Thread: CommitFest Wiki page annoyance
It seems changes to the commit fest wiki pages are going to be overwritten
accidently when editing the page concurrently. At least, this occured to me
as i accidently removed entries done by Laurenz Albe and my changes were
overwritten by Alvaro Herrera again.....
So developers need to be careful when editing their projects there.
-- Thanks
Bernd
Bernd Helmle wrote: > It seems changes to the commit fest wiki pages are going to be > overwritten accidently when editing the page concurrently. At least, > this occured to me as i accidently removed entries done by Laurenz > Albe and my changes were overwritten by Alvaro Herrera again..... > > So developers need to be careful when editing their projects there. > Maybe we need to be using a database backed tracker ... *ducks* cheers andrew
Bernd Helmle wrote: > It seems changes to the commit fest wiki pages are going to be > overwritten accidently when editing the page concurrently. At least, this > occured to me as i accidently removed entries done by Laurenz Albe and my > changes were overwritten by Alvaro Herrera again..... > > So developers need to be careful when editing their projects there. Hmm. I do click "show changes" before each commit, and I make sure no such thing shows up on the diff. Perhaps what's going on is that the software does not check that the version I read is the most recent one :-( so if two of us edit the page at the same time, the one committing last is going to stomp on the changes of the other one. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
Andrew Dunstan wrote: > > > Bernd Helmle wrote: >> It seems changes to the commit fest wiki pages are going to be >> overwritten accidently when editing the page concurrently. At least, >> this occured to me as i accidently removed entries done by Laurenz >> Albe and my changes were overwritten by Alvaro Herrera again..... >> >> So developers need to be careful when editing their projects there. > > Maybe we need to be using a database backed tracker ... I did say, some time before we had any Commitfest page at all, that I considered a Wiki to be a bad substitute for a tracker. With luck we will be able to replace the current interim setup with a real tracker. -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Alvaro Herrera <alvherre@commandprompt.com> writes:
> Bernd Helmle wrote:
>> It seems changes to the commit fest wiki pages are going to be
>> overwritten accidently when editing the page concurrently. At least, this
>> occured to me as i accidently removed entries done by Laurenz Albe and my
>> changes were overwritten by Alvaro Herrera again.....
>>
>> So developers need to be careful when editing their projects there.
> Hmm. I do click "show changes" before each commit, and I make sure no
> such thing shows up on the diff.
Does "show changes" diff against the latest version, or the one you
started to edit?
I can see this behavior becoming a real problem, particularly during
commit fest when multiple people might be updating the page. Can we
do anything to prevent collisions? As a first cut, I'd be happy if
it refused to let me commit if the page had changed from the version
I started to edit.
regards, tom lane
On Wed, Apr 23, 2008 at 2:40 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote: > > Bernd Helmle wrote: > >> It seems changes to the commit fest wiki pages are going to be > >> overwritten accidently when editing the page concurrently. At least, this > >> occured to me as i accidently removed entries done by Laurenz Albe and my > >> changes were overwritten by Alvaro Herrera again..... > >> > > I can see this behavior becoming a real problem, particularly during > commit fest when multiple people might be updating the page. Can we > do anything to prevent collisions? As a first cut, I'd be happy if > it refused to let me commit if the page had changed from the version > I started to edit. > Strange, I thought mediawiki dealt with this problem in a fairly sane way: http://meta.wikimedia.org/wiki/Help:Edit_conflict According to the information at the above link, the wiki should let you know about concurrent edit conflicts and give you an opportunity to resolve them before attempting to commit your changes again. Not sure how the conflicts observed by Bernd could have happened. We're not running some drastically out-of-date version of mediawiki by any chance? Cheers, BJ
On Wed, 23 Apr 2008 02:59:43 +1000 "Brendan Jurd" <direvus@gmail.com> wrote: > According to the information at the above link, the wiki should let > you know about concurrent edit conflicts and give you an opportunity > to resolve them before attempting to commit your changes again. > > Not sure how the conflicts observed by Bernd could have happened. > We're not running some drastically out-of-date version of mediawiki by > any chance? Likely the other way around. I believe it is running against 8.3.1 and only the *latest* versions of mediawiki support it. Joshua D. Drake > > Cheers, > BJ > -- The PostgreSQL Company since 1997: http://www.commandprompt.com/ PostgreSQL Community Conference: http://www.postgresqlconference.org/ United States PostgreSQL Association: http://www.postgresql.us/ Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
On Wed, Apr 23, 2008 at 12:16 AM, Alvaro Herrera <alvherre@commandprompt.com> wrote: > > Perhaps what's going on is that the software does not check that the > version I read is the most recent one :-( so if two of us edit the page > at the same time, the one committing last is going to stomp on the > changes of the other one. > I had a look into the history on the CommitFest page. It looks like Alvaro unintentionally clobbered the previous five edits, which went back about four hours. I've restored those changes now, but it is definitely worrying that such a drastic regression could occur. The only way I can conceive of this happening is that Alvaro was making his edits on a *very* old copy of the edit page. Cheers, BJ
--On Mittwoch, April 23, 2008 03:17:54 +1000 Brendan Jurd
<direvus@gmail.com> wrote:
> I had a look into the history on the CommitFest page. It looks like
> Alvaro unintentionally clobbered the previous five edits, which went
> back about four hours. I've restored those changes now, but it is
> definitely worrying that such a drastic regression could occur.
>
> The only way I can conceive of this happening is that Alvaro was
> making his edits on a *very* old copy of the edit page.
Hmm yes, after looking at the diff again i see that Alvaro committed 23
minutes after my changes, maybe he had opened the wiki page *a lot* earlier
before starting to edit it.....?
-- Thanks
Bernd
"Brendan Jurd" <direvus@gmail.com> writes:
> I had a look into the history on the CommitFest page. It looks like
> Alvaro unintentionally clobbered the previous five edits, which went
> back about four hours. I've restored those changes now, but it is
> definitely worrying that such a drastic regression could occur.
We saw this type of thing happen more than once on the CommitFest:March
page, if memory serves. Apparently there is something very broken about
the copy of mediawiki we are using --- would someone look into that?
regards, tom lane
Brendan Jurd escribió: > On Wed, Apr 23, 2008 at 12:16 AM, Alvaro Herrera > <alvherre@commandprompt.com> wrote: > > > > Perhaps what's going on is that the software does not check that the > > version I read is the most recent one :-( so if two of us edit the page > > at the same time, the one committing last is going to stomp on the > > changes of the other one. > > I had a look into the history on the CommitFest page. It looks like > Alvaro unintentionally clobbered the previous five edits, which went > back about four hours. I've restored those changes now, but it is > definitely worrying that such a drastic regression could occur. > > The only way I can conceive of this happening is that Alvaro was > making his edits on a *very* old copy of the edit page. Really strange. The only way I imagine this could have happened is that the browser got the page from the cache instead of re-fetching it. As I said, I normally verify the diff manually before committing. I'll keep an eye on diffs _after_ committing. -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Tom Lane escribió: > "Brendan Jurd" <direvus@gmail.com> writes: > > I had a look into the history on the CommitFest page. It looks like > > Alvaro unintentionally clobbered the previous five edits, which went > > back about four hours. I've restored those changes now, but it is > > definitely worrying that such a drastic regression could occur. > > We saw this type of thing happen more than once on the CommitFest:March > page, if memory serves. Apparently there is something very broken about > the copy of mediawiki we are using --- would someone look into that? Hmm, perhaps it is the Postgres port that's broken. It wouldn't be really surprising that people mostly use the MySQL port ... -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Alvaro Herrera <alvherre@commandprompt.com> writes:
> Tom Lane escribi�:
>> We saw this type of thing happen more than once on the CommitFest:March
>> page, if memory serves. Apparently there is something very broken about
>> the copy of mediawiki we are using --- would someone look into that?
> Hmm, perhaps it is the Postgres port that's broken. It wouldn't be
> really surprising that people mostly use the MySQL port ...
Conflict detection not working, perhaps? Wouldn't be too surprising.
