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

Re: Integration of debian/ scripts in packages



On Thu, Jun 07, 2001 at 03:50:11PM -0800, Ethan Benson wrote:
> 
> huh?  maybe some packages are done that way, but not all, for some
> packages the diff has ONLY the debian directory which itself contains
> a patches subdirectory, those patches are applied at build time and
> removed at clean time.

Well, e2fsprogs certainly isn't done that way, and the New
Maintainer's Guide doesn't discuss this technique at all.

One of the reasons why I wasn't particularly happy with Debian a
number of years ago (and recommended a large number of people not to
use it) was when I discovered that a large number of changes were
being made to e2fsprogs, and they were all in one single .diff file,
so it was a real pain in the *ss to figure out what the heck was going
on.

When the .diff file gets that's big, it becomes to hard to track what
bugs are the upstream's maintainer, and which ones were introduced by
the Debian specific maintainer.  And when a bug introduced by debian
is forwarded to the author, either by the Debian maintainer because he
doesn't realize it was caused by one of his changes, or by end-users,
and when the upstream author has to paw through kilobytes and
kilobytes of patch file to try to figure out what's going on, it's
very easy for the upstream author to become.... irritable.

(Fortunately, this is no longer a problem, as Yann and I have managed
to resolve matters.  But it's made me determined to make the size of
debian .diff file tend towards zero, having been burned by this in the
past.)

							- Ted

P.S.  To be fair, I should point out that Debian isn't the only
distribution who makes changes to packages which become a huge
headache for the upstream.  The most recent distribution on my
sh*t-list was Mandrake, who changed the command-line UI of fsck, and
the worse yet, made their boot scripts depend on this non-standard
processing of fsck's -t option.  Did they document this anywhere?  NO.
Did they bother to inform me?  NO.

Do end-users of a distribution who try to upgrade to the latest
version of e2fsprogs complain to the distribution when things
mysteriously break?  NO, they complain directly to the upstream
author.  (Despite protestations of Debian maintainers who claim this
never happens, let me assure you, it does.)



Reply to: