Thread: Archive Message-Id links broken?
I don't use the archives that much to know the normal behavior, but I just noticed that the Message-Id links point to the non-Message-Id URL (like msg02010.php). That can't be the point, can it?
Peter Eisentraut wrote: > I don't use the archives that much to know the normal behavior, but I > just noticed that the Message-Id links point to the non-Message-Id URL > (like msg02010.php). That can't be the point, can it? No, it isn't -- the problem is that Mhonarc changed behavior midflight because I disabled email mangling for strings that looked like message-ids. I hadn't given much time to the archives lately, but it should be relatively easy to fix. -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
On Sun, Aug 30, 2009 at 8:03 PM, Alvaro Herrera<alvherre@commandprompt.com> wrote: > Peter Eisentraut wrote: >> I don't use the archives that much to know the normal behavior, but I >> just noticed that the Message-Id links point to the non-Message-Id URL >> (like msg02010.php). That can't be the point, can it? > > No, it isn't -- the problem is that Mhonarc changed behavior midflight > because I disabled email mangling for strings that looked like > message-ids. I hadn't given much time to the archives lately, but it > should be relatively easy to fix. I'm mildly surprised to learn that this is unintentional... I use it semi-regularly as a way of getting a shorter and more legible URL for a message in the archives. Not a big deal, though. ...Robert
Robert Haas escribió: > On Sun, Aug 30, 2009 at 8:03 PM, Alvaro > Herrera<alvherre@commandprompt.com> wrote: > > Peter Eisentraut wrote: > >> I don't use the archives that much to know the normal behavior, but I > >> just noticed that the Message-Id links point to the non-Message-Id URL > >> (like msg02010.php). That can't be the point, can it? > > > > No, it isn't -- the problem is that Mhonarc changed behavior midflight > > because I disabled email mangling for strings that looked like > > message-ids. I hadn't given much time to the archives lately, but it > > should be relatively easy to fix. > > I'm mildly surprised to learn that this is unintentional... I use it > semi-regularly as a way of getting a shorter and more legible URL for > a message in the archives. I guess we could provide both. I vaguely remember someone asking how to get this URL some time ago. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
On Mon, Aug 31, 2009 at 1:57 AM, Alvaro Herrera<alvherre@commandprompt.com> wrote: >> I'm mildly surprised to learn that this is unintentional... I use it >> semi-regularly as a way of getting a shorter and more legible URL for >> a message in the archives. > > I guess we could provide both. I vaguely remember someone asking how to > get this URL some time ago. The message-id url is far more useful generally. It's guaranteed not to change if we reprocess the mail archives -- even if we change mail archiving systems altogether. More importantly it can be used to track down the message even if the mail archives are down or inaccessible. The msg02010.php style url is what you get from our search engine or index listing. I didn't think it was hard to get these urls. How are you *not* getting one these urls in the first place? -- greg http://mit.edu/~gsstark/resume.pdf
Greg Stark escribió: > On Mon, Aug 31, 2009 at 1:57 AM, Alvaro > Herrera<alvherre@commandprompt.com> wrote: > >> I'm mildly surprised to learn that this is unintentional... I use it > >> semi-regularly as a way of getting a shorter and more legible URL for > >> a message in the archives. > > > > I guess we could provide both. I vaguely remember someone asking how to > > get this URL some time ago. > > The message-id url is far more useful generally. It's guaranteed not > to change if we reprocess the mail archives -- even if we change mail > archiving systems altogether. More importantly it can be used to track > down the message even if the mail archives are down or inaccessible. > > The msg02010.php style url is what you get from our search engine or > index listing. I didn't think it was hard to get these urls. How are > you *not* getting one these urls in the first place? The number-based URLs are generated by Mhonarc internally. It doesn't generate the message-id-based URLs, and there's no way to make it emit them, so they are generated by an external program that broke when I changed the Mhonarc config some time ago after Tom complained that it broke URLs that ended in message-ids in the message body. This is all crap of course, which is why we want to replace Mhonarc with a database-backed system. -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
On Sun, Aug 30, 2009 at 9:32 PM, Greg Stark<gsstark@mit.edu> wrote: > On Mon, Aug 31, 2009 at 1:57 AM, Alvaro > Herrera<alvherre@commandprompt.com> wrote: >>> I'm mildly surprised to learn that this is unintentional... I use it >>> semi-regularly as a way of getting a shorter and more legible URL for >>> a message in the archives. >> >> I guess we could provide both. I vaguely remember someone asking how to >> get this URL some time ago. > > The message-id url is far more useful generally. It's guaranteed not > to change if we reprocess the mail archives -- even if we change mail > archiving systems altogether. More importantly it can be used to track > down the message even if the mail archives are down or inaccessible. > > The msg02010.php style url is what you get from our search engine or > index listing. I didn't think it was hard to get these urls. How are > you *not* getting one these urls in the first place? Well, typically what I do is find the message by searching my gmail account, extract the message ID, and use that to construct an archives.postgresql.org/message-id/ style link. You may ask - why do I search my gmail account rather than searching for the message in some other fashion? Well, because that's what I've found tends to work the best. YMMV. ...Robert
