Re: Two theses regarding packages
On 10-Jun-01, 21:59 (CDT), Michael Stone <mstone@debian.org> wrote:
> I've thought about it. Maintaining build depends under the current
> system is a royal pain. The requisite dependencies aren't immediately
> obvious, the specific dependency line required to fill a particular
> dependency is often poorly documented, and it seems that the biggest
> problem comes with packages that, by rights, should be part of the
> default build environemnt. (E.g., perl)
I didn't mean to claim that it was easy for the maintainer to get them
right the first time. What I meant was that when someone reports a
problem with the build-depends, that report will include the information
needed to fix the build-depends. If it's an error along the lines of
"foo: no such file", then you do 'dpkg -S foo' to find the package
containing foo and add that to your build-depend. Often, the bug
reporter will have done that for you. Versioned dependencies need a
little more work, agreed, but I don't think anyone requires you to nail
the version exactly, and if they do, they should help you out with it.
If the bug report does not contain sufficient info to determing the
need package, and the reporter does not respond to requests for more
info, then close the report. But in actuallity, the most likely reporter
is one of the folks running a build daemon, and I've never gotten an
un-useful report from one of them.
The specific hypothetical involved someone *removing* a build-depends
line in response to a bug report, under the theory that lack of a
build-depends is non-RC. That is completely inexcusable.
Steve
--
Steve Greenland <stevegr@debian.org>
(Please do not CC me on mail sent to this list; I subscribe to and read
every list I post to.)
Reply to: