Re: 2nd call for help
-----BEGIN PGP SIGNED MESSAGE-----
On 9 Mar 1998, Manoj Srivastava wrote:
> [IMHO, in no circumstances does it make sense to cross post to
> debian-private, since -private is meant for private stuff, and
> the orhers are not]
Sorry. Could someone PLEASE add a better description to the package
maintainers manual or to the mailing lists text file in /debian/doc? I
got the implication that devel was for general discussion on development
and private was for release-critical issues, the inaccuracy of which has
been made (pointedly) apparant to me several times now.
> How did the changes file get created? If dupload can't parse
> it, there is a good chance that there may be a flaw in the changes
> file. Could you please mail me the changes file (or send it to the
> list?)
The .changes file followed my .sig.
> Did you use dpkg-genchanges to create the file in
> debian/rules? (that is the standard procedure).
This is that funny package I asked about earlier, that got shoved into
Contrib because I couldn't separate the source and binaries without
breaking the package, and the format was odd enough that there was no way
for me to get debian/rules to compile it without manual intervention
before the code freeze. The source code and some minor binaries involved
in preparing the source code are in iraf-common. Precompiled binaries for
the primary computational system are in iraf-ibin. Precompiled binaries
for the NOAO project stuff are in iraf-noaobin. No changes were made to
the source. Binaries were stripped. Some FreeBSD stuff was deleted, as
well as a redundant copy of f2c. The documentation warns that some of the
files from iraf-common are actually used during execution, and until I
find out which I can't yank them out of the binary package. It's a couple
dozen megs of source code to examine (though it'll be a lot easier once
I've got all of the bootstrap and make-style scripts figured out).
Each slave .deb corresponds to a single .tar.gz file that on the
original site was cut into 512k chunks. In the -ibin and -noaobin
packages, the only changes made to the .tar.gz file before repackaging
were permissions changes and binary stripping to appease lintian.
Running a diff on it didn't make a whole lot of sense to me, unless
diff notes changed permissions in a way I wasn't aware of (it obviously
isn't going to be able to do much with the binaries).
The remaining system installs, removes, reinstalls and runs fine
on my system up as far as I know how to test IRAF (I'm unfortunately
not a good enough astronomer to properly stress-test it, and the
astronomers who inspired this effort are sufficiently crunched in
their own research that they aren't going to be able to help me for a
week or two, but I don't think I broke anything). I went hunting
through the policy manual and couldn't find anything on how to deal
with a binary-only release that happens to include its own source
code, so I improvised, hand-wrote a .changes file, packaged up a .deb,
ran it through lintian, fixed anything that looked wrong, sent
overrides requests to Mr. Schwarz and tried a dupload. That's more or
less where I am now, except that I've fixed up the .changes file a
bit.
[Bitch mode on]
This stupid package is starting to take over my life!
[Bitch mode off]
In theory, then, I should download each 512k chunk again, and
record them all as source files, and then try to get as much output as
I can out of a diff to match the packages? Or is this whole mess so
badly confused that at this point I just need to abort and work on
getting the compilable source separated and a debian/rules made for
release 2.1, whenever that may be? I'd been throwing my spare time
into this so far trying to avoid that, so I could go back to the
astronomy dept. around 2.0 release time and tell them, "See, Debian
even has the latest version of IRAF that you can just drop into your
system :)" and see if I ccould con them into putting up a few boxes :)
Still, if I've botched it, I've botched it. I was just hoping that
one of the experts out here could find a way out.
=============================================================================
Zed Pobre <zcp@po.cwru.edu> | PGP key on servers, fingerprint on finger
=============================================================================
-----BEGIN PGP SIGNATURE-----
Version: 5.0
Charset: noconv
iQEVAwUBNQRamdwPDK/EqFJbAQHPPwf/bxxPN9Lz+pV5jvEkZNDFKiNHGb0MkhkZ
hlk0tHskvLrb5cSK5iLU/QhaId813CAY7PlN6EVJd0GZv6hSZ3l5GE7yrf6u0uFY
u6U4QUuKFEMGsIrJU6qfdlvZCRb5deo31Cu+Lgm9Z4ABZOZV6DoFLCXOwFDSr7Ze
WQAzEVV3dDJ0uZeMc2HT3js+frx1bUO8yXfVupELk/CD4LrY/DCgDbQr6f36fUov
kwQ64pjr6K9qGVfAPFEQVfsEUKza5vLrpeudZBtl1FoTAG8Q4CngJGEFPLqt7jHa
hbY8+u1KWXGI8NmAXLBDhnXpQTZ6rUghJCDz4WCOueJaOiRTXmoJ0A==
=Hj0Z
-----END PGP SIGNATURE-----
Reply to: