[Pkg-jed-commit] r683 - in jed-extra/trunk/debian: . examples patches

[Rafael Laboissiere]

* G. Milde <milde at users.sourceforge.net> [2007-05-15 11:14]:

> On 14.05.07, Rafael Laboissiere wrote:
> > $ ls -l jed-extra_2.3.2.orig.tar.gz
> > -rw-r--r-- 1 rafael rafael 459744 2007-05-14 19:37 jed-extra_2.3.2.orig.tar.gz
> 
> Strange, here, after building with utils/build-from-svn-copy.sh (which calls
> get-orig-source under the way) I get a different size:
> 
> pkg-jed/jed-extra/trunk > ls -l jed-extra_2.3.2.orig.tar.gz
> -rw-rw-r-- 1 milde milde 459751 2007-05-15 09:08 jed-extra_2.3.2.orig.tar.gz

This happens because the the tar format stores the user/group information
for each file.  The boxquote.sl file has "root/root" here but should be
something longer in your case.

> but the same md5sum 
> 
> > $ tar tfz jed-extra_2.3.2.orig.tar.gz | md5sum
> > 9790336509fecf6942f56c1b6263bdd6  -
> 
>   9790336509fecf6942f56c1b6263bdd6  -
> 
> is equal. This shows at least that the content is (most probabely)
> identical.

Note that "tar tfz" outputs just the contents of the tarball.  At least, we
got identical lists of files.

> > Should I go ahead and use this tarball?
> 
> I think so. I cannot verify truly independent, as
> jed-extra_2.3.2.orig.tar.gz is build from
> jedmodes-$(UPSTREAM_VERSION).tgz and
> http://www.cis.fu-berlin.de/~phgrau/slang/boxquote.sl
> 
> Do you think it helps to test the md5sum of jedmodes-2.3.2.tgz?
> 
> At the sourceforge shell (where I created jedmodes-2.3.2.tgz, I get
> 
>   #> tar tfz jedmodes-2.3.2.tgz | md5sum
>   80f0ea0528f26d8db1bb68b2288c2726  -

I added a line in debian/rules for outputing the md5sum of the upstream
tarball. Note that this is different from (and better than) the md5sum of
the output of "tar tfz".  It would be great if an md5sum file is added to
the download area at SF, such that an automatic check could be done.  (For
instance, Jed itself provides such a file, cf
ftp://space.mit.edu/pub/davis/jed/v0.99/MD5sums).

> > Another question: is it intentional that the distribution is "experimental",
> > or should we upload the package to unstable?
> 
> "experimental" is a paranoia setting. (We had problems with upload to
> unstable earlier as there were conflicts with the stable version, but
> this should be sorted out by now.)
> 
> If it is ok to upload to unstable, this would be preferable. 

Okay, will upload to unstable.

-- 
Rafael