On Wed, Apr 15, 2009 at 19:05, Jack Repenning <jrepenning_at_collab.net> wrote:
>...
> However, all that aside ... cataloging and assigning Tigris mail-list
> botches at the moment is considerably more "fun" than anyone really
> needs to have. I'm not sure how many, or precisely which, of the known
Stefan Sperling ran into a problem where the ML software munged his
patch. Basically, it *rewrote* the message into something that was
*incorrect*. It said "quoted-printable", but the content was wrong.
There was a URL with an ?something=16<more digits> in it which didn't
get quoted. Thus, MUAs interpreted it as: ...?something<0x16> <more
digits> ... which meant the URL terminated at the end of "something"
rather than including the digits. And (thus) the URL was totally
broken. He also said the patch contents, which were inline, was
monkeyed up incorrectly.
> problems have cropped up in the history of this thread, but it does
> seem to be a pretty impressive sampling. Hyrum is now reviewing, so he
> evidently got the text somehow, but "how" is not clear from the thread.
The "how" is to wait for the text file to come thru so that it is
visible. Or to just download the attachment and review it outside of
the MUA (which was the original pain point making the patch
"unreviewable").
> I think that the original attempt was on the right track, and would
> have worked except for the unfortunate file-name typo. Am I wrong
> about that?
Yah... that is probably the original problem with the mime type for
the attachment.
But there is still the double-quoting of the message body.
Oh. And I'm still waiting to see a solution to correct the "this
message has been converted to an attachment". I thought you said that
was solved by making the message delivery a passthru, and the
conversion would only happen for he web UI. What happened to that?
-g
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1730551
Received on 2009-04-15 19:43:56 CEST