Обсуждение: Re: [pgsql-www] New email list for emergency communications
Stefan Kaltenbrunner wrote: [ Discussion about making infrastructure downtime more visible to people.] > > Right now, when email is down, I just start IM'ing people. A more > > structured system would help me. > > Some people would say a more structured approach to patch/bug tracking > would help them (but maybe not help you) ;-) [ Email moved to hackers.] This is an interesting observation that deserves to be discussed separately, and relates to the patch queue. Many people want to know what is happening with the patch queue but having better reporting doesn't seem to help much the people who are actually processing the queue, and requires more work. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Bruce Momjian wrote: > This is an interesting observation that deserves to be discussed > separately, and relates to the patch queue. Many people want to know > what is happening with the patch queue but having better reporting > doesn't seem to help much the people who are actually processing the > queue, and requires more work. I'm not sure I follow. I tried last week to give some hours for patch review. What actually happened was that I had to wade through a ton of stuff, skipping patches that were already applied, threads that were purely discussion but not patches, threads about earlier versions of some patches. I did manage to find actual patches, but I had then to resort to my own mailbox to get the actual patch to apply and review. Now I can't be sure whether the current commitfest is finished or there are still patches pending review; what patches could use input from me; what patches I could bounce to the next commitfest. Also I don't know what patches we have for the next commitfest. I would gladly give some time to put up a Wiki page to list the patches for the next commitfest. Would that be helpful? -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
Alvaro Herrera wrote: > Bruce Momjian wrote: > > > This is an interesting observation that deserves to be discussed > > separately, and relates to the patch queue. Many people want to know > > what is happening with the patch queue but having better reporting > > doesn't seem to help much the people who are actually processing the > > queue, and requires more work. > > I'm not sure I follow. I tried last week to give some hours for patch > review. What actually happened was that I had to wade through a ton of > stuff, skipping patches that were already applied, threads that were > purely discussion but not patches, threads about earlier versions of > some patches. Yep, that describes the painful process well. I have tried to stay on top of patches already applied but I am doing a clock sweep over the queue and you probably caught a spot I was nearing. The rest is clearly hard to digest. Ideally I could group all related patches together, and separate the pure TODO items from the patches that might be TODO items or might be applied. (Though I do need comments on the pure TODO items too.) The problem is that I know of no easy way to do that and the time to doing it is better spent actually processing/deleting items. > I did manage to find actual patches, but I had then to resort to my own > mailbox to get the actual patch to apply and review. You can download an mbox of the entire queue if that helps. > Now I can't be sure whether the current commitfest is finished or there > are still patches pending review; what patches could use input from me; > what patches I could bounce to the next commitfest. > > Also I don't know what patches we have for the next commitfest. I would > gladly give some time to put up a Wiki page to list the patches for the > next commitfest. Would that be helpful? Uh, I don't think we have any patches for the next commit fest yet. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Bruce Momjian <bruce@momjian.us> writes:
>> Also I don't know what patches we have for the next commitfest. I would
>> gladly give some time to put up a Wiki page to list the patches for the
>> next commitfest. Would that be helpful?
> Uh, I don't think we have any patches for the next commit fest yet.
There are two or three items that we have decided to hold over, although
maybe it is sufficient if they are entered as TODOs.
Also, we *have* received several patches since the start of commit fest.
Somebody needs to check through them and see which are revisions of
patches already in the queue (legit material for this fest) and which
are new stuff (not so much).
regards, tom lane
Bruce Momjian wrote: > Alvaro Herrera wrote: > > I did manage to find actual patches, but I had then to resort to my own > > mailbox to get the actual patch to apply and review. > > You can download an mbox of the entire queue if that helps. Ah, yeah, that's what I did. > > Also I don't know what patches we have for the next commitfest. I would > > gladly give some time to put up a Wiki page to list the patches for the > > next commitfest. Would that be helpful? > > Uh, I don't think we have any patches for the next commit fest yet. We certainly have, for example http://archives.postgresql.org/pgsql-patches/2008-03/msg00245.php Simon has been submitting patches as well. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
Tom Lane wrote: > Bruce Momjian <bruce@momjian.us> writes: > >> Also I don't know what patches we have for the next commitfest. I would > >> gladly give some time to put up a Wiki page to list the patches for the > >> next commitfest. Would that be helpful? > > > Uh, I don't think we have any patches for the next commit fest yet. > > There are two or three items that we have decided to hold over, although > maybe it is sufficient if they are entered as TODOs. > > Also, we *have* received several patches since the start of commit fest. > Somebody needs to check through them and see which are revisions of > patches already in the queue (legit material for this fest) and which > are new stuff (not so much). New stuff, as far as I can tell, but I will recheck. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Alvaro Herrera wrote: > Bruce Momjian wrote: > > Alvaro Herrera wrote: > > > > I did manage to find actual patches, but I had then to resort to my own > > > mailbox to get the actual patch to apply and review. > > > > You can download an mbox of the entire queue if that helps. > > Ah, yeah, that's what I did. > > > > > Also I don't know what patches we have for the next commitfest. I would > > > gladly give some time to put up a Wiki page to list the patches for the > > > next commitfest. Would that be helpful? > > > > Uh, I don't think we have any patches for the next commit fest yet. > > We certainly have, for example > > http://archives.postgresql.org/pgsql-patches/2008-03/msg00245.php > > Simon has been submitting patches as well. I mean we haven't added any for the next one, but I could move them to patches_hold, I guess. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Alvaro Herrera wrote: > > Also I don't know what patches we have for the next commitfest. I would > gladly give some time to put up a Wiki page to list the patches for the > next commitfest. Would that be helpful? > > Yes. My experience is similar to yours. I'm not prepared to spend hours perusing a mailbox again. We have unfortunately been badly underprepared for this. cheers andrew
"Bruce Momjian" <bruce@momjian.us> writes: > Alvaro Herrera wrote: >> >> I'm not sure I follow. I tried last week to give some hours for patch >> review. What actually happened was that I had to wade through a ton of >> stuff, skipping patches that were already applied, threads that were >> purely discussion but not patches, threads about earlier versions of >> some patches. > > Yep, that describes the painful process well. > > I have tried to stay on top of patches already applied but I am doing a > clock sweep over the queue and you probably caught a spot I was nearing. I've also tried to find stuff I could review and failed repeatedly. I have a week now where I could spend a lot of time on this so I'll try again. Am i looking at the right list though? The one I'm looking at still starts with a patch to which Tom commented "This is superseded by a later submission" which I think has been there since pretty much the start of the commitfest. > You can download an mbox of the entire queue if that helps. >> >> Also I don't know what patches we have for the next commitfest. I would >> gladly give some time to put up a Wiki page to list the patches for the >> next commitfest. Would that be helpful? I tried this once already. I dumped the first page of Bruce's queue into a wiki page and laid it out in a column. But I wanted to add links back to the messages in the archive and ran into a problem with that. It also would have been a pain because Bruce's list is divided up into pages. If we modify the archives to have a URL to go straight to a message-id and Bruce generated a list all on one page including comments and message-ids then I could process them all into a wiki page in a few hours. -- Gregory Stark EnterpriseDB http://www.enterprisedb.com Ask me about EnterpriseDB's RemoteDBA services!
Gregory Stark <stark@enterprisedb.com> writes:
> Am i looking at the right list though? The one I'm looking at still starts
> with a patch to which Tom commented "This is superseded by a later submission"
> which I think has been there since pretty much the start of the commitfest.
Yeah, there is still a boatload of stuff that could be removed from the
patch queue. I think that Bruce is intentionally leaving in all traffic
related to an open patch, which is okay by me as long as the obsolete
versions are properly commented. But I wish he'd get on with removing
the stuff that is marked as already committed (some in 8.3!).
> It also would have
> been a pain because Bruce's list is divided up into pages.
That's been a pain in the rear since the start. Bruce, could we combine
it into one page now that it's relatively short?
