The trac mailing list could have more useful content

Create issue
Issue #2523 resolved
Gabriele Bozzola created an issue

I am subscribed to the trac mailing list to follow reported bugs. When I get an email, the only useful information I can extract is the title of the ticket. For example,

http://lists.einsteintoolkit.org/pipermail/trac/2021-April/015293.html

If I click on the link, I end up in a non-render HTML page which is very difficult to parse visually. So, if I want to understand what is going on, I have to visit this website, find the relevant ticket, and find the new post. I think that the mailing list would be much more useful if it actually included the new comments, or if it provided a direct link to the post.

Comments (14)

  1. Roland Haas

    The mailing list emails are HTML (only) emails, which so far seem to have been ok with the usual email clients (which would prefer the HTML part even if there was a plaintext alternative). Which email client are you using?

    The archive was supposed to render HTML parts into plain text as far as I can tell from the mailing list settings. I will try and see if I can find out what is going wrong or (more likely) I may be misunderstanding the mailman options.

  2. Gabriele Bozzola reporter

    I can reproduce this problem with GMail, but maybe I should clarify that I get these emails with the option digest (looking at pipermail I assumed that this detail was not important). This is what I see as subject of the email:

    If I click on the link, I end up in a HTML page that is not rendered.

  3. Roland Haas

    I suspect that the digest is to blame.

    The HTML pages in the link are set up by mailman to show the raw HTML so that no one can send a HTML attachment the mailing list with say malicious javascript code which then gets executed when the HTML page is shown in your browser. While one seems to be able to disable this, this seems to be a mailman wide setting and that seems to be risky.

  4. Roland Haas

    Well its in the HTML content… which is of no use to use since you cannot see the HTML in the digest. Currently the HTML looks like this when properly rendered: https://ekohaes8.ncsa.illinois.edu/~rhaas/2523.html

    The “proper” solution is to generate a “multipart/alternative” email which has both plaintext (for archiving and digest) and HTML (everyone else) content. Due to the way Bitbucket works (and not wanting to enter into the whole mess of creating content preserving plaintext from HTML) the plaintext will be the markdown source which may or may not look nice.

  5. Roland Haas

    Old messages will not have their archived content updated, similarly the digest may take a day or so to show anything. The first email that will look ok is likely going to be the one where you say “I still see the same messages”.

  6. Gabriele Bozzola reporter

    I think that there’s still some minor issue. The email I got today had the following:

    As you can see, the message is not the correct one (but the URL is).

  7. Roland Haas

    I think the email is correct. I accidentally added that comment to the wrong ticket and then deleted the comment. So that comment would not show up in the web-view of the ticket right now but will show up in emails (which are generated the moment one modifies a ticket). We may have to add a “removed comment” case, though I am not sure if Bitbucket lets on hook into that event.

  8. Gabriele Bozzola reporter

    I see! I guess that the current behavior is acceptable, when the comments seems out of place, it is always possible to follow the link (which is what I did) and see exactly what is going on.

  9. Roland Haas
  10. Log in to comment