Trailing slash is stripped from List-Archive: URL

[Mark Sapiro]

In creating the List-Archive: header, CookHeaders.py strips a trailing slash from the URL returned by mlist.GetBaseArchiveURL(). It has always (in 2.1) done this, but it is wrong.

Consider a list's archive TOC page which has entries like the following:

<tr>
            <td>March 2012:</td>
            <td>
              <A href="2012-March/thread.html">[ Thread ]</a>
              <A href="2012-March/subject.html">[ Subject ]</a>
              <A href="2012-March/author.html">[ Author ]</a>
              <A href="2012-March/date.html">[ Date ]</a>
            </td>
            <td><A href="2012-March.txt">[ Text 766 KB ]</a></td>
            </tr>

If the base archive URL is http://www.example.com/private/listname/ and one goes to that URL, the relative hrefs all work fine, but if one goes to http://www.example.com/private/listname instead, those relative hrefs become relative to private/ instead of private/listname/, and one gets 'no such list: 2012-march' errors (or 404 errors in the case of public 'pipermail' URLs).

[Mark Sapiro (msapiro)]

I don't think this issue actually affects public (pipermail) URLs because in those cases the web server is serving static HTML and knows the /listname is a directory. It also can be worked around for private URLs by putting an appropriate rewrite rule in the web server, but this fix is still correct, and shouldn't cause any harm.

[Bernard (unowned60)]

if i am outdoors, GNU-Mailman??

2012/3/25 Mark Sapiro <email address hidden>

> I don't think this issue actually affects public (pipermail) URLs
> because in those cases the web server is serving static HTML and knows
> the /listname is a directory. It also can be worked around for private
> URLs by putting an appropriate rewrite rule in the web server, but this
> fix is still correct, and shouldn't cause any harm.
>
> ** Changed in: mailman
> Status: New => Fix Committed
>
> --
> You received this bug notification because you are subscribed to GNU
> Mailman.
> Matching subscriptions: 24/7
> https://bugs.launchpad.net/bugs/964190
>
> Title:
> Trailing slash is stripped from List-Archive: URL
>
> Status in GNU Mailman:
> Fix Committed
>
> Bug description:
> In creating the List-Archive: header, CookHeaders.py strips a trailing
> slash from the URL returned by mlist.GetBaseArchiveURL(). It has
> always (in 2.1) done this, but it is wrong.
>
> Consider a list's archive TOC page which has entries like the
> following:
>
> <tr>
> <td>March 2012:</td>
> <td>
> <A href="2012-March/thread.html">[ Thread ]</a>
> <A href="2012-March/subject.html">[ Subject ]</a>
> <A href="2012-March/author.html">[ Author ]</a>
> <A href="2012-March/date.html">[ Date ]</a>
> </td>
> <td><A href="2012-March.txt">[ Text 766 KB ]</a></td>
> </tr>
>
> If the base archive URL is http://www.example.com/private/listname/
> and one goes to that URL, the relative hrefs all work fine, but if one
> goes to http://www.example.com/private/listname instead, those
> relative hrefs become relative to private/ instead of
> private/listname/, and one gets 'no such list: 2012-march' errors (or
> 404 errors in the case of public 'pipermail' URLs).
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mailman/+bug/964190/+subscriptions
>