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

Re: Debian bugs belong to the Debian BTS

On Wed, Oct 22, 2003 at 11:38:34PM +0300, Riku Voipio wrote:
> On Wed, Oct 22, 2003 at 02:53:01PM +0200, Adrian Bunk wrote:
> > it's your responsibility as a Debian maintainer to handle bugs in your 
> > packages.
> There is no way you can say to a volutantary worker "I demand to be
> served for free!" Ofcourse, technically closing upstream bugs is wrong.

This is the common "You can't force a volunteer to do anything." 
argument that is completely flawed:

Yes, you can't force a volunteer to do anything.

But you can say:
If you volunteer as a Debian developer, you have to either maintain your 
packages proper or give them away.

This is not meant as a personal offence against Chris, and as a former 
Debian developer I know that it's not always possible to handle bugs as
contemporary as they deserve.

> But it is just as wrong to only reopen them. 

Reopening only reinstates the status quo before the closure, and if the
closing was wrong, the reopening therefore can't be wrong. 

> You should have send them to the kde upstream Bug Tracking System 
> and mark them "forwarded to upstream" in debian BTS. Since you appear
> to be unwilling to help debian project even that small way, 

The bugs were closed with a message that handling these bugs that 
handling them is too much work for "only around 5 Debian maintainers", 
but you impute it's only "small way" to do this work.

> I've promised Chris to read through the bug list in weekend, and submit
> and mark upstream bugs to kde BTS.


That's great news.

> > It's _not_ OK to tell the users of your packages to only file them 
> > upstream.
> This is ludicrous. There is no advantage to anyone if obvious
> upstream issues pass through debian BTS. Packaging bugs
> to debian BTS, upstream bugs to upstream BTS. If unsure, 
> to debian BTS.

That's wrong.

Some examples:
- Chris closed even release critical bugs, a release critical upstream
  bug is a bug that should definitely handled properly before the next 
  stable release [1]
- there are packages in Debian that are orphaned for a long time (in 
  some cases up to 10 years)
- it might be caused by some Debian specific change and some upstream 
  maintainers refuse to handle bugs in modified versions of their 
  software (e.g. if you report on linux-kernel a bug against a 
  distrubution kernel you are often told to report it to the BTS of your 
  distribution - even people working for RedHat will tell you there to 
  report bugs in RedHat kernels to RedHat [2])
- the Debian maintainer usually knows better how to properly report a
  bug to upstream

An example that affects me personally:
- I'm doing some backports for Debian 3.0 and checking the Debian BTS
  is the only good way I know to get a quick overview over the current
  state of the package in unstable; leaving alone the many packages
  where the "upstream BTS" is an email address without a public archive
  (even XFree86 didn't have a BTS until recently), it would be nearly 
  impossible for me to search the BTS's of nearly 200 different source


[1] "handling properly" in some cases might simply mean to downgrade
    bugs, but closing RC bugs without checking them is definitely
[2] Altough the Debian kernel images are _far_ less modified than the
    kernel images of other distributions, there are some changes (and 
    e.g. security fixes require such changes). Herbert Xu usually does a
    good job in forwarding Debian bugs to upstream.


       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Reply to: