[Mailman-Users] Re Problems with "more info" Hyperlink from ListArchives Page

[Mark Sapiro]

Barry Finkel wrote:

>I wrote, in part,
>
>>>...
>>>I have 13 lists.  For each list I checked the URL chains from the
>>>
>>>     https://lists.example.com/mailman/listinfo
>>>
>>>Mailman "home page", and for five of the lists I see URL problems.
>>>
>>>...
>>>
>>>Note that the Mailman list archives page lacks a "/" between "mailman"
>>>and "listinfo".  For each of these bad lists I ran
>>>
>>>     /usr/lib/mailman/bin/withlist -l -r fix_url $list -u lists.example.com
>>>
>>>but there were no changes to the URLs.
>>>...
>
>and "Richard Rae" <Richard_Rae at xyratex.com> replied:
>
>>You might need to run the fix_url command
>>
>>~/bin/withlist -l -r fix_url listname
>
>Is this any different than the command I ran?  I tried the
>
>     fix_url
>
>command without the
>
>     -u lists.example.com
>
>parameter, and the URLs did not change.


Yes, it is different, but that is not the current problem (or maybe it
was, but more on that in a paragraph).

The -u option sets the url_host in case that list's url_host is
different from DEFAULT_URL_HOST, which I gather in your case it is not.

Here's what I think the issue is. When you originally changed
DEFAULT_URL_PATTERN for the https scheme you left off the trailing
slash and added cgi-bin/as in

DEFAULT_URL_PATTERN = 'https://%s/cgi-bin/mailman'
(when the 'mailman' list was created)

also, apparently at some point you had other things like

DEFAULT_URL_PATTERN = 'https://%s/cgi-bin/mailman/'
(bad3 through bad 6)
and

DEFAULT_URL_PATTERN = 'https://%s/'
(bad2)


You later corrected that to

DEFAULT_URL_PATTERN = 'https://%s/mailman/'

and (perhaps later still) ran fix_url which fixed the hidden
web_page_url attribute on these six lists.

However, the archive pages are static HTML pages. Current index and TOC
pages are constantly rebuilt as messages are added to the archives,
but pages that haven't been updated since before you ran fix_url will
have bad links as you observe.

There are a few ways to fix this. The easiest is to run 'bin/arch
--wipe' on all the problem lists, but if these lists have archives
going back to older Mailman versions (I'm not sure how old is a
problem, but maybe 2.0.x, maybe pre 2.1.4), there are some potential
issues in that there may be unescaped "From " lines in message bodies
that are taken as mbox message separators by the archiver
(bin/cleanarch can help find/fix these). Also, there is a possibility
that messages may get renumbered in the archive which is not a problem
in itself, but it invalidates any saved URLs linking to individual
archive messages.

If you don't want to run 'bin/arch --wipe' you can ignore the problem.
The main TOC page and the current period's index pages should get
fixed the next time a message is archived for the list, and new index
pages will be built correctly. If older index pages are a concern,
they can be edited with a script.

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan