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

Re: Lintian based autorejects



On Tue, Nov 03, 2009 at 09:28:01AM +0900, Charles Plessy wrote:
> I reproduce here the one from Brian May, which I think is very
> relevant and was asked in many different ways, and always ignored:
> 
>     20091102084201.GA15187@microcomaustralia.com.au">http://lists.debian.org/msgid-search/20091102084201.GA15187@microcomaustralia.com.au
> 
>     I want to perform a NMU upload on a package, say to fix a security issue, or some
>     other major issue (as allowed by NMUs).

Actually, it's a couple of days that I'm pondering about that, trying to
make up my mind about that, I'll try only now to post that. The issue is
surely relevant.

On Mon, Nov 02, 2009 at 05:00:53PM -0800, Russ Allbery wrote:
> Surely the answer to that question is obvious: fix the bugs Lintian is
> finding that prevent upload.  They're the equivalent of RC bugs (not
> precisely the same, but similar), which if you're already doing an NMU are
> certainly fair game.

I don't think it is that simple. <nitpick>For once, we need to refine
some of our guidelines (that's the easy part). Devref §5.11.1 authorizes
to upload only changes that fix RC bugs older than X days, so if lintian
is complaining about issues not corresponding to RC bugs (e.g. because
it is a new check in lintian, not yet reported), in theory you shouldn't
upload with specific delays.</nitpick> (As I anticipated, that's the
easy part, just fix the guidelines.)

In general though, I wonder whether that would be the right
approach. Why should the NMUer be forced to fix other issues than her
own (RC) itch, considering that other (indubitably grave) issues were in
the archive _before_? The ideal solution would be to have dak know the
previous state and do not accept _regressions_ wrt the previous set of
fatal upload errors (according to the proposed wording).  I'm not sure
it is worth though, maybe Russ' solutions is the one NMUers would
implement anyhow and hence is overkilling to look for something more
complicated.

> This answer is independent of what we decide should go into that set of
> checks.

ACK.

Cheers.

PS Charles: I do believe I've limited myself to one post per day on the
   average, and at least one of them was a simple "thank you". Also,
   this thread have been rather acceptable, if you exclude the
   debate/flame on Manoj' bug filing. The fact that archive admins stop
   in participating is most likely related to their ongoing meeting, I'm
   confident they'll come back to some of the raised criticisms in the
   next few days; anyhow, the fact that they stop participating is not
   an argument for stop discussing among us "else".

-- 
Stefano Zacchiroli -o- PhD in Computer Science \ PostDoc @ Univ. Paris 7
zack@{upsilon.cc,pps.jussieu.fr,debian.org} -<>- http://upsilon.cc/zack/
Dietro un grande uomo c'è ..|  .  |. Et ne m'en veux pas si je te tutoie
sempre uno zaino ...........| ..: |.... Je dis tu à tous ceux que j'aime

Attachment: signature.asc
Description: Digital signature


Reply to: