> Collin, (and John)(and Alex)
>
> > My guess is that it would be in the code that handles the approved
> > header in the body of the message, but I am not sure. I tried running
> > the test case by hand through the 1.92 resend with no luck. If anybody
> > else tracks it down, I would appreciate looking at the patch so I can
> > make sure that it is really fixed and not merely masked in the 1.93
> > resend.
>
> I *think* I patched a simillar problem I was having a couple of weeks
> ago. When an owner of a moderated list posted to the list with the
> Approved: command in the BODY of the message, header information (such
> as bulk, sender) was getting mixed in with the text of the message.
> This header info was being inserted after the first (or second, I dont
> remember) line of the message. And for some reason the subject line
> got blown away in the process also.
>
> This behaviour did not occur when the listowner placed the Approved:
> command in the header of the message.
>
> I came up with the following patch (for 1.92), its a hack, and hasn't
> been fully tested but it seems to be working well. Basically the
> header information looked like it was being rewritten when a valid
> Approved line was detected, however if this Approved: line is in the
> body of the message then there is no need to print the header info
> again.
>
> Comments?
>
[patch set out]
This sounds like the problem (and the area of code I was looking at) about
a year ago when we first brought up majordomo here--I gave up trying to put
the "Approved" stuff in the body for just the reasons given above.
I was (and am) too poorly trained to attempt very much perl code--and Brent
was up to his ears it appeared, and I had to get something going--we finally
got the approve program and process to run, but it seems like I had to
make some mods to the code that handles "Approved" anywhere run, when
I got the header version running I lost interest in the task.
But I remember thinking that the "Approved-in-the-body" code looked like
a kludge that probably worked at most at one site with some really site-
local conditions. And as I recall, I decided the reason I was too frightened
to proceed was that the code--having found "Approved:" somewhere, assummed
that it had found it in the headers.
I'm not sure if I have added anything here, but for what ever it is worth,
the description above sounds really familiar.
Oh, and it seems like I determined that the problem is exacerbated if you
"forward" a piece to the list without deleting the MA-inserted blank line
beore the text--the code (it seems like) keyed on the blank line as "eoh",
even if it had already _found_ the end of the headers.
--
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
. .
- L. F. (Larry) Sheldon, Jr. -
. Unix Systems Administration .
- Creighton University Computer Center-Old Gym Most of us can do more -
. 2500 California Plaza than we think we can, .
- Omaha, Nebraska, U.S.A. 68178 but usually less than -
. lsheldon@creighton.edu we think we do. .
- 402 280-2254 (work) -
. 402 977-2946 (pager) "Bits & Pieces" .
- 402 332-4622 (residence) -
. .
.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.
References:
|
|
