Python to use a non open source bug tracker?

A

Aahz

Disrespectful? Because I say that I don't agree with some specific
requirement, trying to discuss and understand the rationale behind it?

*snort* Now you're trying to dodge responsibility for your own words.
Normally I'd be inclined to cut you some slack because English isn't your
native language, but it's clear that you're deliberately trying to write
emotionally to encourage other people to flail around. Here are again
some of the things you've written in this thread:

Does this smell "Bitkeeper fiasco" to anyone else than me?
I am impressed you do not want to understand the "why"
most bug reports submitted by users go totally ignored for several years,

That adds up to a clear summation of disprespect in my view.
 
I

Ilias Lazaridis

Ben said:
Oh my. You have *seriously* misjudged this group if you think that
comment will give you any net gain in discussions here.

I'm not interested in any gain (except within the systems that I
produce).

My intention was mainly to place myself on the side of the OP, before
the 'Herd of Savages' (the 'Regular Posters') get's out of control and
eat's him.

..
 
T

Terry Reedy

Giovanni Bajo said:
tracker. I was claiming that, if such a group was ever formed, it was
better
spent on bug triage rather than keeping their keys ready all day long to
quick-fix any server breakage in minutes.

This could be made into an argument for accepting the Jira offer so we
don't 'waste' *any* more Python-knowledgable volunteer time on admin.
However, thinking about it more, I think that wrestling with a software
system like Roundup and interacting with sometimes naive and non-responsive
bug submitters are two different skills and likely to attract different
volunteers.

[snip]
Either close directly any nonsense, or ask for more feedback to the
poster,
until the bug/patch/rfe is sufficiently clear to be handled, or 3 months
are
passed and you close the bug for "no further feedback from the poster".
If this
would dramatically reduce the number of open bugs, then yes, Python
really
needs someone to do bug triaging.

I have thought this for some time based on my occasional efforts at
'first-response' reviewing. But I have not tried to do anything because of
the difficulty of working with the SF tracker. Perhaps submissions by new
submitters should start in 'limbo' until rejected or accepted into active
open status. I hope that whichever new tracker we get will allow for
automated followups at determined intervals, such as 3 mos or whatever.
It might be not a good use of your time at all, since you are a
developer. But
having a database with 938 open bugs most of which are
incomplete/nonsense/unconfirmed is much less useful than it could be.

Perhaps when the new tracker is set up, you can help scratch the 'too many
open bugs' itch.
It also
raises the bar for new developers: it's much harder to just "pick one"
and fix
it. I know because I tried sometimes, and after half an hour I couldn't
find
any bug that was interesting to me and "complete" enough to work on it. I
also
noticed that most bugs are totally uncommented like nobody cared at all.
This
is where my thought about Python missing bug triaging started.

s/most/some/

When I read a bug with no comment I sometimes put extra energy into
thinking of something to say or ask just so the reporter will know the
report has been read.

Terry Jan Reedy
 
?

=?ISO-8859-15?Q?=22Martin_v=2E_L=F6wis=22?=

Giovanni said:
Are you ever going to try and make a point which is not "you are not entitled
to have opinions because you do not act"? Your sarcasm is getting annoying. And
since I'm not trolling nor flaming, I think I deserve a little bit more of
respect.

You seem to imply that people who are willing to do roundup admin could
regularly, easily do bug triage, and also would be willing to do so. I
can attest that this assumption is sooooooo remote from the truth that
I can't really believe you really meant it.

I have called for people doing bug review and providing patches many
many times, and most of these calls got unheard.

Regards,
Martin
 
?

=?ISO-8859-1?Q?=22Martin_v=2E_L=F6wis=22?=

Paul said:
According to Harald Armin Massa's PostgreSQL talk at EuroPython, the
PostgreSQL people manage all their bugs via mailing lists. Given that
trends in revision control point towards completely decentralised
solutions, I wonder whether there's anything to learn from less
centralised (or more flexible) approaches to bug management.

From my experience with GCC, I can only report that this is definitely
not working. There used to be a mailing list (e-mail address removed), and
reports got either answered immediately, or not at all. People who
thought they were responsible put the mails in some folder, and then
never found the time to come back. This is why I set up a bug tracker
for GCC.

Regards,
Martin
 
F

Fredrik Lundh

Martin said:
not working. There used to be a mailing list (e-mail address removed), and
reports got either answered immediately, or not at all. People who
thought they were responsible put the mails in some folder, and then
never found the time to come back.

you need tools to help you track the bugs and their status, but you can
handle issue registration, discussion, and most maintenance stuff using
good old mail just fine.

</F>
 
S

skip

Fredrik> you need tools to help you track the bugs and their status, but
Fredrik> you can handle issue registration, discussion, and most
Fredrik> maintenance stuff using good old mail just fine.

Which is something SourceForge has yet to learn. At work we use a system
called RT (http://www.bestpractical.com/rt/). While it's not perfect, it
does allow submissions and responses via email. That feature alone puts it
miles ahead of SF in my mind.

Skip
 
P

Paul Rubin

Which is something SourceForge has yet to learn. At work we use a system
called RT (http://www.bestpractical.com/rt/). While it's not perfect, it
does allow submissions and responses via email. That feature alone puts it
miles ahead of SF in my mind.

I'm on the other side--I think spam has destroyed the usefulness of
email as a communications medium and I don't want to depend on
anything having to do with email any more. I hate the way SF requires
registering an email address and then it emails you every update to
your SF issues. As a low-intensity user, I sort of tolerate it. But
if I used SF more, I'd have to direct all the SF email to a spam
bucket and never look at it. At most I'd want it to send about one
email per week. But I'd much rather have a personalized RSS feed that
delivers updates about the bugs that I'm following.

I also notice that the PyPy mailing list now delivers mostly spam, so
I've had to direct that to a spam bucket.
 
?

=?ISO-8859-1?Q?Michael_Str=F6der?=

Fredrik> you need tools to help you track the bugs and their status, but
Fredrik> you can handle issue registration, discussion, and most
Fredrik> maintenance stuff using good old mail just fine.

Which is something SourceForge has yet to learn. At work we use a system
called RT (http://www.bestpractical.com/rt/). While it's not perfect, it
does allow submissions and responses via email. That feature alone puts it
miles ahead of SF in my mind.

I'd also prefer to at least be able to respond to tracker items via
e-mail. I'm traveling a lot by train working off-line most times. An
e-mail spool is unbeatable in this situation. So it's ok for me to add
an issue to a tracker while being online. But the follow-ups should be
possible via e-mail. SF pretty much sucks in this regard (and has other
issues too).

The main reason why I'm following this discussion is that I'd also
prefer to move away from SF for python-ldap. One reason is availability
and the other one is the really bad user interface.

I could imagine to spend some spare time when this infrastructure can
also be used for python-ldap. And pydns would be another candidate to be
moved away from SF.

Ciao, Michael.
 
?

=?ISO-8859-1?Q?Michael_Str=F6der?=

Paul said:
I'm on the other side--I think spam has destroyed the usefulness of
email as a communications medium and I don't want to depend on
anything having to do with email any more.

E-mail spam is an issue but the python.org infrastructure already has to
do spam filtering for mailing lists. Or does it simply resend all mail?

Ciao, Michael.
 
P

Paul Rubin

Michael Ströder said:
E-mail spam is an issue but the python.org infrastructure already has to
do spam filtering for mailing lists. Or does it simply resend all mail?

The problem is that the lists (or at least the pypy list) got mirrored
somewhere without having the addresses obscured. Spam robots found
the mirrors and scanned for email addresses as they usually do. The
addresses then circulate between the spammers and get on more and more
lists. It doesn't matter whether the list reflectors forward spam or
not. Spammers send crap to the addresses directly.
 
M

Magnus Lycka

Fredrik said:
you're not on the infrastructure list, I hear.

I tried to figure out where that list is, so I could have
a look at the archives, but I didn't find it in the (for
me) obvious places. Could someone please provide a link
to the archives for this mailing list, or aren't there
any public archives of them? Only for PSF members?
python.org could still need a few more roundup volunteers,
> but it's not like nobody's prepared to contribute manhours.
> don't underestimate the community.

So, how many have offered to help? Is this information
available in some public repository?

I don't know how much work it actually takes to maintain
a roundup installation for the Python project, but I know
that in general, not all people manage to follow through
on everything they commit to do, even if they have good
intentions, so I'd be a bit worried to move to roundup if
only two or three people had offered to run it, even if
that might nominally be enough. Of course, this depends
on who those people would be... Ten seems like a bit too
many though. I somehow suspect that less work would get
done in a group of ten than in a group of six people...

It seems to me that an obvious advantage with either Roundup
or Trac, is that if the Python project used it, the Python
project would have a significant impact on how this product
developed. Even if the Jira people seem eager to please us,
I'm pretty convinced that it will be easier to get Roundup
or Trac improved to fit our particular needs.

