I added some code to elogd which always places <> around the "MAIL FROM:" in the SMTP communication. So I hope that problem should be solved from now on. Thank you for your detailed analysis and your hints. 

I'm not sure if this works, since your test

i < (int) sizeof(cookie)

still will stop parsing cookies if there is one which is too long. So I added your test plus changed the parsing to:

for (i = 0; *p && *p != ';' && *p != '\r' && *p != '\n' ; )

   if (i < (int) sizeof(cookie)-1)

      cookie[i++] = *p++;

   cookie[i] = 0;

The modification is in the curren SVN revision (# 2162). So have a look and check that it works.

The Windows Binaries are not recompiled for each minor modification. You should check the SVN revistion (the number which shows up at the very bottom of the page, which is 2153 in this case). It might well be that the Windows Binaries do not yet contain the modification for extended SMTP error report.

Furthermore, TLS is not supported in elogd, only plain SMTP is possible. 

Importing dates from CSV files was not implemented for a good reason: There are dozens of date formats (such as "1/22/09", "Jan 1 2009", "Mon 1 January '09") and it is almost impossible to handle all formats. I changed now the code to catch at least some formats, such as the internal RFC2822 compliant date which is also used internall by elog. An example of this format is "Thu, 22 Jan 2009 09:26:00". This date format will also be used if you export a logbook via a CSV file. Another format I added is "1/22/09". So give a try to the new code (I made an intermediate release for windows V2.7.5-3) and see if it works for you.

You need something along these lines:

Attributes = ..., Confirmed
Options Confirmed = boolean
Preset Confirmed = 1 

No, this is not possible. 

> For authentication via the Windows IIS I have "basic authentication" with no TLS encryption required - I'm 
> not sure if that meets the provider's "SMTP authentication" requirement.  When I was using the previous version 
> of ELOG (sorry, I did not note the version number) the username was encrypted along with the password.  With 
> the 2168 build, the password is still encrypted but the usernamer is not; should the password be in "plain 
> text" as well?  

I changed only the screen output. The username is printed not not encrypted, to be able to check it by eye. The 
username sent over the network is and was always encrypted.

The only clue I have is to compare the network traffic you get with an email client (outlook, thunderbird ??) 
with the network traffic you get from elogd. To monitor the traffic, I use normally WireShark (formerly Ethereal, 
www.wireshark.org). You have to filter for "tcp.port == 25" which is SMTP, then capture your traffic, right-
click on it and select "Follow TCP stream". The result you see in the attachment. The strings produced by elogd 
are pink, the strings from the SMTP server are blue. In my case I don't have authentication (our SMTP server does 
not support this). In you case, you have to compare the traffic from a functioning email client with elogd.

> Hello,
> 
> We are now running elog-2.7.5 in a Scientific Linux 4 (RHEL4) system, which uses ImageMagick 6.0.7.1.  After uploading an image, the image 
> manipulation buttons don't work and complain:
> convert: unrecognized option '-set'
> 
> We are using an RPM built from source, and it looks like I should be able to just change "-set comment ..." to "-comment" as in:
> ------
> [dab66@lnx100 tmp]% diff elogd.c elogd.c.new 
> 22601c22601
> <       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -set comment ' %d' '%s'", file_name, new_rot,
> ---
> >       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -comment ' %d' '%s'", file_name, new_rot,
> 22607c22607
> <       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -set comment ' %d' '%s'", file_name, new_rot,
> ---
> >       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -comment ' %d' '%s'", file_name, new_rot,
> 22618c22618
> <       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -set comment ' %d' '%s'", file_name, cur_rot,
> ---
> >       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -comment ' %d' '%s'", file_name, cur_rot,
> 22624c22624
> <       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -set comment ' %d' '%s'", file_name, cur_rot,
> ---
> >       sprintf(cmd, "convert '%s' -rotate %d -thumbnail %d -comment ' %d' '%s'", file_name, cur_rot,
> ------
> 
> Is there any better way for us to fix this, or is anything else needed?

Well, I believe your modification won't work. I just tried it with ImageMagick 6.3.8 and it failed. Try to rotate your picture four times, and 
it should be back to the old position. When I use "-comment" instead "-set comment" on a PNG file, this did not work. From the man page it 
tells me:

Version: ImageMagick 6.3.8 01/25/08 Q16 http://www.imagemagick.org
Copyright: Copyright (C) 1999-2008 ImageMagick Studio LLC

Usage: convert [options ...] file [ [options ...] file ...] [options ...] file

Image Settings:
  ...
  -comment string      annotate image with comment
  ...

Image Operators:
  ...
  -set property value  set an image property
  ...


So indeed "-comment" and "-set comment" are two different things. I'm afraid you have to upgrade your ImageMagick package in order to make 
this work.