Solved - Re: [Mailman-Users] mailmanlistinfo without slash

[John Fleming]

----- Original Message ----- 
From: "Mark Sapiro" <msapiro at value.net>
To: "John Fleming" <john at wa9als.com>; "Mark Sapiro" <msapiro at value.net>; 
<mailman-users at python.org>
Sent: Friday, January 28, 2005 3:00 PM
Subject: Re: [Mailman-Users] mailmanlistinfo without slash


> John Fleming wrote:
>>
>>----- Original Message ----- 
>>From: "Mark Sapiro" <msapiro at value.net>
> <snip>
>>>
>>> bin/arch --wipe sked /dev/null
>>>
>>> or whatever your other list names might be.
>>>
>>> I tested that this time and it works :-)
>>
>>
>>The above seemed to work.  I restarted Mailman, and then when clicking the
>>Archives link on the Listinfo page, I was taken to
>>Forbidden
>>You don't have permission to access /pipermail/sked/ on this server.
>>
>>I thought maybe I just needed to rebuild the archive, so I did that, but
>>same error.  I looked at the permissions of
>>/var/lib/mailman/archives/private/sked.mbox and it is the same as my other
>>lists that don't give the FORBIDDEN error - group-list, owner www-data, 
>>2775
>>(set GID).  Which page exactly has incorrect permissions?
>
>
> /var/lib/mailman/archives/private/sked/index.html
>
> The web accessable archive is in the sked/ directory. The sked.mbox/
> directory only contains the sked.mbox file which is the archive in
> mbox format.
>>
>>Here's the bin/arch run:
>>
>>Luke:/var/lib/mailman/bin# ./arch --wipe sked /dev/null
>>Traceback (most recent call last):
>>  File "./arch", line 187, in ?
>>    main()
>>  File "./arch", line 166, in main
>>    shutil.rmtree(mlist.archive_dir())
>>  File "/usr/lib/python2.3/shutil.py", line 142, in rmtree
>>    raise exc[0], (exc[1][0], exc[1][1] + ' removing '+arg)
>>OSError: [Errno 2] No such file or directory removing
>>/var/lib/mailman/archives/private/sked
>>Luke:/var/lib/mailman/bin#
>
> It looks like there was no existing archive when you ran
>
> bin/arch --wipe sked /dev/null
>
> so the command failed and didn't build a new archive. The archive is
> gone because the previous
>
> bin/arch --wipe sked
>
> which failed on the lack of the sked.mbox file had removed the archive
> before failing.
>
> So now try
>
> bin/arch sked /dev/null
>
> i.e. without the --wipe. This should build a new, empty archive.

I think that's right.  I've done several things since my last post, and now 
everything's working great and life is good!  I'd like to summarize for my 
own record and in case there's anyone as inept as me lurking here:

1.  The original problem was that on a list's archive page, there was an 
incorrect link to "More Information".  That link was missing a slash, i.e. 
/mailmanlistinfo/listname.  Several ways to handle that were mentioned, but 
I ended up saving the lists' .pck files and then deleting the lists and 
recreating and substituting the old .pck files.  That fixed the original 
problem, and the missing slash appeared. However, there were now some 
virtual domain/weird link problems.

When I went to list A's archive page, the More Information link was to 
another domain, the main domain of the site.  If I clicked this link, I 
actually went to the correct page for list A, even with the correct domain 
(although technically incorrect in the link).  When I started looking around 
and messing with other lists, things got really confusing!

First I was told I should use fix_url.py.  Then we decided I would also need 
to rebuild the archive after running fix_url.py to get the link corrected on 
the Archives page.

3.  After some learning of how to use withlist, fix_url, and arch, I 
methodically went through each of my lists (I only have 4), running first 
fix_url and then rebuilding the archive.  This got confusing to this newbie 
because when I rebuilt the archive, the ownership of the archive directory 
changed to the user I was running as rather than www-data.  The owner of the 
archive has to be www-data or apache wouldn't let me surf that page later.

4.  So for each list, I ran:

bin/withlist -r -l fix_url listname --urlhost=desireddomain

bin/arch --wipe listname /dev/null

Then I changed the ownership of /archives/listname (and maybe 
/archives/listname.mbox) to www-data.  These were both group list and didn't 
need to be changed.

Then I restarted mailman (unnecessary?)

Now when I click on the Archives link on a listinfo page, the link contains 
the proper slashes, and I am taken to a page that says there are no 
archives.  The link for More Information on that archive page uses the 
correct domain!  YAY!  Life is good!

I'm sure most of you knew all of this, but I've learned a lot.  FAQ or no, 
when someone tells a newbie to run fix_url.py, you'd be surprised how 
difficult it can turn out to be.  ;-)  Thanks especially to Mark and Tokio 
and to everyone for your patience and bandwidth.  - John



-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.300 / Virus Database: 265.8.1 - Release Date: 1/27/2005