That's valuable in two ways:
1) The Python project would get a bug tracker which is
developed with the needs of the Python project as a
prime concern. (Being the major "customer" of a product
has benefits. Jira on the other hand, might get more
and more integrated with other Java stuff that we don't
really care about.
2) We'd help making a good Python product even better, and
probably more widely used, thus spreading the use of
Python even further.
 
G

Georg Brandl

Magnus said:
I tried to figure out where that list is, so I could have
a look at the archives, but I didn't find it in the (for
me) obvious places. Could someone please provide a link
to the archives for this mailing list, or aren't there
any public archives of them? Only for PSF members?

The archives are viewable for list members. The list info is at
http://mail.python.org/mailman/listinfo/infrastructure
So, how many have offered to help? Is this information
available in some public repository?

Not yet, as it seems.

Georg
 
S

skip

Michael> E-mail spam is an issue but the python.org infrastructure
Michael> already has to do spam filtering for mailing lists. Or does it
Michael> simply resend all mail?

Email sent to most mailing lists hosted on mail.python.org are passed
through a SpamBayes instance before being forwarded to the list's members.

Skip
 
P

Paul Boddie

Magnus said:
It seems to me that an obvious advantage with either Roundup
or Trac, is that if the Python project used it, the Python
project would have a significant impact on how this product
developed. Even if the Jira people seem eager to please us,
I'm pretty convinced that it will be easier to get Roundup
or Trac improved to fit our particular needs.

Yes, because Roundup and Trac are open source projects: there is no
barrier to prevent the users taking the code in a direction appropriate
to their own needs. And just to make it clear that I'm not picking on
Jira, it should be noted that even with their apparent willingness to
make a useful "community" product (and their otherwise remarkable open
source credentials), the Launchpad developers can't offer the kinds of
assurances implicitly provided by Roundup, Trac or any of their open
source brethren.
That's valuable in two ways:
1) The Python project would get a bug tracker which is
developed with the needs of the Python project as a
prime concern. (Being the major "customer" of a product
has benefits. Jira on the other hand, might get more
and more integrated with other Java stuff that we don't
really care about.

As has been said already, there's supposedly no guarantee that people
will want to develop Roundup at a hectic tempo in order to satisfy the
needs/desires of the Python developers. But then again, other pieces of
infrastructure have a high community investment, notably Mailman (which
uses Jira as its issue tracker, as it turns out).
2) We'd help making a good Python product even better, and
probably more widely used, thus spreading the use of
Python even further.

It seems to me that with all the fuss about marketing Python [1],
instead of ranting about how other products and technologies are
stealing all the thunder, one might instead want to start closer to
home. In this respect, several opportunities are being missed or
squandered either because people think marketing is all about press
releases, or they want Python to retain its stealth label (the
"competitive advantage" people mention constantly).

Take python.org as the place to start. One can claim all one likes
about how Web applications aren't special enough to warrant special
mentions or coverage in the context of persuading people about Python's
advantages, but many people presumably visit python.org and wonder...

* How they can develop Web applications using Python in a way they
recognise either from intuition or previous experience. Where can
they find a good solution and get started quickly?

* Whether python.org, as some kind of content platform, is some kind
of convenient answer to their own Internet/intranet site project.
Can they download the code and run the same kind of thing
themselves?

The answers aren't too clear to these questions. I've revisited some of
the material available via python.org [2] in order to attempt to
provide clearer answers to the first question, but the topic of
standardisation is currently stagnant (so it's every framework for
itself), and the community is split between hyping the most popular
frameworks whilst emphasizing the modest achievements that led to WSGI
(which doesn't really answer the first question entirely). Meanwhile,
despite the python.org codebase presumably running various commercial
sites, it would surprise me if there would ever be a convenient
downloadable package of that codebase available prominently from
python.org itself (even though the components are all openly
available). So the Python project - the power behind content management
solutions like Zope, Plone and (at a different angle) MoinMoin - offers
an incoherent response to the second question.

Then, there are the other recommendations under the "Using Python
For..." heading - advocacy points to show how Python can be really
useful - which mentions under "Software Development" the following:
Buildbot, Trac, Roundup and IDEs. If one ignores the current issue
tracker debate for a moment and follows the "Software Development"
link, one reaches a general Python applications page which mentions
amongst other "choices for web development" the CPS project, and
following the provided link swiftly delivers another advocacy own-goal:
"We're switching to JAVA!" state the CPS people proudly, still
blissfully unaware that "Java" isn't an acronym; "Read why" they
suggest.

