According to Geoff Hutchison:
> I couldn't really care less. Berolist is certainly easy to set up, but
> there's a lot of good list software out there. As long as we don't change
> things significantly as far as setup, I don't mind. Adding subscribe and
> unsubscribe addresses might help though. (You know, [EMAIL PROTECTED],
> [EMAIL PROTECTED])
I'm torn on this. On the one hand, it would be nice to have something
more solid. On the other hand, BeroList mostly works OK, and it's all
set up now. Finding, evaluating, configuring, testing and cutting over
to another package could be a big job, and it's not one I'd like to take
on at this point. With that in mind, I decided to dive into the BeroList
code to see how hard it would be to fix. I've got two untested patches
so far, and some questions about how to handle the third problem.
> >Personally, after looking at the BeroList source, I'm not really impressed
> >with the quality of it. However, it does seem that my beefs are fixable.
>
> Ditto. I haven't received any sort of message from the author or the
> mailing list about some of the problems.
I don't know how many users he deals with, or how much feedback he gets,
but we can certainly send him our fixes, if we go ahead with them.
I've fixed the handling of multi-line headers. His code used to skip
over them properly, but then truncated the saved header after the first
newline. I now save the header after it's made into one line, and I
allow spaces as well as tabs for continuation lines. My second patch
allows an unsubscribe command as the first line of the message body.
> >the easy fix would be to set sender=list in the config files for the
> >mailing lists, but then replies would go to the list. It would take a
> >bit of hacking to allow a separate envelope FROM address, but it's feasible.
>
> This is why I didn't set that option. Then again, I figured I wouldn't want
> to... :-)
Another way might be to set sender=list, and replyto=sender, so that
even though messages will now be "From: [EMAIL PROTECTED]", the Reply-To
header will show the original sender. We'd still get the silly "* From"
lines inserted, though.
The other option is to use listaddress in the MAIL FROM line, instead of
the sender address. That would make errors go back to the list server.
It already sets Errors-To to listaddress, but most mailers ignore that
header (it's obsolete). I'm leaning towards this option. What do you
think? Should it be config file selectable?
> >sticking with BeroList, I'll try to hack it in the next day or two. If
> >instead the plan is to upgrade to a better quality package soon, I'll save
> >myself the bother.
>
> As I said, I couldn't care less. I don't think Andrew cares either, he
> suggested it because it seemed easy to set up and maintain. As I haven't
> looked at UNIX list software in a while, I'm open to suggestions. Mailman?
I'm open to suggestions too, but I have no existing list server
experience, so if we're going to switch, I think we'd need another
volunteer to take on the job. I think it'd be worthwhile, but in
the meantime, here are my fixes so far to BeroList, included below.
How should we go about testing this? I think it might be a good idea
to run the test version only on htdig3-dev, until we get the bugs out,
then switch the other lists over to it.
--- list.c.multilinebug Sat Jan 23 19:10:17 1999
+++ list.c Wed Apr 7 11:10:56 1999
@@ -884,13 +884,14 @@ int main(int argc, char *argv[])
additional_header=salloc(slen(header)+2);
header_line=salloc(slen(header)+2);
while(strchr(header_copy,'\n')!=NULL) {
- strcpy(header_line,header_copy);
/* Handle multi-line headers (Received, etc.) */
- while(*(strchr(header_copy,'\n')+1)==9) { /* Multi Line */
- strcpy(strchr(header_copy,'\n'),strchr(header_copy,'\n')+2);
+ while((a = strchr(header_copy,'\n'))!=NULL &&
+ (*(a+1)=='\t' || *(a+1)==' ')) { /* Multi Line */
+ strcpy(a,noleadingspaces(a+2));
}
+ strcpy(header_line,header_copy);
header_copy=strchr(header_copy,'\n')+1;
*(strchr(header_line,'\n')+1)=0;
while(strchr(header_line,'\t')!=NULL)
--- list.c.nounsubmsg Wed Apr 7 11:09:00 1999
+++ list.c Wed Apr 7 17:08:49 1999
@@ -1024,6 +1024,11 @@ int main(int argc, char *argv[])
} else if(scasecmp(noleadingspaces(subject),"unsubscribe")==0) {
unsubscribe(to,from,NULL);
control=1;
+ } else if(strncasecmp(a=noleadingspaces(msg),"unsubscribe",11)==0 &&
+ (*a == '\r' || *a == '\n')) {
+ /* we'll allow "unsubscribe" alone on first line of message
+body. */
+ unsubscribe(to,from,NULL);
+ control=1;
} else if(control==1) {
/* Mode is -control, but command couldn't be processed... */
free(mail);
--
Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]>
Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil
Dept. Physiology, U. of Manitoba Phone: (204)789-3766
Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED] containing the single word "unsubscribe" in
the SUBJECT of the message.
