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

several package building glitches



I set aside last weekend to work on the repackaging of gmp into the
packages that I recently uploaded. During the transition I ran into a
couple of issues that I thought it would be good to bring to the group.

1.
Because the package(s) changed their names in the process (gmp_1.3.2
became gmp1_1.3.2 and gmp_2.0.2 became gmp2_2.0.2), as well as split into
runtime and development binary packages, I unpacked the source and renamed
the directory correctly. I then let dpkg-buildpackage put it back into the
tar file form. It did not seem that I could get away with simply renaming
the original source because of the new pathname. I can't see any way that
libraries that carry their so number in their name (within Debian) that
are not used by the original author's tarball can use the pristine source
features of dpkg. Have I missed something?

2.
When I first tried to unpack the old package source on my libc6 system I
got the error:

patch:*****can't find .dpkg-org

while, with the same packages on my libc5 system there was no problem
unpacking the source.

I did the obvious thing and upgraded to the latest patch from hamm and
things worked fine, as expected.

I know that Ian J. doesn't like the idea of adding dependencies to
dpkg-dev (although I don't quite see why that one package should have any
special dispensation), but this seems to be an obvious one. The new dpkg
incorporates a feature of patch that isn't available in the older version.
The maintainer who instituted that change should easily have realized that
it would require the "newer" patch program. Why isn't it a good idea to
make that new dpkg-dev package depend on the newer patch?

3.
This is the one that, I think, should be fixed. When you change the name
of the package in the changes and control files, if you don't delete the
debian/files file you get:

One warning about the old package name in files not being found in the
control file.

One error about the section in the control file not matching the section
in files.

Now, it seems to me that if changing the name of the package is only a
warning, that changing section or priority should only generate a warning!

I can not think of any data in the files file, that, when changed in the
control file, should cause a fatal error. The files file get's rebuilt
appropriately if it is deleted, so if non-existance isn't fatal, then why
should contradiction be so?

4.
I botched typing my "pass phrase" into pgp three times in a row once and
it got me to wondering how many times I might have been allowed to make a
mistake before pgp would dump me?

Waiting is,

Dwarf
-- 
_-_-_-_-_-_-                                          _-_-_-_-_-_-_-

aka   Dale Scheetz                   Phone:   1 (904) 656-9769
      Flexible Software              11000 McCrackin Road
      e-mail:  dwarf@polaris.net     Tallahassee, FL  32308

_-_-_-_-_-_- If you don't see what you want, just ask _-_-_-_-_-_-_-


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .


Reply to: