Обсуждение: BUG #18784: Bugs and BugTracking

Поиск
Список
Период
Сортировка

BUG #18784: Bugs and BugTracking

От
PG Bug reporting form
Дата:
The following bug has been logged on the website:

Bug reference:      18784
Logged by:          Bharat DB
Email address:      bharatdb@cdac.in
PostgreSQL version: 17.0
Operating system:   Linux
Description:

Dear PostgreSQL Community,

We hope this message finds you well.

While going through the pgsql-bugs i found out there is no proper website or
a
system to monitor bugs or to track bug status ,i have also referred this
bug

https://www.postgresql.org/message-id/flat/5618263A.40806%40agliodbs.com#a8cc0869f5a3273df07c59fb8e2dde05
 but i can't find that  website is working now.

So,we propose creating a centralized repository that:

* Consolidates bug reports related to PostgreSQL-based applications and
their
integrations.
* Includes bugs identified and pushed from the PostgreSQL pgsql-bugs
mailing
list, categorized as open or closed for better tracking.
* Facilitates discussions and resolutions within the community.

This repository aims to reduce duplication of effort, improve transparency,
and
provide a unified platform for managing bug-related information. We are
also
pushing relevant data (e.g., author, message ID, date and time, body
message,
attachments, and URLs) from the pgsql-bugs list to GitHub for improved
organization and access.

I have created and dummy repo and tried to push some bugs ,Kindly find the
attachment of demo data below.

We kindly request your feedback for this initiative.

Your input would be invaluable in ensuring that this repository serves as
a
constructive addition to the PostgreSQL ecosystem.

 Looking forward to your thoughts and suggestions.

Thank you for your time and support.

Warm regards,
Vasuki M
CDAC Chennai
Email:bharatdb@cdac.in
Phone:6374605133


Re: BUG #18784: Bugs and BugTracking

От
Álvaro Herrera
Дата:
Hello

On 2025-Jan-24, PG Bug reporting form wrote:

> While going through the pgsql-bugs i found out there is no proper website or
> a system to monitor bugs or to track bug status ,i have also referred
> this bug
> https://www.postgresql.org/message-id/flat/5618263A.40806%40agliodbs.com#a8cc0869f5a3273df07c59fb8e2dde05
>  but i can't find that  website is working now.

This refers to the following
https://granicus.if.org/pgbugs/
which was an experiment with a bug tracker from 2015 that gave no fruit.

> I have created and dummy repo and tried to push some bugs ,Kindly find the
> attachment of demo data below.

I didn't see any attachment to this email, but I doubt that a system
that doesn't feed directly from the mailing list traffic is going to go
anywhere with this community.  A manually maintained github repo seems
unworkable.

I described what I see as working during last year's pgconf.dev
developer's meeting, namely a system very similar to our patch
tracker^W^W commitfest app, which would contain links to pgsql-bugs
threads.  The app, in addition to listing the relevant thread links for
each bug, would allow simple metadata such as affected Postgres
branches, SHA1 git commit IDs (indicating point at which each bug was
fixed on each branch), maybe a simple tagging system.  No discussion
would take place on the app itself, since that should continue to be
relegated to mailing lists.

It would be possible to create bugs in that system in two ways: one,
using the already existing bug reporting form, which would assign the
bug number, post to pgsql-bugs and create the bug metadata in the new
system; two, by manually accepting a link to a thread in some mailing
list (probably pgsql-hackers, but it could be any other
@lists.postgresql.org list), whereby a new bug number would be assigned
but no additional mail is sent.

I think this wouldn't require much new code, because, as I said, it
would be very similar to the commitfest app, so it's quite likely that
we could reuse a lot of that.  It would be a separate app, however,
because though they are similar, the requirements are different enough.


Why do I think this would be successful?  Many years ago, there was a
proposal that would use Debian's bugtracker, which is essentially a
glorified mailing list with per-bug archives, plus a lot of smarts to
know when to close a bug.  This had the most traction among this
community, but the idea died down when it was clear that 1) the Debian
community was not really open to sharing their stuff, and 2) nobody
offered to maintain this.

-- 
Álvaro Herrera         PostgreSQL Developer  —  https://www.EnterpriseDB.com/



Re: BUG #18784: Bugs and BugTracking

От
Tom Lane
Дата:
=?utf-8?Q?=C3=81lvaro?= Herrera <alvherre@alvh.no-ip.org> writes:
> I described what I see as working during last year's pgconf.dev
> developer's meeting, namely a system very similar to our patch
> tracker^W^W commitfest app, which would contain links to pgsql-bugs
> threads.

Yeah, given our existing workflow (which no one seems in a hurry
to change), what we fundamentally want is an index to mailing list
threads about bugs, plus status information.

> Why do I think this would be successful?  Many years ago, there was a
> proposal that would use Debian's bugtracker, which is essentially a
> glorified mailing list with per-bug archives, plus a lot of smarts to
> know when to close a bug.  This had the most traction among this
> community, but the idea died down when it was clear that 1) the Debian
> community was not really open to sharing their stuff, and 2) nobody
> offered to maintain this.

Even more years ago (roughly around 2000 IIRC), we actually did try
to use a bug tracker, but the experiment only lasted a few weeks.
It failed because (a) that tracker was not adapted to mailing-list-
based workflow, and (b) we didn't have people who were interested
in moderating/curating the tracker's contents.  If you don't have
some people who will keep status entries up-to-date, close out junk
bugs, and such, the tracker descends to uselessness very quickly.

Our community was far smaller then than it is today, and I bet
finding people for (b) would be easier.  But we still need
something that meets criterion (a), and GitHub ain't it.

            regards, tom lane