It's tempting to label what I've written above as just some
opportunistic criticism of the maintenance level of the python.org
content, that the core developers should just choose their tools and
get on with things, and that this thread has attempted to politicize
the decision under discussion from the start. Indeed, as someone who
merely browses python-dev, perhaps I shouldn't care how the core
developers track their bugs: if they struggle to manage that
information in future, why should I care? Well, the reason I should
care is related to the reason why the core developers should care about
more than purely technical issues: the wider community and the core
developers do not exist by themselves in isolation; the well-being of
the community is related to how Python is managed and portrayed by the
custodians of the language, and the well-being of the development
effort is related to how much community effort can be directed towards
improving the language and its image. If this were not so, Python would
have vanished like many of its contemporaries.

Perhaps the decision makers evaluated the above and much more in depth,
although us outsiders are not in a position to say, but perhaps the
discussion around the decision wouldn't have been so inflammatory in
places if there had been an acknowledgement of this "bigger picture" of
the community, its influences and that in a large open source project
no moderately significant decision is without a political dimension.

Paul

[1] http://www.artima.com/forums/flat.jsp?forum=106&thread=150515
[2] http://wiki.python.org/moin/WebFrameworks
 
A

A.M. Kuchling

... Meanwhile, despite the python.org codebase presumably running
various commercial sites, ...

Nothing should have given you this impression! python.org's
formatting is handled through a custom script called Pyramid, and if
you poke around with enough determination you can find the SVN
repository's URL. But it's never been released as a tarball, and
isn't used by any other sites.

As an experiment I tried formatting the python.org site using
rest2web; that looks promising, if I ever figure out how sidebars
work, and may someday replace Pyramid.

--amk
 
P

Paul Boddie

A.M. Kuchling said:
Nothing should have given you this impression! python.org's
formatting is handled through a custom script called Pyramid, and if
you poke around with enough determination you can find the SVN
repository's URL. But it's never been released as a tarball, and
isn't used by any other sites.

I believed that Pollenation had deployed other sites using the same
technology. Certainly, they appear to be members of the Nevow community
on which Pyramid somehow seems to be based. Once upon a time, I did
promise to try and make Debian packages available for Pyramid so that
people might be more inclined to use the toolchain and thus contribute
python.org content, but given the complexity of the dependencies
(Twisted 2, some YAML parser that conflicts with the Python bindings of
the most widely-deployed YAML parser, Nevow, other stuff) I decided
that my time was arguably better spent elsewhere, unfortunately.
As an experiment I tried formatting the python.org site using
rest2web; that looks promising, if I ever figure out how sidebars
work, and may someday replace Pyramid.

The python.org redesign was another matter of major community
controversy, of course, although I don't have as much to say on that
matter. One day, the python.org Wiki may have invaded enough of the
visible "reference space" on the site to make the façade so thin as to
be dispensible, although I'm certainly not compaigning for such a thing
to happen.

Paul
 
B

Ben Finney

Paul Boddie said:
Indeed, as someone who merely browses python-dev, perhaps I
shouldn't care how the core developers track their bugs: if they
struggle to manage that information in future, why should I care?
Well, the reason I should care is related to the reason why the core
developers should care about more than purely technical issues: the
wider community and the core developers do not exist by themselves
in isolation; the well-being of the community is related to how
Python is managed and portrayed by the custodians of the language,
and the well-being of the development effort is related to how much
community effort can be directed towards improving the language and
its image. If this were not so, Python would have vanished like many
of its contemporaries.

Perhaps the decision makers evaluated the above and much more in depth,
although us outsiders are not in a position to say, but perhaps the
discussion around the decision wouldn't have been so inflammatory in
places if there had been an acknowledgement of this "bigger picture" of
the community, its influences and that in a large open source project
no moderately significant decision is without a political dimension.

Thank you.
 
B

Ben Finney

Magnus Lycka said:
So, how many have offered to help? Is this information available in
some public repository?

I don't yet know of private discussions leading to it, but Brett
Cannon has made an unofficial announcement that Roundup has been
picked:

I am making an unofficial announcement here that it looks like we
will be able to go with Roundup as the issue tracker for
python-dev. Now this does not mean people should stop volunteering
by emailing infrastructure at python.org! We have not finalized
which of the volunteers will be asked to help admin the Roundup
installation so if you want to help please email us with your
timezone, rough amount of time you can donate per week, and your
Roundup experience.

This announcement is unofficial because there has been an offer
for professional Roundup hosting. We are awaiting the details of
the offer before deciding how to proceed. Once we have decided how
we are going to handling hosting there will be an official
announcement with more details.

<URL:http://sayspy.blogspot.com/2006/10/looks-like-we-will-be-going-with.html>
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,982
Messages
2,570,189
Members
46,734
Latest member
manin

Latest Threads

Top