Mailing List archive corruption?
Hi, When I look at the mailing list archive for python-dev, I see some odd stuff at the bottom of the page: http://mail.python.org/pipermail/python-dev/2010-January/thread.html#95232 Anyone know what's happened? Regards, Vinay Sajip
On Jan 19, 2010, at 03:50 PM, Vinay Sajip wrote:
When I look at the mailing list archive for python-dev, I see some odd stuff at the bottom of the page:
http://mail.python.org/pipermail/python-dev/2010-January/thread.html#95232
Anyone know what's happened?
WTF? I think the archives were recently regenerated, so there's probably a fubar there. CC'ing the postmasters. -Barry
On Jan 19, 2010, at 11:07 AM, Barry Warsaw wrote:
On Jan 19, 2010, at 03:50 PM, Vinay Sajip wrote:
When I look at the mailing list archive for python-dev, I see some odd stuff at the bottom of the page:
http://mail.python.org/pipermail/python-dev/2010-January/thread.html#95232
Anyone know what's happened?
WTF? I think the archives were recently regenerated, so there's probably a fubar there. CC'ing the postmasters.
That happens if messages had unescaped "From" lines in the middle of them. No doubt, you've now broken every link anyone had ever made into the python-dev archives, because now all the article numbers are different. BTDT...unfortunately... Pipermail really is quite crappy, sigh. Anyhow, when I did that, I went back to a backup to get the original article numbers, and edited the mbox file escaping From lines or adding additional empty messages until the newly regenerated article numbers matched the originals. I'd highly recommend going through that painful process, since I suspect a *lot* of people have links to the python-dev archive. Hope you have a backup (or can find caches on google or archive.org or something). James
On Jan 19, 2010, at 11:24 AM, James Y Knight wrote:
No doubt, you've now broken every link anyone had ever made into the python-dev archives, because now all the article numbers are different. BTDT...unfortunately... Pipermail really is quite crappy, sigh.
I've been trying for 10+ years to get folks interested in helping me fix this (and a few other warts) in Pipermail, to not much success. ;/
Anyhow, when I did that, I went back to a backup to get the original article numbers, and edited the mbox file escaping From lines or adding additional empty messages until the newly regenerated article numbers matched the originals. I'd highly recommend going through that painful process, since I suspect a *lot* of people have links to the python-dev archive. Hope you have a backup (or can find caches on google or archive.org or something).
bin/cleanarch uses a set of heuristics to find unescaped From lines and fix them. It's generally pretty good, but it certain can change message numbers (and sadly, their urls). -Barry
On Tue, 19 Jan 2010 11:24:57 -0500, James Y Knight
On Jan 19, 2010, at 11:07 AM, Barry Warsaw wrote:
On Jan 19, 2010, at 03:50 PM, Vinay Sajip wrote:
When I look at the mailing list archive for python-dev, I see some odd stuff at the bottom of the page:
http://mail.python.org/pipermail/python-dev/2010-January/thread.html#95232
Anyone know what's happened?
WTF? I think the archives were recently regenerated, so there's probably a fubar there. CC'ing the postmasters.
That happens if messages had unescaped "From" lines in the middle of them.
No doubt, you've now broken every link anyone had ever made into the python-dev archives, because now all the article numbers are different. BTDT...unfortunately... Pipermail really is quite crappy, sigh.
Anyhow, when I did that, I went back to a backup to get the original article numbers, and edited the mbox file escaping From lines or adding additional empty messages until the newly regenerated article numbers matched the originals. I'd highly recommend going through that painful process, since I suspect a *lot* of people have links to the python-dev archive. Hope you have a backup (or can find caches on google or archive.org or something).
The Python issue tracker does, for one. -- R. David Murray www.bitdance.com Business Process Automation - Network/Server Management - Routers/Firewalls
participants (5)
-
Barry Warsaw -
James Y Knight -
Nick Coghlan -
R. David Murray -
Vinay Sajip