I copied a list’s mbox to a new server and regenerated the archive.
Unfortunately threading is not rendered satisfyingly, here are the two
Mailman/pipermail versions:
Old: http://lists.textdrive.com/pipermail/textmate/2008-January/thread.html
New: http://lists.macromates.com/textmate/2008-January/thread.htmlThe new one is using wrong indent levels for letters.
On 6/17/08, Allan Odgaard wrote:
I copied a list's mbox to a new server and regenerated the archive.
Unfortunately threading is not rendered satisfyingly, here are the two Mailman/pipermail versions:
Your threading got broken somehow, but I can't tell you how. This would need further investigation.
One thing that strikes me is to ask whether you copied the "cooked" text archives or the raw mbox archives, before regeneration?
-- Brad Knowles <brad@shub-internet.org> LinkedIn Profile: <http://tinyurl.com/y8kpxu>
On 17 Jun 2008, at 06:01, Brad Knowles wrote:
On 6/17/08, Allan Odgaard wrote:
I copied a list's mbox to a new server and regenerated the archive.
Unfortunately threading is not rendered satisfyingly, here are the two Mailman/pipermail versions:
Your threading got broken somehow, but I can't tell you how. This
would need further investigation.One thing that strikes me is to ask whether you copied the "cooked"
text archives or the raw mbox archives, before regeneration?
Raw mbox file.
And new letters sent to the list also show up with wrong indent in the
web archive.
Allan Odgaard wrote:
Raw mbox file.
And new letters sent to the list also show up with wrong indent in the
web archive.
I'm only guessing, but I think there must have been some issue with the mbox file that caused the archive database file to be messed up.
Does a new list work OK?
What does bin/cleanarch say about the mbox file? Did bin/arch report anything odd?
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 18 Jun 2008, at 17:45, Mark Sapiro wrote:
Allan Odgaard wrote:
Raw mbox file.
And new letters sent to the list also show up with wrong indent in
the web archive.I'm only guessing, but I think there must have been some issue with
the mbox file that caused the archive database file to be messed up.Does a new list work OK?
A new list show the same problem.
What does bin/cleanarch say about the mbox file? Did bin/arch report anything odd?
I don’t recall so. But since this happens both for new letters to this
list, and for a completely new list I created a few days ago, I don’t
think the problem is the mbox.
I’ll see if I can get access to another Mailman version and generate
the archive from there.
If I didn’t state already, this is Mailman 2.1.9 under Ubuntu
(installed via aptitude).
Allan Odgaard wrote:
On 18 Jun 2008, at 17:45, Mark Sapiro wrote:
Does a new list work OK?
A new list show the same problem.
What does bin/cleanarch say about the mbox file? Did bin/arch report anything odd?
I dont recall so. But since this happens both for new letters to this
list, and for a completely new list I created a few days ago, I dont
think the problem is the mbox.Ill see if I can get access to another Mailman version and generate
the archive from there.If I didnt state already, this is Mailman 2.1.9 under Ubuntu
(installed via aptitude).
I have never before seen a report like this. Since it occurs for a new list, there is apparently something wrong with the Mailman installation itself. If it were a problem with the Ubuntu package, I think it would have been seen elsewhere, so I think it's likely specific to your installation, but I have no idea what might cause it.
Pipermail does threading by Message-ID and by In-Reply-To and References: headers, so assuming that those are normal and people aren't hijacking threads, I haven't a clue.
-- Mark Sapiro <mark@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better use your sense - B. Dylan
On 18 Jun 2008, at 22:46, Mark Sapiro wrote:
[...] I have never before seen a report like this. Since it occurs for a new list, there is apparently something wrong with the Mailman installation itself. If it were a problem with the Ubuntu package, I think it would have been seen elsewhere, so I think it's likely specific to your installation, but I have no idea what might cause it.
I installed same version of Mailman locally (OS X) and generated the
archive there, from the same mbox, and here it shows up fine.
So definitely there is something on the server causing the problem. I
tried reverting mm_cfg.py to the default version, but that did not fix
anything.
Looks like I’ll have to go through the pipermail source to solve this…
On 19 Jun 2008, at 07:00, Allan Odgaard wrote:
On 18 Jun 2008, at 22:46, Mark Sapiro wrote:
[...] I have never before seen a report like this. Since it occurs for a
new list, there is apparently something wrong with the Mailman installation itself. If it were a problem with the Ubuntu package, I think it would have been seen elsewhere, so I think it's likely specific to your installation, but I have no idea what might cause
it.I installed same version of Mailman locally (OS X) and generated the
archive there, from the same mbox, and here it shows up fine. [...]
Doing a diff on the generated archives, the only problem with the bad
one is that it has the <UL> and </UL>’s placed wrong.
It tends to be very eager with <UL> so everything gets down to third
level…
On 6/19/08, Allan Odgaard wrote:
Doing a diff on the generated archives, the only problem with the bad one is that it has the <UL> and </UL>'s placed wrong.
It tends to be very eager with <UL> so everything gets down to third level
That could be an issue with the Python libraries. Depending on which version of Python and which version of which libraries you have, they might munge things like that.
Of course, I've never heard of this particular type of munging before, but certainly I've heard of weird stuff happening when you have the wrong version of libraries installed.
-- Brad Knowles <brad@shub-internet.org> LinkedIn Profile: <http://tinyurl.com/y8kpxu>
participants (3)
-
Allan Odgaard -
Brad Knowles -
Mark Sapiro