On Sun, Aug 30, 2009 at 10:47 PM, Alvaro Herrera<alvherre@commandprompt.com> wrote: > Greg Stark escribió: >> On Mon, Aug 31, 2009 at 1:57 AM, Alvaro >> Herrera<alvherre@commandprompt.com> wrote: >> >> I'm mildly surprised to learn that this is unintentional... I use it >> >> semi-regularly as a way of getting a shorter and more legible URL for >> >> a message in the archives. >> > >> > I guess we could provide both. I vaguely remember someone asking how to >> > get this URL some time ago. >> >> The message-id url is far more useful generally. It's guaranteed not >> to change if we reprocess the mail archives -- even if we change mail >> archiving systems altogether. More importantly it can be used to track >> down the message even if the mail archives are down or inaccessible. >> >> The msg02010.php style url is what you get from our search engine or >> index listing. I didn't think it was hard to get these urls. How are >> you *not* getting one these urls in the first place? > > The number-based URLs are generated by Mhonarc internally. It doesn't > generate the message-id-based URLs, and there's no way to make it emit > them, so they are generated by an external program that broke when I > changed the Mhonarc config some time ago after Tom complained that it > broke URLs that ended in message-ids in the message body. > > This is all crap of course, which is why we want to replace Mhonarc with > a database-backed system. Database? You mean like Excel? ...Robert
Robert Haas escribió: > On Sun, Aug 30, 2009 at 10:47 PM, Alvaro > Herrera<alvherre@commandprompt.com> wrote: > > This is all crap of course, which is why we want to replace Mhonarc with > > a database-backed system. > > Database? You mean like Excel? Huh. You know, Mhonarc has something that it calls "database" which is actually a bunch of Perl files with a dump of internal variables describing the mailbox. The fact that it has to read and keep in memory all those files is the reason it doesn't scale well past a month's worth of email. Perhaps it could cope with a year, provided you have a lot of RAM; but clearly it won't work well if we want to handle our complete archive. -- Alvaro Herrera http://www.CommandPrompt.com/ PostgreSQL Replication, Consulting, Custom Development, 24x7 support
Robert Haas escribió: > You may ask - why do I search my gmail account rather than searching > for the message in some other fashion? Well, because that's what I've > found tends to work the best. YMMV. I don't think it's a surprise that Google works much better at searching than the alternatives, really. Still, it's a pity that Gmail doesn't readily expose the message-id and you have to waste time digging it out. Did someone request this to Google? -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
On Mon, Aug 31, 2009 at 10:24 AM, Alvaro Herrera<alvherre@commandprompt.com> wrote: > Robert Haas escribió: >> On Sun, Aug 30, 2009 at 10:47 PM, Alvaro >> Herrera<alvherre@commandprompt.com> wrote: > >> > This is all crap of course, which is why we want to replace Mhonarc with >> > a database-backed system. >> >> Database? You mean like Excel? > > Huh. You know, Mhonarc has something that it calls "database" which is > actually a bunch of Perl files with a dump of internal variables > describing the mailbox. The fact that it has to read and keep in memory > all those files is the reason it doesn't scale well past a month's worth > of email. Perhaps it could cope with a year, provided you have a lot of > RAM; but clearly it won't work well if we want to handle our complete > archive. Go ahead, ignore my witty (?) remark and come back with a serious response. There seems to be a whole lot of project infrastructure that is a bit long in the tooth. Of course, it's hard to get up the enthusiasm to address the problems since it all basically works, so the effort/reward ratio is a bit on the low side. ...Robert
On Mon, Aug 31, 2009 at 1:12 PM, Robert Haas<robertmhaas@gmail.com> wrote: > On Mon, Aug 31, 2009 at 10:24 AM, Alvaro > Herrera<alvherre@commandprompt.com> wrote: >> Robert Haas escribió: >>> On Sun, Aug 30, 2009 at 10:47 PM, Alvaro >>> Herrera<alvherre@commandprompt.com> wrote: >> >>> > This is all crap of course, which is why we want to replace Mhonarc with >>> > a database-backed system. >>> >>> Database? You mean like Excel? >> >> Huh. You know, Mhonarc has something that it calls "database" which is >> actually a bunch of Perl files with a dump of internal variables >> describing the mailbox. The fact that it has to read and keep in memory >> all those files is the reason it doesn't scale well past a month's worth >> of email. Perhaps it could cope with a year, provided you have a lot of >> RAM; but clearly it won't work well if we want to handle our complete >> archive. > > Go ahead, ignore my witty (?) remark and come back with a serious response. > > There seems to be a whole lot of project infrastructure that is a bit > long in the tooth. Of course, it's hard to get up the enthusiasm to > address the problems since it all basically works, so the > effort/reward ratio is a bit on the low side. By which I mean the high side. ...Robert