Hi,
I often search the archives using "foo bar site:lists.puredata.info" in Google (since the search engine of the archive is pretty useless). Usually it works fine.
However, today I tried a few searches, and 100% of the times, when I click on a search result returned by Google, it takes me to a totally unrelated message.
For example if you search for "berlin site:lists.puredata.info", the first search result is "[PD] [PD-announce] Blank Pages Berlin - search for a place" but if you click on it, it takes you to: "[PD] Re : Split multichannel audio file into stereo pairs"
Any idea what may be going on??
Is there any other mirror hosting the archives so that I could try searching there instead of lists.puredata.info?
Thanks m.
It seems the problem is indeed in the archive, regardless of Google. I've just tried searching the archive directly, and the results are complete nonsense: not the usual nonsense (e.g. returning an unrelated message as a result just because the keyword appears in the "next message" title, for example) but complete junk.
Is anybody in the list involved in the maintenance of the mailing list and have any idea whether this is going to be fixed?
thanks m.
On 06/29/2010 06:52 PM, Matteo Sisti Sette wrote:
Hi,
I often search the archives using "foo bar site:lists.puredata.info" in Google (since the search engine of the archive is pretty useless). Usually it works fine.
However, today I tried a few searches, and 100% of the times, when I click on a search result returned by Google, it takes me to a totally unrelated message.
For example if you search for "berlin site:lists.puredata.info", the first search result is "[PD] [PD-announce] Blank Pages Berlin - search for a place" but if you click on it, it takes you to: "[PD] Re : Split multichannel audio file into stereo pairs"
Any idea what may be going on??
Is there any other mirror hosting the archives so that I could try searching there instead of lists.puredata.info?
Thanks m.
I noticed the problem as well. I guess what happens is that the message numbering has changed for some reason. Probably the last time google was crawling http://lists.puredata.info/pipermail/pd-list/2009-06/070918.html it actually still contained the post about Blank Pages.
Don't know how mailman works internally, but probably the message-IDs are not persistent (when you move archives around or delete some messages? I don't know).
Roman
On Tue, 2010-06-29 at 18:52 +0200, Matteo Sisti Sette wrote:
Hi,
I often search the archives using "foo bar site:lists.puredata.info" in Google (since the search engine of the archive is pretty useless). Usually it works fine.
However, today I tried a few searches, and 100% of the times, when I click on a search result returned by Google, it takes me to a totally unrelated message.
For example if you search for "berlin site:lists.puredata.info", the first search result is "[PD] [PD-announce] Blank Pages Berlin - search for a place" but if you click on it, it takes you to: "[PD] Re : Split multichannel audio file into stereo pairs"
Any idea what may be going on??
Is there any other mirror hosting the archives so that I could try searching there instead of lists.puredata.info?
On 2010-06-30 00:15, Roman Haefeli wrote:
I noticed the problem as well. I guess what happens is that the message numbering has changed for some reason. Probably the last time google was crawling http://lists.puredata.info/pipermail/pd-list/2009-06/070918.html it actually still contained the post about Blank Pages.
Don't know how mailman works internally, but probably the message-IDs are not persistent (when you move archives around or delete some messages? I don't know).
due to the recent outage of the mailing list archives, i had to rebuilt the list archives which eventually renumbered the pages. i'm aware of the problem, but don't have any solution ready at hand. (what's worse, if the same thing happens again, i don't know of any simple solution to rebuild the archives without renumbering again)
i don't even know exactly why the archiver renumbers differently every time i rebuild. i guess it has to do with wrong sender times in the mail-headers.
anyhow, the google search index will of course re-adapt itself within reasonable time. the namazu index will have to be rebuild manually.
sorry for the inconvenience.
fgmasdr IOhannes
On Wed, 2010-06-30 at 10:17 +0200, IOhannes m zmoelnig wrote:
On 2010-06-30 00:15, Roman Haefeli wrote:
I noticed the problem as well. I guess what happens is that the message numbering has changed for some reason. Probably the last time google was crawling http://lists.puredata.info/pipermail/pd-list/2009-06/070918.html it actually still contained the post about Blank Pages.
Don't know how mailman works internally, but probably the message-IDs are not persistent (when you move archives around or delete some messages? I don't know).
due to the recent outage of the mailing list archives, i had to rebuilt the list archives which eventually renumbered the pages. i'm aware of the problem, but don't have any solution ready at hand. (what's worse, if the same thing happens again, i don't know of any simple solution to rebuild the archives without renumbering again)
i don't even know exactly why the archiver renumbers differently every time i rebuild. i guess it has to do with wrong sender times in the mail-headers.
anyhow, the google search index will of course re-adapt itself within reasonable time. the namazu index will have to be rebuild manually.
sorry for the inconvenience.
No problem for me.
** Many thanks ** for going through the trouble of hosting the mailing list.
Roman
On Tue, 29 Jun 2010, Matteo Sisti Sette wrote:
However, today I tried a few searches, and 100% of the times, when I click on a search result returned by Google, it takes me to a totally unrelated message.
But isn't it also that the message you want is always in the same folder (representing a month of messages), and then, what do you find the difference of message-number to be ? Is it nearly always the same value ? This might tell you how to do a temporary workaround.
Meanwhile, I just do a plain-text search through the whole collection of the mbox files (that I update every few months whenever I need it, since a few years), but it doesn't always work, because plain-text is just very close to the mbox email format, not exactly the same.
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard, Montréal, Québec. téléphone: +1.514.383.3801
-
IOhannes m zmoelnig -
Mathieu Bouchard -
Matteo Sisti Sette -
Roman Haefeli