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

Re: Support on fixing bugs.



On Sat, Dec 12, 1998 at 10:28:04PM +0100, Martin Schulze wrote:

> > Ugh. I was working on a new release too. I wanted to send it to you - I will
> > not do an NMU as long as I am able to reach the original maintainer. I am not
> > that self-confident yet as I am a new maintainer ;-)
> 
> Hups?  I'm confused.  Phil, didn't you wrote to me just a few days ago that
> you were giving the ppp package to Torsten?

He told me that he will probably have enough time again in the future but he
would give me the ppp-package if I am using it heavily. I think he has still
better possibilities to test the package at an ISP...

> > I suggested Debian for a server in university and told them all the advantages
> > - along with other things of the BTS of course. Too bad they really looked at
> > it... They told me that they would not consider such a buggy distribution as
> > the operating system for a server.
>
> This is a simple-minded decision.  Please take a deep look at the bugs itself.
> Like Hamish already said, all major bugs are fixed already.  The bug tracking
> system contains a lot of bugs that are upstream bugs.  They are also present
> within other distributions.  They just avoid to let the user know.

WE know that but they do believe in what I say.

> Please subscribe to debian-bugs-dist and debian-bugs-closed and you will
> notice that a lot of bugs are being fixed.  There are also people working
> on longstanding bugs (e.g. like you ad Othmar, like myself who has fixed
> a bug which was about three years old but need a major redesign of the
> affected package).

I am subscribed to debian-bugs-dist and debian-bugs-closed.

> However you are probably correct that there are too many bugs.  But also there
> are a lot of bugs that are *very* difficult to reproduce, understand and to fix.
> Of course, there are also a lot of bugs that are easy to fix where the
> maintainer just lacks maintenance for his package(s).

This is exactly the problem I see. I think it does not really matter if we
have a minor bug which was inherited from the upstream source, but many bugs
are only packaging issues...

> > "we will provide an integrated system of high-quality, 100% free software".
> 
> This is true.  There is another phrase, too.  "We won't hide our bugs."
> I feel much better, if I *know* the bugs and are able to fix them myself
> if they get too important for me.  If I just notice a bug and can live with
> it for a while, I just report it, without a fix and without fixing it locally,
> just waiting for the maintainer to do his job.  Normally this is ok.

Right. But if I am a maintainer of a package I try to fix packaging bugs ASAP.
Often this are bugs as "the file foo belongs to path bar".

> Several other distributions just hide their bugs.  You can report them, but
> they won't get fixed.  Nothing is publicaly documented.  I'm sorry but I only
> have very little confidence for them.

AOL

> > So we (Othmar and I) thought we will simply do something to get some bugs 
> > tracked down and started to hunt some bugs. I don't remember why we focused 
> > on ppp in the first time but it does not really matter.
> 
> I guess it's because both of you need it to get on the net. :-)

Yes, but it is working perfectly. I did not notice all the bugs in using it :)

> Anyway, I highly appreciate this effort.  I would really love if I could spend
> some time as well, besides these package where I need a fix, improvement or
> whatever.
> 
> > I am hoping we can get an initiative which works on bugs of maintainers that
> > do not have enough time to cope with the load of bugs on their packages. Like
> > the people who support Linus in developing our kernel - the maintainer as a
> > benevolent dictator :)
> 
> I feel that initially the debian-qa list was started for this reason.
> Quality assurance is an important matter for any piece of software.  It is
> currently downgraded into a list flooded by the bug tracking system and more
> or less dead otherwise.  You may want to revive it for your needs.
> 
> This is the official description:
> 
>  Quality assurance is important for a distribution.
>  This list addresses this quality.

Hmm, interesting. 

> > I am not that kind of person who complains about something not working he got
> > for free. Instead I will try to change it to the better. That's why I like
> > free software - it's so much easier to patch a program you have the sources
> > for. I did not really have a problem with ppp - most of my problems are with
> > SuSE which we are using at university now :-(
> 
> Because it is bugfree since there is no public bug tracking system?
> 
> I'm not going to comment on this.

It everything but bug free. It has a shadow package without shadow support...
I am working on fixing bugs introduced by SuSE all day long. The package for
the kernel source had no include directory... Very funny!

cu
    Torsten

Attachment: pgpQX70q4baZy.pgp
Description: PGP signature


Reply to: