Dan Harkless [Tue, 6 Feb 2001 03:46:27 +0000 (03:46 +0000)]
* -L isn't sufficient for specifying the path of the Cyrus SASL
shared library. That'll allow us to link successfully, but on
many/most OSes that won't allow us to find libsasl at runtime. On
Solaris, we need to specify the library path with -R as well (or
else the user will have to use the $LD_LIBRARY_PATH kludge, which
is considered harmful). This fix should be extended to other OSes
as well.
* Print whether we have SASL support in the "nmh configuration"
summary configure prints out.
Shantonu Sen [Sun, 24 Dec 2000 10:04:58 +0000 (10:04 +0000)]
Fixed up the hesiod library test so that the location of res_send
actually helps to determine whether -lresolv needs to be passed
as a hesiod lib. Also, fixed the krb5 test to look for
-lk5crypto if it exists, since that's the new name for -lcrypto.
The old -lcrypto remains in case the new version isn't found.
Dan Harkless [Wed, 20 Dec 2000 03:30:14 +0000 (03:30 +0000)]
Just changed a "can" to a "could" since I don't know if most POP3 servers are
like qpopper in not allowing people with APOP activated to log in via normal
POP3.
Shantonu Sen [Thu, 14 Dec 2000 01:30:43 +0000 (01:30 +0000)]
Resolve the circular dependency of libmh on libmts on libmh. The
files mts/generic/client.c and mts.c are moved to sbr/, and mts/generic/mts.h
is moved to h/mts.h. Thus, libmh is self-contained. All header includes
have been appropriately updated, and the Makefiles and configure script
no longer build mts/generic.
Dan Harkless [Tue, 12 Dec 2000 06:19:25 +0000 (06:19 +0000)]
When Shantonu made the new libmts.a, he swapped $(MTSLIB) and libmh.a in
sbr/Makefile.in so that libmh.a comes first, but this causes the build to fail
on Solaris, because libmts.a has to get ruserpass() out of libmh.a. Swapping
them back to the way Ken Hornstein's patch (which I applied on Jul 20) put them,
with libmh.a correctly coming second. If there are times when libmts.a needs to
come second, then it would appear there's a circular dependency and someone
(Shantonu?) did an mts merge incorrectly.
Shantonu Sen [Fri, 8 Sep 2000 00:30:58 +0000 (00:30 +0000)]
Moved code from zotnet/mts to mts/generic. Also, unified generic
mts code and smtp/sendmail into libmts.a as an intermediate, instead
of having two static libs.
Shantonu Sen [Wed, 6 Sep 2000 22:40:03 +0000 (22:40 +0000)]
Took out parsing of military time, which was causing issues when it
encountered something like (MET), which it doesn't recognize as a timezone,
and which was then read as military characters.
Dan Harkless [Thu, 10 Aug 2000 21:32:33 +0000 (21:32 +0000)]
Changed my mind again on scan.MMDDYY and scan.YYYYMMDD. Put the replied /
encrypted column back in, because it's useful to "always" (at least when you
aren't using replied / encrypted) have a space in between the message number and
date, so you can use awk (with the default field separators) to grab message
numbers from scan output.
Dan Harkless [Thu, 10 Aug 2000 20:33:25 +0000 (20:33 +0000)]
Decided that limiting the message number columns to 3 on my scan.MMDDYY and
scan.YYYYMMDD (to try to regain space taken by extra date info) was
ill-conceived. It's not that tough to get past 999 messages, though I imagine
it's rather rare to exceed 9999. Changed these to 4. I had already removed the
"replied / encrypted" column in YYYYMMDD, but this time I removed it on MMDDYY
as well (I've never seen it used...).
Dan Harkless [Sat, 5 Aug 2000 16:00:49 +0000 (16:00 +0000)]
Kimmo's etc/Makefile.in change note implied that install was previously
completely broken. Not the case. I asked Kimmo about it and he clarified that
it failed when building outside the source tree. Fixed the entry to reflect.
Dan Harkless [Mon, 24 Jul 2000 23:03:38 +0000 (23:03 +0000)]
When Shantonu wrote the new, more portable dtimep.lex, he left out the #ifdef
DSTXXX stuff for some reason. Not a good idea, as that code is required for
proper printing of numeric-offset timezones that have daylight saving time.
Without that code, -0700 during DST gets printed as MST instead of PDT.
Dan Harkless [Fri, 21 Jul 2000 02:35:43 +0000 (02:35 +0000)]
Applied Kurt J. Lidl <lidl@eng.us.uu.net>'s $MAILHOST patch:
I have a small patch that would be nice to be included -- basically, it
allows the usage of the "MAILHOST" environment variable, without having to
have HESIOD turned on. I need this functionality for my environment, where
we have identical /usr/local on all my machines (so I cannot just hardcode
into the mts.conf file), and I have multiple POP mail servers for my users.
Modified inc.man to reflect that along with "pophost:" and -host, $MAILHOST can
now activate POP mail inclusion as well.
Dan Harkless [Fri, 21 Jul 2000 02:25:59 +0000 (02:25 +0000)]
Last pass at README.developers -- Kimmo's 5-step commit was overkill. You only
need 3 steps, since configure.in is the only autoconf file with the RCS $Id
keyword.
Dan Harkless [Thu, 20 Jul 2000 20:49:00 +0000 (20:49 +0000)]
Ken Hornstein's SASL patch was not integrated properly with Ruud's new merged
mts/sendmail code. Kimmo has since fixed nmh so it compiles, but according to
Ken, the SASL stuff still does not work. Integrating a patch from him for this.
Dan Harkless [Thu, 20 Jul 2000 20:32:31 +0000 (20:32 +0000)]
One more pass at README.developers now that it's clear that my
previously-suggested one-line autoconf-file commit can cause unnecessary local
makes and an out-of-sync stamp-h.in file, but would not cause problems for other
people using the CVS files.
Dan Harkless [Thu, 20 Jul 2000 20:14:39 +0000 (20:14 +0000)]
Moved Kimmo's new "--with-hash-backup" to be output with the rest of the --with
options in the configure --help output, rather than being in between two random
--enable options.
Dan Harkless [Tue, 18 Jul 2000 02:15:03 +0000 (02:15 +0000)]
Clarified and made some corrections to Kimmo's README.developers changes (BTW,
if anyone can explain why the RCS Ids are able to cause problems with the
dependencies, please fill in the explanation -- I never encountered a problem
with the old single-commit method).
Dan Harkless [Tue, 18 Jul 2000 02:07:51 +0000 (02:07 +0000)]
I don't understand why Kim's split-up of the autoconf-file cvs commits should be
necessary, so I left a little note hoping for someone to explain technically why
that should be the case. I certainly never ran into any problems with my old
single-commit version (which I added back in along with an explanation of why it
supposedly won't always work).
Also Kim's text implied that the _only_ reason for the given order is the RCS
Ids -- not true -- the timestamps are the main reason a particular order is
necessary. Made a couple of other clarifications as well.