6:22 a.m.
Hi all!
FESCo this week finally decided the details how sub-groups like EPEL
should report their doings. In short: there should be weekly reports
that must be send to fedora-maintainers for public discussion. The
reports also get send to the (private) FESCo-list, to make sure FESCo
members don't miss the reports in the noise.
This EPEL report is the first one that follows this new scheme -- note
that I'll send it to epel-devel-list as well (thus three lists in
total!), to make sure non-fedora-contributors that might be interested
in EPEL are also aware of what happened and are able join the
discussions (they can't on fedora-maintainers, as that's a private list
open only for contributors). My preferred discussion ground for replies
to this reports would be epel-devel-list, as that's open to everyone --
but I assume some discussions in reply to this reports will happen
either on both lists, or just on one. You have been warned. (Side note:
I hope we find a better solution after the mailing list reorganisation,
which got discussed months ago, but was put on hold until we have new
hardware in place and F7 out the door).
And another note: In the past EPEL had reports that were meeting-centric
-- e.g. mainly the summary of the meetings that happen on Wednesday at
17:00 UTC in #fedora-meeting on freenode. The reports in the future will
likely be meeting focused with a summary of the meeting as well, but
I'll try to integrate some general notes about what happened in
EPEL-land over the last week added as well, so non-EPEL contributors can
easily follow all the important happenings.
Find the report below!
CU
knurd
----
= Weekly EPEL Summary =
Week 19/2007
== Most important happenings ==
* Axel Thimm left the EPEL SIG and the EPEL Steering Committee; see
https://www.redhat.com/archives/epel-devel-list/2007-May/msg00019.html
for details. The Steering Committee is looking for a replacement.
* the Steering Committee elected knurd as its chairmen; stahnma will
act as backup
* now that the RHEL5 rebuild is finished knurd will send a "if you
waited for a start signal to build your packages for EPEL this is it"
mail over the next few days to fedora-maintainers
* repotags -- some discussion in the meeting again. It looks like it
will we'll continue without repotags (final decision probably in next
weeks meeting, after this summary has been posted and discussed). If you
want repotags please *speak up now* and *help* to find a technical
solution that is not only fine for the EPEL Steering Committee, but also
acceptable for the Fedora Packaging Committee and FESCo -- from
discussions on list and on IRC it looks like that some members of those
groups tend to be against using repotags (see this weeks FESCo meeting
for example) or want to see something cooperation statements signed by
EPEL and 3rd party repos before they are willing to accept repotags.
In other words: it looks like it will be a whole lot of work to go for
repotags in EPEL -- some people might be willing to accept repotags if
someone does the wor out the details how to realize them in a way that
is acceptable for the different Committees. But it seems nobody is in
sight to do it. So if you want repotags and are willing to do that work
speak up now -- then there is a chance EPEL might get repotags in the
near future.
== EPEL SIG Meeting ==
=== Attending ===
From the Steering Committee: dgilmore (partially), knurd, mmcgrath
(partially), nirik, stahnma
Other contributors that joined the meeting: rdieter, smooge
=== Summary ===
* knurd was elected as chair by the steering committee in a wiki voting
over the past week; stahnma will act as a backup / kind of
"vice-president". See
http://fedoraproject.org/wiki/EPEL/SteeringCommittee/Voting for details
* thimm leaving / vacant seat in the Steering Committee : knurd will
mail those people that thimm suggested and ask them if they are willig
to become meber of the steering committee. Then we'll decide how to move on.
* mass rebuild for RHEL5 -- knurd did rebuild the remaining arch
packages over the weekend; issue solved. knurd will send a "if you
waited for a start signal to build you packages for epel then this is
it" mail to fedora-maintainers over the next few days . knurd and
dgilmore will work something out to make branching lots of packages easier.
* repotags/interaction with 3rd party repos -- nirik> "we can't keep
discussing it forever. " Some discussions to "put our feed down" and go
as planed without repotags. See "Most important happenings" for more details
* 00:30 and later; some discussions around commitment to EPEL4
stahnma> | is anyone concerned about the lack of package for RHEL 4?
It's seems like RHEL 5 has many a but clean up here and there and
RHEL 4 has few
knurd> | stahnma, well, we have to see how it evolves ; seems some
contributors only want to do RHEL5 and later
nirik> | there are probibly some things that won't work on rhel4,
since versions are old, etc.
See full log for all the details
* knurd did a some cleanup and additions to the EPEL docs in the wiki
and asks people to look over the changes
* QA process/testing repo/upload -- we need someone to look into that
(bodhi, push scrips) closer. Any volunteers?
=== Full meeting log ===
https://www.redhat.com/archives/epel-devel-list/2007-May/msg00031.html
== Stats ==
=== EPEL 5 ===
Number of source packages: 225
Number of binary packages: 373
Packages added over the past week:
* boolstuff | Disjunctive Normal Form boolean expression library
* php-pear-Structures-DataGrid-DataSource-MDB2 | DataSource driver
using PEAR::MDB2 and an SQL query
* php-pear-Structures-DataGrid-DataSource-RSS | DataSource driver using
RSS files
* php-pear-Structures-DataGrid-Renderer-Pager | Renderer driver using
PEAR::Pager
* pmount | Enable normal user mount
* python-Coherence | Python framework to participate in digital living
networks
* python-decoratortools | Use class and function decorators -- even in
Python 2.3
=== EPEL 4 ===
Number of source packages: 158
Number of binary packages: 312
Packages added over the past week:
* boolstuff | Disjunctive Normal Form boolean expression library
* python-Coherence | Python framework to participate in digital living
networks
7:45 a.m.
On Sat, 2007-05-12 at 13:22 +0200, Thorsten Leemhuis wrote:
Find the report below!
Overall, looks great. A few comments below.
= Weekly EPEL Summary =
Week 19/2007
== Most important happenings ==
* Axel Thimm left the EPEL SIG and the EPEL Steering Committee; see
https://www.redhat.com/archives/epel-devel-list/2007-May/msg00019.html
for details. The Steering Committee is looking for a replacement.
Unfortunate. We wish Axel the best.
* the Steering Committee elected knurd as its chairmen; stahnma
will
act as backup
Could we use Real Names instead of IRC nicks? I know who knurd is but I
don't know who stahnma is.
* now that the RHEL5 rebuild is finished knurd will send a "if
you
waited for a start signal to build your packages for EPEL this is it"
mail over the next few days to fedora-maintainers
* repotags -- some discussion in the meeting again. It looks like it
will we'll continue without repotags (final decision probably in next
weeks meeting, after this summary has been posted and discussed). If you
want repotags please *speak up now* and *help* to find a technical
solution that is not only fine for the EPEL Steering Committee, but also
acceptable for the Fedora Packaging Committee and FESCo -- from
discussions on list and on IRC it looks like that some members of those
groups tend to be against using repotags (see this weeks FESCo meeting
for example) or want to see something cooperation statements signed by
EPEL and 3rd party repos before they are willing to accept repotags.
Just some clarification. Yes, FESCo overall didn't see a good reason to
use repotags. If EPEL chooses to do so, FESCo won't stand in the way.
It would be unfortunate, however, if EPEL and FESCo packaging guidelines
diverged.
josh
8:10 a.m.
Josh Boyer schrieb:
On Sat, 2007-05-12 at 13:22 +0200, Thorsten Leemhuis wrote:
> Find the report below!
Overall, looks great. A few comments below.
thx
BTW, I put it into the wiki now, too.
http://fedoraproject.org/wiki/EPEL/Reports
> * the Steering Committee elected knurd as its chairmen; stahnma
will
> act as backup
Could we use Real Names instead of IRC nicks? I know who knurd is but I
don't know who stahnma is.
I tried that in the past and it makes stuff harder and time consuming to
write in my experience (I have the nicks from the meeting log already
and don't have to type them again or look out how peoples names are
spelled correctly). So I prefer to use nicks as writing summaries and
reports is boring enough already.
> * repotags -- some discussion in the meeting again. It looks
like it
> will we'll continue without repotags (final decision probably in next
> weeks meeting, after this summary has been posted and discussed). If you
> want repotags please *speak up now* and *help* to find a technical
> solution that is not only fine for the EPEL Steering Committee, but also
> acceptable for the Fedora Packaging Committee and FESCo -- from
> discussions on list and on IRC it looks like that some members of those
> groups tend to be against using repotags (see this weeks FESCo meeting
> for example) or want to see something cooperation statements signed by
> EPEL and 3rd party repos before they are willing to accept repotags.
Just some clarification. Yes, FESCo overall didn't see a good reason to
use repotags. If EPEL chooses to do so, FESCo won't stand in the way.
FYI: I was a bit against repotags in the past, but my position these
days after all this discussions is similar.
Or, to be more verbose: If someone works out the details on how to
realize repotags in Fedora then (depending on how the proposal looks
like) I'll likely abstain from a vote or might even support to use
repotags, as long as a simple "cp FC-6/foo.spec EL-5/" remains possible.
But I don't have the energy to work out the details. Anyone willing to
work them out?
It would be unfortunate, however, if EPEL and FESCo packaging
guidelines
diverged.
That what I want to prevent, too.
CU
thl
2:04 p.m.
On Sat, 2007-05-12 at 15:10 +0200, Thorsten Leemhuis wrote:
Josh Boyer schrieb:
> On Sat, 2007-05-12 at 13:22 +0200, Thorsten Leemhuis wrote:
>> Find the report below!
> Overall, looks great. A few comments below.
thx
BTW, I put it into the wiki now, too.
http://fedoraproject.org/wiki/EPEL/Reports
>> * the Steering Committee elected knurd as its chairmen; stahnma will
>> act as backup
> Could we use Real Names instead of IRC nicks? I know who knurd is but I
> don't know who stahnma is.
I tried that in the past and it makes stuff harder and time consuming to
write in my experience (I have the nicks from the meeting log already
and don't have to type them again or look out how peoples names are
spelled correctly). So I prefer to use nicks as writing summaries and
reports is boring enough already.
>> * repotags -- some discussion in the meeting again. It looks like it
>> will we'll continue without repotags (final decision probably in next
>> weeks meeting, after this summary has been posted and discussed). If you
>> want repotags please *speak up now*
I do. But maybe you are just asking for "I do"'s from list participants
that do not have a third party repo? Rationale for support has been
already hashed to death in many threads.
>> and *help* to find a technical
>> solution that is not only fine for the EPEL Steering Committee, but also
>> acceptable for the Fedora Packaging Committee and FESCo -- from
>> discussions on list and on IRC it looks like that some members of those
>> groups tend to be against using repotags (see this weeks FESCo meeting
>> for example) or want to see something cooperation statements signed by
>> EPEL and 3rd party repos before they are willing to accept repotags.
> Just some clarification. Yes, FESCo overall didn't see a good reason to
> use repotags. If EPEL chooses to do so, FESCo won't stand in the way.
FYI: I was a bit against repotags in the past, but my position these
days after all this discussions is similar.
Or, to be more verbose: If someone works out the details on how to
realize repotags in Fedora then (depending on how the proposal looks
like) I'll likely abstain from a vote or might even support to use
repotags, as long as a simple "cp FC-6/foo.spec EL-5/" remains possible.
But I don't have the energy to work out the details. Anyone willing to
work them out?
I presume this is internal help? (meaning how to tweak the build system
to support them transparently without any impact to the spec files
themselves?)
> It would be unfortunate, however, if EPEL and FESCo packaging
guidelines
> diverged.
That what I want to prevent, too.
-- Fernando
3:06 a.m.
Fernando Lopez-Lezcano schrieb:
On Sat, 2007-05-12 at 15:10 +0200, Thorsten Leemhuis wrote:
> Josh Boyer schrieb:
>>> * repotags -- some discussion in the meeting again. It looks like it
>>> will we'll continue without repotags (final decision probably in next
>>> weeks meeting, after this summary has been posted and discussed). If you
>>> want repotags please *speak up now*
I do. But maybe you are just asking for "I do"'s from list participants
that do not have a third party repo?
Actually I'm interested in opnions from everyone; but I'm especially
interested in opnions from Fedora/EPEL contributos.
Rationale for support has been
already hashed to death in many threads.
And others that don't see a sense or benefit in repotags (or even think
they do harm) have expressed their opinion as well.
>>> and *help* to find a technical
>>> solution that is not only fine for the EPEL Steering Committee, but also
>>> acceptable for the Fedora Packaging Committee and FESCo -- from
>>> discussions on list and on IRC it looks like that some members of those
>>> groups tend to be against using repotags (see this weeks FESCo meeting
>>> for example) or want to see something cooperation statements signed by
>>> EPEL and 3rd party repos before they are willing to accept repotags.
>> Just some clarification. Yes, FESCo overall didn't see a good reason to
>> use repotags. If EPEL chooses to do so, FESCo won't stand in the way.
> FYI: I was a bit against repotags in the past, but my position these
> days after all this discussions is similar.
> Or, to be more verbose: If someone works out the details on how to
> realize repotags in Fedora then (depending on how the proposal looks
> like) I'll likely abstain from a vote or might even support to use
> repotags, as long as a simple "cp FC-6/foo.spec EL-5/" remains possible.
> But I don't have the energy to work out the details. Anyone willing to
> work them out?
I presume this is internal help? (meaning how to tweak the build system
to support them transparently without any impact to the spec files
themselves?)
No, more meaning: Find a technical and political solution that is
acceptable for EPEL SIG and the Fedora Packaging Committee (and FESCo,
if it wants to ).
And let me warn you: there seems to be a opposition against repotags on
that way from what I can see, so this process might take weeks and
likely lots of discussions and mails. There are high chances to get
frustrated and burned out on the way afaics.
Actually realizing what got decided in the end is probably easy.
CU
thl
11:50 a.m.
[sorry for the delay, maybe too late but here it goes anyway - was way
too busy then and I'm now in the middle of a trip visiting family so
email access is sporadic...]
On Sun, 2007-05-13 at 10:06 +0200, Thorsten Leemhuis wrote:
Fernando Lopez-Lezcano schrieb:
> On Sat, 2007-05-12 at 15:10 +0200, Thorsten Leemhuis wrote:
>> Josh Boyer schrieb:
>>>> * repotags -- some discussion in the meeting again. It looks like it
>>>> will we'll continue without repotags (final decision probably in
next
>>>> weeks meeting, after this summary has been posted and discussed). If
you
>>>> want repotags please *speak up now*
> I do. But maybe you are just asking for "I do"'s from list
participants
> that do not have a third party repo?
Actually I'm interested in opnions from everyone; but I'm especially
interested in opnions from Fedora/EPEL contributos.
> Rationale for support has been already hashed to death in many threads.
And others that don't see a sense or benefit in repotags (or even think
they do harm) have expressed their opinion as well.
Yes, of course, I did not mean only supporting opinions had been
expressed, that would make this a no-brainer :-)
My personal, biased and most probably incomplete reading of the opposing
opinions include "I wish this would just go away"[*], "this introduces
problems in ordering of packages between repositories", "we are being
forced to do this and we don't want to", "it is a bad solution to the
problem", "it would be a problem to set this up in the epel build
system" and "political problems in consistence between epel and the rest
of fedora". I think most if not all the technical points have had
reasonable technical responses from the proponents (and long time users)
of repotags.
[*] I'm _not_ trying to flame, I actually read this in one of the irc
sessions and I think several emails as well.
>>>> and *help* to find a technical
>>>> solution that is not only fine for the EPEL Steering Committee, but
also
>>>> acceptable for the Fedora Packaging Committee and FESCo -- from
>>>> discussions on list and on IRC it looks like that some members of those
>>>> groups tend to be against using repotags (see this weeks FESCo meeting
>>>> for example) or want to see something cooperation statements signed by
>>>> EPEL and 3rd party repos before they are willing to accept repotags.
>>> Just some clarification. Yes, FESCo overall didn't see a good reason
to
>>> use repotags. If EPEL chooses to do so, FESCo won't stand in the way.
>> FYI: I was a bit against repotags in the past, but my position these
>> days after all this discussions is similar.
>> Or, to be more verbose: If someone works out the details on how to
>> realize repotags in Fedora then (depending on how the proposal looks
>> like) I'll likely abstain from a vote or might even support to use
>> repotags, as long as a simple "cp FC-6/foo.spec EL-5/" remains
possible.
>> But I don't have the energy to work out the details. Anyone willing to
>> work them out?
>
> I presume this is internal help? (meaning how to tweak the build system
> to support them transparently without any impact to the spec files
> themselves?)
No, more meaning: Find a technical and political solution that is
acceptable for EPEL SIG and the Fedora Packaging Committee (and FESCo,
if it wants to ).
And let me warn you: there seems to be a opposition against repotags on
that way from what I can see, so this process might take weeks and
likely lots of discussions and mails. There are high chances to get
frustrated and burned out on the way afaics.
Yeah, I understand that. No offense people, but after reading some of
the recurring threads I was ready to go jump out of the nearest window.
Given the previous threads I personally don't have much hope that
repotags will be ever accepted by the Fedora community and regretfully I
don't have any magical political solution to the problem - it does look
to me that most of the objections are political in nature.
-- Fernando
1:02 p.m.
On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
Yeah, I understand that. No offense people, but after reading some
of
the recurring threads I was ready to go jump out of the nearest window.
I hope that you live on the ground floor :)
Given the previous threads I personally don't have much hope
that
repotags will be ever accepted by the Fedora community
be careful: don't confuse the whole community with a couple of people
sitting behind the steering wheel. fpc had signaled that they would
pass any decision of epel to introduce repotags and fesco couldn't
care less, they don't like the idea of repotags (especially from their
non-rhel pov), but they wouldn't veto any such decision either.
In fact I would argue that the Fedora and - what's more important in
this context - the RHEL community were quite accustomed and happy with
repotags.
But repotags have a very week point: It takes one player to taint the
system.
and regretfully I don't have any magical political solution to
the
problem - it does look to me that most of the objections are
political in nature.
Well, for what it's worth the repotag issue was never introduced in any
non-political way. It is the way that repos used to visually mark
their packages like marking cattles. They don't produce more or less
milk either, but when a cattle breaks free and stampedes your meadow
to dirt you know what farmer to blame.
Same for packages. It was never a "technical" issue that Dag's repotag
"dag" was higher than ATrpms' "at". Neither that "rf"
was higher that
"ccrma" and so on. These "technical" arguments are rather lame, and
if
the people having these arguments really belived in them, they would
in fact want a repotag of "zzz_my_repo". There are no technical
benefits or drawbacks with repotags.
So it started by politics (a common request of many existing and to
become software repos for RHEL) and ended with politics (epel
effectively rejecting to play along). And even hiding behind being a
homo faber, e.g. ignoring all but technical sides of any issue, is
already a political statement. There is nothing wrong with politics
per se, it would just had been nice if politics would had been layed
out into nice inter-repo cooperation.
Instead we now have the same political rift that Fedora itself had 3
years ago. Fedora's rifts are still not completely healed, let's see
how long it will take until epel vs the rest of world rifts start to
heal. It is just sad to see the exact same mistakes recurring
again. And even sader it is to be in the middle of it again.
Anyway, Nando, as much as I personally liked and promoted repotags,
it's currently more promising to beat a dead horse than to get epel to
consider repotags, after all epel already voted on this and nothing
has changed since. So I wouldn't invest too much time into this, if I
were you, because you will just burn out as the rest of us. I know I
burnt out now for the umpteenth time. :/
--
Axel.Thimm at ATrpms.net
1:06 p.m.
On Fri, 2007-05-18 at 20:02 +0200, Axel Thimm wrote:
On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano
wrote:
> Yeah, I understand that. No offense people, but after reading some of
> the recurring threads I was ready to go jump out of the nearest window.
I hope that you live on the ground floor :)
Don't worry, I can hopefully resist the rige :-)
> Given the previous threads I personally don't have much hope
that
> repotags will be ever accepted by the Fedora community
be careful: don't confuse the whole community with a couple of people
sitting behind the steering wheel. fpc had signaled that they would
pass any decision of epel to introduce repotags and fesco couldn't
care less, they don't like the idea of repotags (especially from their
non-rhel pov), but they wouldn't veto any such decision either.
Sorry, I did not mean to do that, there have been voices of support, of
course. But those opposed do speak for the community (whatever that is)
and they have the wheel.
In fact I would argue that the Fedora and - what's more important
in
this context - the RHEL community were quite accustomed and happy with
repotags.
But repotags have a very week point: It takes one player to taint the
system.
> and regretfully I don't have any magical political solution to the
> problem - it does look to me that most of the objections are
> political in nature.
Well, for what it's worth the repotag issue was never introduced in any
non-political way. It is the way that repos used to visually mark
their packages like marking cattles. They don't produce more or less
milk either, but when a cattle breaks free and stampedes your meadow
to dirt you know what farmer to blame.
Same for packages. It was never a "technical" issue that Dag's repotag
"dag" was higher than ATrpms' "at". Neither that "rf"
was higher that
"ccrma" and so on. These "technical" arguments are rather lame, and
if
the people having these arguments really belived in them, they would
in fact want a repotag of "zzz_my_repo". There are no technical
benefits or drawbacks with repotags.
That particular one not even qualifies as an "argument" (IMHO), it can
be disproved as a problem in a few lines - reductio ad absurdum - if the
person at the other end is listening (and that's the problem).
For the record, the only "technical issue" that I felt was technical in
the list I wrote about was technical issues with respect to how to
implement this in the build system. The rest are non-technical from my
point of view (they are wishful thinking, or distractions from the issue
at hand, just plain bad logic or criticism about the quality of the
solution with no concrete plans for a better alternative - except for
the suggestion of somehow using the pgp signatures for that purpose).
[MUNCH]
Anyway, Nando, as much as I personally liked and promoted repotags,
it's currently more promising to beat a dead horse than to get epel to
consider repotags, after all epel already voted on this and nothing
has changed since.
Yup, I know. Don't worry, I don't _have_ that much time to invest :-)
When I see an email asking about this issue I'll try to answer (if there
is anything to say and I have time to say it), if I see objections I'll
do my best to answer them (I already did that a couple of times and
there were no responses to my answers). That's it...
-- Fernando
7:58 p.m.
On Sat, 2007-05-19 at 11:06 -0700, Fernando Lopez-Lezcano wrote:
On Fri, 2007-05-18 at 20:02 +0200, Axel Thimm wrote:
> On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
> > Yeah, I understand that. No offense people, but after reading some of
> > the recurring threads I was ready to go jump out of the nearest window.
>
> I hope that you live on the ground floor :)
Don't worry, I can hopefully resist the rige :-)
I have no clue what "rige" might be but I'll resist it anyway, I
actually tried to type "urge" and that came out...
-- Fernando
8:05 p.m.
On Sat, May 19, 2007 at 05:58:41PM -0700, Fernando Lopez-Lezcano wrote:
On Sat, 2007-05-19 at 11:06 -0700, Fernando Lopez-Lezcano wrote:
> On Fri, 2007-05-18 at 20:02 +0200, Axel Thimm wrote:
> > On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
> > > Yeah, I understand that. No offense people, but after reading some of
> > > the recurring threads I was ready to go jump out of the nearest window.
> >
> > I hope that you live on the ground floor :)
>
> Don't worry, I can hopefully resist the rige :-)
I have no clue what "rige" might be but I'll resist it anyway, I
actually tried to type "urge" and that came out...
I was wondering about that - "rige" is the name of one of my cats. She
never really tried to push people out of windows until now. :=)
--
Axel.Thimm at ATrpms.net
8:19 a.m.
New subject: repotags (was Re: EPEL report 2007, week 19)
On 18.05.2007 20:02, Axel Thimm wrote:
On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano
wrote:
[...]
fpc had signaled that they would
pass any decision of epel to introduce repotags
Hmm, I got different signals from Spot, which leads the Fedora Packaging
Committee.
See:
https://www.redhat.com/archives/fedora-devel-list/2007-April/msg01393.html
Quoting
----
But I'd like to point out that I'm still fundamentally opposed to the
repotag, as I've yet to hear what problem it solves, aside from the
"other repos want it" problem.
~spot
--
And below is something from #fedora-devel from some weeks ago after I
send
https://www.redhat.com/archives/fedora-devel-list/2007-April/msg01232.html
----
14:59 < knurd> | rdieter, spot, @FPC -- will you talk about the
repotag issue in todays meeting?
14:59 < spot> | knurd: if someone submits a draft.
15:00 < knurd> | spot, well, you seem to be still opposed to the idea
15:00 < spot> | http://fedoraproject.org/wiki/Packaging/Committee
15:01 < spot> | not to be rude, but, you proposed "lets do
repotags!"
15:01 < spot> | i'm far far more interested in "what problem are
we trying to solve?"
15:02 < spot> | daniel_hozac: no, that's the disttag
15:02 < rdieter> | spot: the problem is clear (hopefully not
debatable): same pkg in multiple repos.
15:02 < knurd> | spot, I only try to solve the problem "outside
people yell at EPEL and are makeing it look bad because it doesn't use
repotags"
15:02 < spot> | rdieter: so, repotag doesn't really solve that
problem.
15:02 < daniel_hoza> | spot: but it's also the repotag for all Fedora
packages?
15:02 < knurd> | spot, one of them is in FPC
15:02 < knurd> | spot, so all I want is politics
15:03 < rdieter> | spot: agreed, but it helps. it's not 100% a
technical issue either (far from it). sigh.
15:04 < spot> | rdieter: i'm not sure that it helps. if
foo-1.0.0-1.epel trumps foo-1.0.0-1.rf ... won't rf get mad?
15:05 < knurd> | spot, .rf is higher then .epel
15:05 < rdieter> | spot: repotag trumping is better than identical
EVR's.
15:05 < spot> | rdieter: i suppose that depends on which repo you
wanted that package to come from
15:07 < spot> | i'm not sure if adding them is harmless
15:07 < spot> | it complicates release ordering
15:08 < spot> | and n-v-r comparison
15:08 < spot> | and adds hacks to the buildsystem
15:08 < knurd> | spot, or spec files
15:08 < rdieter> | spot: if adding repotag to the end of Release
causes problems, it is only exposing problems that already existed.
15:08 < spot> | rdieter: no, its exposing problems i
intentionally avoided when i wrote the first naming drafts
15:09 < spot> | i didn't want any unnecessary non-numeric
characters in the n-v-r
15:09 < spot> | the repotag is "unnecessary non-numeric
characters"
15:09 < knurd> | spot, +1
15:10 < spot> | so far, the only technical reason I've heard is
that it helps us mix repos, which we're not trying to do anyways,
because our tools don't help us mix
these repos with or without repotags
15:10 < spot> | which leaves us with political reasons, in that
the other 3rd party RHEL repos want us to do it.
