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

Re: package update lag on powerpc



>>>>> "DJ" == Daniel Jacobowitz <dan@debian.org>

Daniel,

    DJ> I feel like I've answered this question a thousand
    DJ> times....  There is a build daemon now, but only for
    DJ> potato, and I do not have time to fix all of its failures
    DJ> by hand.  There are also packages which build but present
    DJ> warnings which indicate serious problems, and I mark to
    DJ> look at later.  I've posted the list a few times.  No one
    DJ> else seems quite inspired enough to look at them, either.

I vaguely remember having seen such a list (on debian-powerpc, I'm
not on debian-devel), but it was quite a while ago.

I don't know about anyone else, but I would be happy to take a
look at such a list and try to figure out some of the problems.
At the very least, perhaps we could get some bugs filed against
the problem packages so that their maintainers could take a look
at them and try to fix them -- one of the problems I've seen with
a number of packages is that both the developers and the Debian
maintainers live solely in the Intel world, where their packages
compile perfectly, and tend to think you're crazy if you tell them
otherwise.

At the moment, it's difficult to tell what versions of a package
are available.  The Web site package search helpfully tells you
*only* about i386 packages.  Apt isn't very helpful for
determining this information, either -- all it tells you is that
there's an updated binary package available, not what version it
is or whether the latest source matches that binary package.
Console apt (or whatever it's called now) will show you the
version you have installed and the latest version available as a
binary package, but nothing about the source.  Comparing the
binary package's version with the version of the latest source
file requires you to look in the binary and source archives
manually.  There's certainly no way to know whether a binary
package isn't available because the builder (daemon) is slow or
the package has show-stopping bugs unless and until you download
the source and try to compile it yourself.

I'm not a coder, so I probably can't do much about any seriously
broken packages (e.g., freeamp, gdb), but I can certainly take a
look at some of the less broken ones and see if I can get them to
work.

Maybe a better solution than simply posting a list of problem
packages to a mailing list would be to maintain a Web page that
listed the problem packages, perhaps with links to the failed
build logs (I'm assuming that such things exist).  That way, an
interested person could take a look at the build logs and get an
idea of what might be wrong, which would make it easier to know
whether they might be able to fix the problem or not.

Another advantage of such a list (via the Web or by e-mail) is
that it would give people a very clear idea of why such-and-such a
package isn't available on PowerPC.  In any case, such a list
would only really be useful if it were updated frequently -- at
least once a week.

Given that potato is due for release pretty soon, and there are (I
suspect) many packages that don't compile on PowerPC yet that do
compile on Intel (and are therefore probably slated for
inclusion), it would be nice to be able to make some contribution
to getting a working potato ready for primetime on PowerPC.

What would be done about getting any necessary changes integrated
into the package is a somewhat separate question -- I imagine that
such patches could be sent to the appropriate Debian maintainers
for integration, or, failing that, could be made available on the
Web.


Just a thought...

   C.

+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
 Behind the counter a boy with a shaven head stared vacantly into space, 
 a dozen spikes of microsoft protruding from the socket behind his ear.
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
   C.M. Connelly               c@eskimo.com                   SHC, DS
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 


Reply to: