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

Re: mgetty 1.1.24 RPM (fwd)

Hi all,
I forwarded this from the mgetty list to try to figure out what the problem
is. Besides the obvious technical problems there seem to be MAJOR communication
problems between the redhat development team and the mgetty development
team. I have seen problems in the past and present (fixed with rawhide rpms)
with mgetty rpms.

A question for the redhat guys, Do you feed changes that are made to
packages back to the original development team? If so any idea what the
problem is here and if not why not?

Does anyone from redhat care about this?

......Tom			Cluelessness: There Are No Stupid Questions,
tdiehl pil net			But There Are LOTS of Inquisitive Idiots. :-)

---------- Forwarded message ----------
Date: Fri, 9 Mar 2001 14:47:19 +0100
From: Gert Doering <gert greenie muc de>
Reply-To: mgetty muc de
To: Heiner Kordewiner <organisationsleitung wohnzimmerrechenzentrum de>,
     mgetty muc de
Cc: Nalin Dahyabhai <nalin redhat com>, Preston Brown <pbrown redhat com>,
     Bill Nottingham <notting redhat com>, Than Ngo <than redhat com>
Subject: Re: mgetty 1.1.24 RPM from RedHat Question


On Fri, Mar 09, 2001 at 02:22:26PM +0100, Heiner Kordewiner wrote:
> >On Sun, Feb 25, 2001 at 05:05:10PM +0000, Rslo Linux wrote:
> >> After the installing of 1.1.24 RPM from RedHat rawhide section, most
> >> problems were solved. But I was wondering it mgetty is pre-compiled with
> >> "Auto_ppp" enabled?
> yes, but without FIDO...  ;-)

Which isn't used so widely anyway these days :-)

> |      375 Sep 19 20:02 mgetty-1.1.14-echo.patch
> |      532 Sep 19 20:03 mgetty-1.1.14-logrotate.patch
> |     2468 Sep 19 20:03 mgetty-1.1.14-policy.patch
> |      368 Sep 19 20:04 mgetty-1.1.21-giftopnm.patch
> |      289 Sep 19 20:04 mgetty-1.1.21-paths.patch
> |      392 Sep 19 20:04 mgetty-1.1.21-void.patch
> |     2036 Sep 19 20:02 mgetty-1.1.22-config.patch
> |     8143 Jan 15 18:11 mgetty-1.1.22-elsa.patch
> |      715 Jan 16 16:39 mgetty-1.1.24-faxprint.patch
> |      429 Sep 19 20:06 mgetty-1.1.5-makekvg.patch

I'm not sure whether all of them are really *used* these days.

> |    12991 Jan 16 16:40 mgetty.spec
> |   930400 Jan  8 17:13 mgetty1.1.24-Jan07.tar.gz
> Is there realy *no* way to syncronize the offical mgetty/vgetty and the
> "redhat-mgetty"?

It's very hard, as RedHat does not bother to even send their patches
into my direction.  They just change things, quite often *break* things in
the process, and then release something like "RedHat 7.0", leaving me
all the questions.  Yes, I'm still pissed.  Very much so.

> Maybe their config.patch ("-DAUTO_PPP") is essential ;-)

Well, that's policy, and I won't mind.  I ship without AUTO_PPP, but
that's just an option - turn it on or not.

> but I think (at least) makekvg.patch , elsa.patch (Marc?) , giftopnm.patch ,
> logrotate.patch and faxprint.patch should be applied to the *offical*
> release without damage?

I'm not exactly sure what's in those patches, but at least the elsa.patch
should be obsolete (as Juergen is continually working with Marc to make
the ELSA support even better - they just took a vgetty prerelease and
sort-of-merged it), and maybe also the giftopnm.patch (I assume that
it changes faxspool, but that could be done with faxspool.filter just
fine without changing code).

I will have to check the current set of patches, and complain again.

> And: Maybe if (all!) the Redhat-patches were less destructive (btw.
> "#ifdef"), they could make their way to the offical release and everyone
> would be happy?

Many of them are just plain broken.

> BTW: I could not find any Redhat-people on this Mailinglist are there any?

They do not care.  After all, people with mgetty problems will go to the
mgetty list, won't they?  So if they don't read it, they do not have to
fix what they broke.

> Following Redhats changelog from the last few months:
> |
> | * Tue Jan 16 2001 Nalin Dahyabhai <nalin redhat com>
> | - use mkdtemp() when printing faxes

I never got any notice about this.  But then, I do not really ship
anything that is officially used to print faxes, just sample apps.

> | * Mon Jan 15 2001 Preston Brown <pbrown redhat com>
> | - fix misdetection of USR voice modem detection <cjj u washington edu>

I never got any notice about this.  Which is especially bad as USRs *are*
nasty, and "fixing USR voice modem detection" (for one specific model)
can easily break others.  We have been through that a number of times,
with Marc doing changes for one model and breaking another.  So this
is a dangerous thing...

> | * Thu Sep  7 2000 Nalin Dahyabhai <nalin redhat com>
> | - make sure all scripts quote variables where possible (#17179)

This, for example, broke faxspool quite nastily.  Faxspool calls
ghostscript as 'gs $DRIVER'.  $DRIVER can actually be *two* arguments,
and their quoteing it made this 'gs "$DRIVER"' - which results in a
syntax error.

This wasn't even tested, just released and burned onto the 7.0 CD.

> | * Mon Aug  7 2000 Nalin Dahyabhai <nalin redhat com>
> | - fix excl patch to keep everything from segfaulting all the time (#11523,11590)

Which is pretty interesting: if their patch causes *segfaults*, maybe
they should have dropped it altogether?

> | * Thu Jul 12 2000 Than Ngo <than redhat de>
> | - add new V250modem patch from ELSA (thanks to Jürgen Kosel)

I'm not sure whether that was ever meant to go into an binary
distribution - making it behave differently than the official
mgetty version carrying the same number.

USENET is *not* the non-clickable part of WWW!
Gert Doering - Munich, Germany                             gert greenie muc de
fax: +49-89-35655025                        gert doering physik tu-muenchen de

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