15:11 < spot> | the only way I can see supporting repotags in the
FPC is part of a much broader agreement between EPEL and the other repos
15:12 < spot> | to collaborate and focus on supporting mixing the
repos.
15:12 < knurd> | spot, that would mean Fedora probably as well
15:12 < knurd> | spot, why only EPEL then
15:12 < spot> | if all the repos say "mixing is our goal. we will
do X and Y and Z to make sure that users don't have pain."
15:12 < spot> | knurd: because as Dag and Karambir told me, they
don't care about Fedora.
15:13 < spot> | they don't do new packages for Fedora anymore
15:13 < knurd> | spot, but others do
15:13 < spot> | knurd: afaik, atrpms is the only one of
significant size
15:13 < spot> | other than dribble/livna, which already have
mixing as a marked goal
15:14 < knurd> | spot, and dries, which is part of rpmforge:
http://ftp.belnet.be/packages/dries.ulyssis.org/fedora/fc6/i386/RPMS.dries/
15:30 < spot> | rdieter: ok, thanks.
----
Spot's signals are the reasons why I don't want to invest more of *my*
time to get repotags for EPEL.
and fesco couldn't
care less, they don't like the idea of repotags (especially from their
non-rhel pov), but they wouldn't veto any such decision either.
Likely. But just FYI: FESCo itself decided against repotags in the past
for EPEL (in the early EPEL days), too.
CU
thl
9 a.m.
New subject: repotags (was Re: EPEL report 2007, week 19)
On Sun, 20 May 2007, Thorsten Leemhuis wrote:
On 18.05.2007 20:02, Axel Thimm wrote:
> On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
> [...]
> fpc had signaled that they would
> pass any decision of epel to introduce repotags
Hmm, I got different signals from Spot, which leads the Fedora Packaging
Committee.
See:
https://www.redhat.com/archives/fedora-devel-list/2007-April/msg01393.html
Quoting
----
But I'd like to point out that I'm still fundamentally opposed to the
repotag, as I've yet to hear what problem it solves, aside from the
"other repos want it" problem.
Ok, this is silly. spot should have joined the discussion instead of
iterating over the same arguments brought up on this list.
1. repotags makes packages unique (cross-repositories)
(yes, this means you have to believe in a world where there is more
than just EPEL, stop reading here if you don't believe in that)
2. repotags do not affect the version comparison in a useful way
(since there is no point in comparing release tags between
repositories anyway, there is no relation)
3. repotags help identify packages in every output
(you can talk about extra headers and changing RPM, we don't have that
today and unless Red Hat commits to backporting RPM support for
repotags we won't have that tomorrow either)
With a Fedora hat, I can understand that people think 'who cares, we never
needed it in Fedora, so why do we need it in EPEL' ? But that is
shortsighted.
Asking to do rpm -qpi or anything else just worsten the experience for
users and is a waste of time for anyone trying to help out.
Sure, RPM could be changed to support it outside of the release-tag.
Sure, Yum is not the best tool to support mixing repositories because it
does not allow easily to select specific packages from specific repos.
All that can be improved, but without the mindset that there is more than
one repository, all these changes will never happen. And I personally
blame Fedora's single repo mindset as the main reason why the current
situation is bad. Before Yum existed (apt) and with Smart these issues
did/do not exist especially because they were designed to select 'other'
versions of a package.
The big problem with EPEL is that people that use Fedora mainly are in
charge of making the rules, and spot's expression is right on that spot.
(Did spot ever use RPMforge or ATrpms on RHEL/CentOS from within Yum, Apt
or Smart ?)
Since it is pretty clear that CentOS Extras and EPEL will be seperate
projects packaging the same RPMs, I hope some people wake up and realise
that the Fedora mindset does not belong in EPEL.
Kind regards,
-- dag wieers, dag(a)wieers.com, http://dag.wieers.com/ --
[all I want is a warm bed and a kind word and unlimited power]
9:01 a.m.
New subject: repotags (was Re: EPEL report 2007, week 19)
On Sun, May 20, 2007 at 03:19:50PM +0200, Thorsten Leemhuis wrote:
On 18.05.2007 20:02, Axel Thimm wrote:
> On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
> [...]
> fpc had signaled that they would
> pass any decision of epel to introduce repotags
Hmm, I got different signals from Spot, which leads the Fedora Packaging
Committee.
See:
https://www.redhat.com/archives/fedora-devel-list/2007-April/msg01393.html
Quoting
----
But I'd like to point out that I'm still fundamentally opposed to the
repotag, as I've yet to hear what problem it solves, aside from the
"other repos want it" problem.
~spot
a) spot is not an inanimate object
b) quoting only half the quotes always biases towards your view, of
course. There is also:
"Indeed. I'm not necessarily opposed to the repotag being part of
%{dist} for EPEL, but I am opposed to making %{dist}
mandatory. [...] ~spot" (and you were part of that thread)
c) contrary to epel's leadership structure fpc works democratic w/o an
overpowered chairman, so even if spot would vote against it, there
are still N-1 other voters
And BTW I'm in no mood to start picking up infinite threads with you
again. Really. You want to present as if all the world, but you
blocked repotags, that's fine. In reality it was really you as a major
opponent to repotags that brought us here.
Do I sound like I'm pissed off? Maybe I do, and maybe I am. I just
don't like playing hide and seek or "pass the old maid".
--
Axel.Thimm at ATrpms.net
1:36 p.m.
New subject: repotags (was Re: EPEL report 2007, week 19)
On 20.05.2007 16:01, Axel Thimm wrote:
On Sun, May 20, 2007 at 03:19:50PM +0200, Thorsten Leemhuis wrote:
> On 18.05.2007 20:02, Axel Thimm wrote:
>> On Fri, May 18, 2007 at 09:50:09AM -0700, Fernando Lopez-Lezcano wrote:
>> [...]
>> fpc had signaled that they would
>> pass any decision of epel to introduce repotags
> Hmm, I got different signals from Spot, which leads the Fedora Packaging
> Committee.
> See:
> https://www.redhat.com/archives/fedora-devel-list/2007-April/msg01393.html
> Quoting
> ----
> But I'd like to point out that I'm still fundamentally opposed to the
> repotag, as I've yet to hear what problem it solves, aside from the
> "other repos want it" problem.
> ~spot
a) spot is not an inanimate object
Sorry, my english skills fail to parse that. What do did you mean?
http://www.thefreedictionary.com/inanimate
Did you mean something like "Spot can change his opinion"?
b) quoting only half the quotes always biases towards your view, of
course. There is also:
"Indeed. I'm not necessarily opposed to the repotag being part of
%{dist} for EPEL, but I am opposed to making %{dist}
mandatory. [...] ~spot" (and you were part of that thread)
Sorry, the thread is some weeks old I don't remember each every detail.
And I at least give pointers to the mails where somebody wrote something
when I quote him.
Further: I for one think a repotags only makes sense if it's used
everywhere. Dist isn't used everywhere and not mandatory, thus I'd
prefer not to abuse it.
c) contrary to epel's leadership structure fpc works democratic
w/o an
overpowered chairman,
The EPEL chairmen has no special powers. He just organizes and
coordinates -- that includes voting and is part of being a chairmen job
in my understanding (and done like that in FESCo, too).
so even if spot would vote against it, there
are still N-1 other voters
I just mentioned what impression I got from talking to the FPC chairmen.
Not more. Not less.
Further: I know from Jesse's statement from a recent FESCo meeting that
he is also against repotags. But yes, I don't know the other peoples
opinion. Do you? I actually wanted to know what FPC thinks about the
repotags (see the #fedora-devel quote in this thread) but didn't get one
until today.
And BTW I'm in no mood to start picking up infinite threads with
you
again. Really. You want to present as if all the world, but you
blocked repotags, that's fine.
We did not understood us there. I back then was against a blind vote
"EPEL uses repotags" without knowing the consequences for out packagers
(e.g. the technical implementation details). So I voted for "no", even
if I was not opposed heavily to repotags in general.
In reality it was really you as a major
opponent to repotags that brought us here.
If someone that wanted repotags did now what I try to do now (e.g. get
some row technical details in place before a voting) or worked out a
solution with the Fedora Packaging Committee before doing the EPEL vote
I might would have abstained a vote or might even have given a "+1" if
the solution was sane.
Do I sound like I'm pissed off? Maybe I do, and maybe I am. I
just
don't like playing hide and seek or "pass the old maid".
Axel, I really wanted and want to have you participating in EPEL. I even
wanted you in the Steering Committee -- that why I proposed you for it
when I presented the proposal for a Steering Committee to FESCo. I still
hope you can reconsider your leave and join us again.
Cu
thl
6184
days inactive
6192
days old
epel-devel@lists.fedoraproject.org
13 comments
5 participants
participants (5)
-
Axel Thimm -
Dag Wieers -
Fernando Lopez-Lezcano -
Josh Boyer -
Thorsten Leemhuis