But right now I'm just asking that somebody from -www find out exactly
what version we are running.
regards, tom lane
On Tue, 22 Apr 2008 13:35:41 -0400 Alvaro Herrera <alvherre@commandprompt.com> wrote: > Tom Lane escribió: > > "Brendan Jurd" <direvus@gmail.com> writes: > > > I had a look into the history on the CommitFest page. It looks > > > like Alvaro unintentionally clobbered the previous five edits, > > > which went back about four hours. I've restored those changes > > > now, but it is definitely worrying that such a drastic regression > > > could occur. > > > > We saw this type of thing happen more than once on the > > CommitFest:March page, if memory serves. Apparently there is > > something very broken about the copy of mediawiki we are using --- > > would someone look into that? > > Hmm, perhaps it is the Postgres port that's broken. It wouldn't be > really surprising that people mostly use the MySQL port ... It could be, or it could be that the postgres port on 8.3 is broken :). I don't know which version we installed for wiki, but in order to get mediawiki running for postgresqldocs, I had to use HEAD of mediawiki. However I note that Greg submitted his 8.3 compatibility before the latest release of mediawiki which is 1.12.0. In theory, we should be fine in that regard. Sincerely, Joshua D. Drake -- The PostgreSQL Company since 1997: http://www.commandprompt.com/ PostgreSQL Community Conference: http://www.postgresqlconference.org/ United States PostgreSQL Association: http://www.postgresql.us/ Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
"Brendan Jurd" <direvus@gmail.com> writes: > I had a look into the history on the CommitFest page. It looks like > Alvaro unintentionally clobbered the previous five edits, which went > back about four hours. I've restored those changes now, but it is > definitely worrying that such a drastic regression could occur. > > The only way I can conceive of this happening is that Alvaro was > making his edits on a *very* old copy of the edit page. I accidentally did the same thing to Tom a while back. In that case I had explicitly pulled up an old version of the page to copy out part of it. I believe what then happened is I accidentally pasted the part I was recovering back into that same old version rather than the current one and hit save. -- Gregory Stark EnterpriseDB http://www.enterprisedb.com Get trained by Bruce Momjian - ask me about EnterpriseDB'sPostgreSQL training!
Tom Lane wrote: > Alvaro Herrera <alvherre@commandprompt.com> writes: > > Tom Lane escribió: > >> We saw this type of thing happen more than once on the > >> CommitFest:March page, if memory serves. Apparently there is > >> something very broken about the copy of mediawiki we are using --- > >> would someone look into that? > > > Hmm, perhaps it is the Postgres port that's broken. It wouldn't be > > really surprising that people mostly use the MySQL port ... > > Conflict detection not working, perhaps? Wouldn't be too surprising. > But right now I'm just asking that somebody from -www find out exactly > what version we are running. That's easy enough: http://wiki.postgresql.org/wiki/Special:Version So we're on 1.12alpha (r31813) Now what that actually *means* is a whole different question :-) I don't know mediawiki nearly well enough to commment. //Magnus
Magnus Hagander wrote: > Tom Lane wrote: >> Alvaro Herrera <alvherre@commandprompt.com> writes: >>> Tom Lane escribió: >>>> We saw this type of thing happen more than once on the >>>> CommitFest:March page, if memory serves. Apparently there is >>>> something very broken about the copy of mediawiki we are using --- >>>> would someone look into that? >>> Hmm, perhaps it is the Postgres port that's broken. It wouldn't be >>> really surprising that people mostly use the MySQL port ... >> Conflict detection not working, perhaps? Wouldn't be too surprising. >> But right now I'm just asking that somebody from -www find out exactly >> what version we are running. > > That's easy enough: http://wiki.postgresql.org/wiki/Special:Version > > So we're on 1.12alpha (r31813) > > Now what that actually *means* is a whole different question :-) I > don't know mediawiki nearly well enough to commment. we are running a mediawiki version from the 1_12 branch downloaded about 1-2 weeks before it got released as 1.12. upgrading to the final release is an open ticket on pmt but I don't think there are many changes between our version and the current release. Stefan
On Tue, 22 Apr 2008 19:52:02 +0200 Magnus Hagander <magnus@hagander.net> wrote: > That's easy enough: http://wiki.postgresql.org/wiki/Special:Version > > So we're on 1.12alpha (r31813) > > Now what that actually *means* is a whole different question :-) I > don't know mediawiki nearly well enough to commment. O.k. the latest is 1:12.0 stable... we need to get that updated. Joshua D. Drake > > //Magnus > -- The PostgreSQL Company since 1997: http://www.commandprompt.com/ PostgreSQL Community Conference: http://www.postgresqlconference.org/ United States PostgreSQL Association: http://www.postgresql.us/ Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate