Unicode chars in list archive

[Jan Willem Stumpel]

This subject sometimes comes up, but not nearly frequently enough
I think.

The web archives of this list at
http://mail.nl.linux.org/linux-utf8/ do not preserve utf-8
characters intact. This is absurd given the subject of this list.

Is there no way of remedying this situation? Could this list be
hosted somewhere else?

Regards, Jan

[Egmont Koblinger]

Post by Jan Willem Stumpel
This subject sometimes comes up, but not nearly frequently enough
I think.
The web archives of this list at
http://mail.nl.linux.org/linux-utf8/ do not preserve utf-8
characters intact. This is absurd given the subject of this list.

Oh yes... (sigh)

Short summary of one particular thread trying to get this bug fixed:

Dec 2004: bug caught by me in Mhonarc 2.6.10, reported to author, and fixed
by the author in CVS nearly immediately.

Mar 2005: I asked the maintainers of nl.linux.org to apply this fix. No
reply for two weeks.

Mar 2005, two weeks later: sent this mail:
http://mail.nl.linux.org/linux-utf8/2005-03/msg00032.html
to this list. No worthful answer arrived to the list, but at least Bruno
Haible (co-founder of the list with Markus) forwarded this mail (cc'ing me)
to Rik van Riel (site admin) with a preface asking if he could fix this. Rik
replied:
"I will fix this once there's a new version of mhonarc
that has the bugfix. Playing around with a CVS version
might not be the best idea on such a large mail archive..."

Apr 2005: Bruno wrote me:
"Markus and I are the initiators of the linux-utf8 list. I'm glad that Rik
does the hosting and admin service for us, for five years. I don't think
that moving to a new list address (a small trouble for everyone) is worth
the trouble. Also, you know, anyone can create a list archive on his own.
For example, gmane.org carries archives of linux-utf8, but many messages
seem to be missing. The easiest way to fix the problem is therefore to
create an alternative archive of the list somewhere else."

May 2005: mhonarc 2.6.11 is released, including the fix. I sent this mail to
Rik and Bruno:
"Back to this topic we discussed two months ago:
There's a new official mhonarc release available which fixes the UTF-8 bugs.
Please upgrade to it and re-generate the mail archives so that the accented
letters show up correctly on the web archive."

No reply since then.


To summarize: I have a feeling that those people who could do anything for
this bug to get fixed do not care at all. :-( It's a two-byte fix that
should be applied to mhonarc, or it should be upgraded to a newer version
that fixes lots of other bugs, and the archives need to be re-generated,
which probably takes 2 minutes from the sysadmin's time (to find out the
proper command). Definitely much less time that I spent so far catching this
bug in mhonarc and typing all these e-mails...

Jan, I perfectly agree with you: "This is absurd."

[Rik van Riel]

Post by Jan Willem Stumpel
This subject sometimes comes up, but not nearly frequently enough
I think.
The web archives of this list at
http://mail.nl.linux.org/linux-utf8/ do not preserve utf-8
characters intact. This is absurd given the subject of this list.
Is there no way of remedying this situation?

Egmont was right. It was only a few minutes of sysadmin time
to apply the patch and figure out a command to regenerate the
archives. I really should have done this earlier, my apologies.

Turns out it will take a while to actually regenerate them
though. Just for linux-utf8 the script has been running for
20 minutes now and it's only just made it to 2004 :)

I suspect I might not regenerate the archives for the other
100-odd mailing lists unless there's a specific request...

[Jan Willem Stumpel]

Post by Rik van Riel
Egmont was right. It was only a few minutes of sysadmin time
to apply the patch and figure out a command to regenerate the
archives.

Fantastic, thanks