regards, tom lane
Gregory Stark wrote: > > > "Bruce Momjian" <bruce@momjian.us> writes: > > > Alvaro Herrera wrote: > >> > >> I'm not sure I follow. I tried last week to give some hours for patch > >> review. What actually happened was that I had to wade through a ton of > >> stuff, skipping patches that were already applied, threads that were > >> purely discussion but not patches, threads about earlier versions of > >> some patches. > > > > Yep, that describes the painful process well. > > > > I have tried to stay on top of patches already applied but I am doing a > > clock sweep over the queue and you probably caught a spot I was nearing. > > I've also tried to find stuff I could review and failed repeatedly. I have a > week now where I could spend a lot of time on this so I'll try again. > > Am i looking at the right list though? The one I'm looking at still starts > with a patch to which Tom commented "This is superseded by a later submission" > which I think has been there since pretty much the start of the commitfest. Yes, those emails are kept for background material for the later patch. > > You can download an mbox of the entire queue if that helps. > >> > >> Also I don't know what patches we have for the next commitfest. I would > >> gladly give some time to put up a Wiki page to list the patches for the > >> next commitfest. Would that be helpful? > > I tried this once already. I dumped the first page of Bruce's queue into a > wiki page and laid it out in a column. But I wanted to add links back to the > messages in the archive and ran into a problem with that. It also would have > been a pain because Bruce's list is divided up into pages. Yep, sorry. > If we modify the archives to have a URL to go straight to a message-id and > Bruce generated a list all on one page including comments and message-ids then > I could process them all into a wiki page in a few hours. Ideally we could have a message id link to the official archives. Off list tell me what you want and I will generate it. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Tom Lane wrote: > Gregory Stark <stark@enterprisedb.com> writes: > > Am i looking at the right list though? The one I'm looking at still starts > > with a patch to which Tom commented "This is superseded by a later submission" > > which I think has been there since pretty much the start of the commitfest. > > Yeah, there is still a boatload of stuff that could be removed from the > patch queue. I think that Bruce is intentionally leaving in all traffic > related to an open patch, which is okay by me as long as the obsolete > versions are properly commented. But I wish he'd get on with removing > the stuff that is marked as already committed (some in 8.3!). I am doing a clock sweep through the list, and it takes a week for each cycle. > > > It also would have > > been a pain because Bruce's list is divided up into pages. > > That's been a pain in the rear since the start. Bruce, could we combine > it into one page now that it's relatively short? Sure. Done. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
"Bruce Momjian" <bruce@momjian.us> writes: >> I tried this once already. I dumped the first page of Bruce's queue into a >> wiki page and laid it out in a column. But I wanted to add links back to the >> messages in the archive and ran into a problem with that. It also would have >> been a pain because Bruce's list is divided up into pages. > > Yep, sorry. > >> If we modify the archives to have a URL to go straight to a message-id and >> Bruce generated a list all on one page including comments and message-ids then >> I could process them all into a wiki page in a few hours. > > Ideally we could have a message id link to the official archives. Off > list tell me what you want and I will generate it. Well I was going to let this commit-fest go forward and then try to get what I would want from you when a) the list is down to a manageable size and b) isn't actively being used. What I wanted was a page that included all comments, the message-id of each message, the author of each message, and included everything on one page. The other half of this was modifying the archives so that we had a URL we could pass a message-id to. I think we all agreed that would be a wonderful thing to have regardless of how we tackle this list anyways. Magnus or Dave? Is there anything I can help with to get the URL to message-ids going? -- Gregory Stark EnterpriseDB http://www.enterprisedb.com Ask me about EnterpriseDB's 24x7 Postgres support!
Gregory Stark wrote: > "Bruce Momjian" <bruce@momjian.us> writes: > > >> I tried this once already. I dumped the first page of Bruce's queue into a > >> wiki page and laid it out in a column. But I wanted to add links back to the > >> messages in the archive and ran into a problem with that. It also would have > >> been a pain because Bruce's list is divided up into pages. > > > > Yep, sorry. > > > >> If we modify the archives to have a URL to go straight to a message-id and > >> Bruce generated a list all on one page including comments and message-ids then > >> I could process them all into a wiki page in a few hours. > > > > Ideally we could have a message id link to the official archives. Off > > list tell me what you want and I will generate it. > > Well I was going to let this commit-fest go forward and then try to get what I > would want from you when a) the list is down to a manageable size and b) isn't > actively being used. > > What I wanted was a page that included all comments, the message-id of each > message, the author of each message, and included everything on one page. You can get the message-ids by grep'ing the mbox file, and the comments are all: http://js-kit.com/rss/momjian.us/msgid-* e.g. http://js-kit.com/rss/momjian.us/msgid-200710222037.l9MKbCJZ098744@wwwmaster.postgresql.org So, once you grep the message ids, you can just pull the comments using the above URL with your current message-id. Is there anything else? > The other half of this was modifying the archives so that we had a URL we > could pass a message-id to. I think we all agreed that would be a wonderful > thing to have regardless of how we tackle this list anyways. Agreed. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
Bruce Momjian wrote: > > If we modify the archives to have a URL to go straight to a message-id and > > Bruce generated a list all on one page including comments and message-ids then > > I could process them all into a wiki page in a few hours. > > Ideally we could have a message id link to the official archives. Off > list tell me what you want and I will generate it. Hmm, there's the permalink stuff I mentioned via Jabber. I think you could do this right away, and not wait until the next commitfest. The other thing Greg is saying is to remove paginated output. You can set mhonarc to put all messages in a single page. I'm not sure this is such a hot idea for all users, because such a page would be humongous; but perhaps runnning mhonarc twice is a good idea? -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
Alvaro Herrera wrote: > Bruce Momjian wrote: > > > > If we modify the archives to have a URL to go straight to a message-id and > > > Bruce generated a list all on one page including comments and message-ids then > > > I could process them all into a wiki page in a few hours. > > > > Ideally we could have a message id link to the official archives. Off > > list tell me what you want and I will generate it. > > Hmm, there's the permalink stuff I mentioned via Jabber. I think you > could do this right away, and not wait until the next commitfest. > > The other thing Greg is saying is to remove paginated output. You can > set mhonarc to put all messages in a single page. I'm not sure this is > such a hot idea for all users, because such a page would be humongous; > but perhaps runnning mhonarc twice is a good idea? Done already --- we can always revert it. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
"Bruce Momjian" <bruce@momjian.us> writes: > You can get the message-ids by grep'ing the mbox file, and the comments > are all: > > http://js-kit.com/rss/momjian.us/msgid-* > > e.g. > > http://js-kit.com/rss/momjian.us/msgid-200710222037.l9MKbCJZ098744@wwwmaster.postgresql.org Huh, fascinating. it's actually an RSS feed. I can fool around with this. I should be able to snarf in the mbox and format a wiki page with the comments from js-kit and links to the message-id. -- Gregory Stark EnterpriseDB http://www.enterprisedb.com Ask me about EnterpriseDB's 24x7 Postgres support!
Gregory Stark wrote: > "Bruce Momjian" <bruce@momjian.us> writes: > > > You can get the message-ids by grep'ing the mbox file, and the comments > > are all: > > > > http://js-kit.com/rss/momjian.us/msgid-* > > > > e.g. > > > > http://js-kit.com/rss/momjian.us/msgid-200710222037.l9MKbCJZ098744@wwwmaster.postgresql.org > > Huh, fascinating. it's actually an RSS feed. > > I can fool around with this. I should be able to snarf in the mbox and format > a wiki page with the comments from js-kit and links to the message-id. Alvaro reminded me that '@' can be used in a URL, just not in the host name part, so I have converted over to using messsge-ids for permanent URLs, e.g.: http://momjian.us/mhonarc/patches/44DA31B1.3090700@enterprisedb.com.html I will keep the MD5 URLs too as well until this commit fest is over. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +
On Mon, Mar 24, 2008 at 06:39:25PM -0400, Andrew Dunstan wrote: > perusing a mailbox again. We have unfortunately been badly underprepared > for this. Surely that there is an emerging consensus to that effect means that it's not as unfortunate as it might be? I seem to recall the original announcement suggesting this was an experiment. I wouldn't expect the first couple rounds to go without a hitch; as long as there is procedural improvement the next time, that's a good thing, right? A
Andrew Sullivan <ajs@crankycanuck.ca> writes:
> On Mon, Mar 24, 2008 at 06:39:25PM -0400, Andrew Dunstan wrote:
>> perusing a mailbox again. We have unfortunately been badly underprepared
>> for this.
> Surely that there is an emerging consensus to that effect means that it's
> not as unfortunate as it might be? I seem to recall the original
> announcement suggesting this was an experiment. I wouldn't expect the first
> couple rounds to go without a hitch; as long as there is procedural
> improvement the next time, that's a good thing, right?
Yeah, we expected to have glitches. I think we now have a much better
idea what sort of status-tracking support we need for future fests.
regards, tom lane
Tom Lane wrote: > Andrew Sullivan <ajs@crankycanuck.ca> writes: > >> On Mon, Mar 24, 2008 at 06:39:25PM -0400, Andrew Dunstan wrote: >> >>> perusing a mailbox again. We have unfortunately been badly underprepared >>> for this. >>> > > >> Surely that there is an emerging consensus to that effect means that it's >> not as unfortunate as it might be? I seem to recall the original >> announcement suggesting this was an experiment. I wouldn't expect the first >> couple rounds to go without a hitch; as long as there is procedural >> improvement the next time, that's a good thing, right? >> > > Yeah, we expected to have glitches. I think we now have a much better > idea what sort of status-tracking support we need for future fests. > > > Yes. I'm not meaning to whine, sorry if it comes over like that. It looks to me like we need a sort of prep phase for a commit-fest, so the people switching into commit-fest mode when it starts can do so with little friction. cheers andrew
Andrew Dunstan wrote: > >> Surely that there is an emerging consensus to that effect means that it's > >> not as unfortunate as it might be? I seem to recall the original > >> announcement suggesting this was an experiment. I wouldn't expect the first > >> couple rounds to go without a hitch; as long as there is procedural > >> improvement the next time, that's a good thing, right? > >> > > > > Yeah, we expected to have glitches. I think we now have a much better > > idea what sort of status-tracking support we need for future fests. > > > > > > > > Yes. I'm not meaning to whine, sorry if it comes over like that. It > looks to me like we need a sort of prep phase for a commit-fest, so the > people switching into commit-fest mode when it starts can do so with > little friction. Yea, I think we are still learning, and polishing our tools. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://postgres.enterprisedb.com + If your life is a hard drive, Christ can be your backup. +