work.
>
> ${prefix}/man/man1/* was incorrect on RedHat-7 since man pages are in
> the share directory. I changed this to ${_mandir}/man1/*. I was in the
> mood of doing some good to curl and spend an hour to do the modification
> properly. I used _libdir, _includdir, _bindir and _mandir where
> appropriate.

Ah. Redhat 7... If I'm just mistaking, that is more correct according to the
Linux Hierarchy Standard, isn't it?

> > The dist-version lacked no less than 26 entries (as listed with
> > 'tar -vztf <file>').
>
> I did not check that.

I was a bit too quick. It wasn't quite that many missing, I hadn't noticed
that two patches were rejected and after I corrected them the number of
missing entries was decreased significantly. Sorry,

> I just made sure that the release built had the necessary files to build
> a working distribution. Do you mean that make distcheck + make rpms does
> not work for you ?

I haven't tried it, I do most developing of curl on solaris without any rpm
tools.

> > Anyway, I'll commit most of your fixes and continue working on them until
> > they do fine. BTW, I build releases with the 'maketgz' script.
>
> I noticed this and I assumed you would continue to do so. I thought,
> however, that it would not do any harm and will help some people like me
> to be able to use the standard way to build distributions (make dist). If
> you want to use make dist rpms for the published distributions, I'd be
> happy to do the necessary work so that make dist exactly match what is
> built by maketgz. However, if you don't use make dist yourself the make
> dist will always be out of sync with maketgz since you won't update the
> Makefile.am EXTRA_DIST but only the FILES list.

I need to figure out a way to build release archives that work for me and
still keeps the 'make dist' things work. I will probably just convert the
script to make some fiddling first and then invoke 'make dist'. That should
make all of us happy! ;-)

A separate mail is coming on this subject, I feel we need to gather our
efforts.

> > Why do you need to rebuild curl?
>
> I don't have write access to the CVS tree, therefore when I apply a
> patch I do the following. patch, make test, make dist rpms, copy rpm to
> other machines, rpm -ivh to install the patched version, send the patch
> to the list, wait for it to be accepted, cvs update from time to time.
>
> Do you suggest another way of doing this ?

Hm, well, no. :-) Not if you really need to update those other machines using
RPMs.

> And BTW, thanks for your continuous efforts for providing curl ;-)

Thanks for helping me making curl better!

-- 
  Daniel Stenberg -- curl project maintainer -- http://curl.haxx.se/