From: newtophp2000 on
Robert Hicks wrote:
> I am running the same script on 3 different HP servers. One of the
> servers is erroring out with the following:
>
> rhicks(a)lims-dev1$ ./fstats.tcl
> key message-id not in header
> while executing


I faced this very problem recently. No one responded to my post at
first, so after some trial and error, I stumbled upon the solution.
While I did it manually at the time, Jeff Hobbs suggested the same
thing and how it could be done easily via the TDK.

The trick is one of the packages that smtp relies on is not on the
server where you see this error. You can check this manually, or if
you have TDK, use the Scan feature of tclapp. It will resolve all
dependencies and present you with all the packages that you need to
have.

I don't have the source code with me now to give you the list of the
required packages but check to see if you have Trf, mime, smtp, and md5
(and maybe a few others) on the server where you see the problem. If
you can't resolve it, email me and I will send you the list.

From: Don Porter on
newtophp2000(a)yahoo.com wrote:
> The trick is one of the packages that smtp relies on is not on the
> server where you see this error.
....
> I don't have the source code with me now to give you the list of the
> required packages but check to see if you have Trf, mime, smtp, and md5
> (and maybe a few others) on the server where you see the problem. If
> you can't resolve it, email me and I will send you the list.

Since you've done the work tracking the problem down, can you send
error reports to those packages that use commands from other packages
they do not [package require] ?

If each bit of code [package require]s all the packages that provide
any of the commands it uses, then users will get error messages like
"can't find package foo" instead of cryptic things like reported in
this thread.

--
| Don Porter Mathematical and Computational Sciences Division |
| donald.porter(a)nist.gov Information Technology Laboratory |
| http://math.nist.gov/~DPorter/ NIST |
|______________________________________________________________________|
From: newtophp2000 on
Don Porter wrote:
> Since you've done the work tracking the problem down, can you send
> error reports to those packages that use commands from other packages
> they do not [package require] ?
>
> If each bit of code [package require]s all the packages that provide
> any of the commands it uses, then users will get error messages like
> "can't find package foo" instead of cryptic things like reported in
> this thread.


Sure, I would be glad to this weekend. I should make the disclaimer
that what I did was not very scientific. I was simply scanning all the
source code in tcllib to see who was doing package require/provide. As
I recall, what made it complex was that there was conditional package
require's. So, that may account for silent error messages that slip
by.

Here is the original thread in case anyone is interested:
http://groups.google.com/group/comp.lang.tcl/browse_frm/thread/237d7ff67aa4f223/73d5a442066714e5?q=smtp+starpacks&rnum=1#73d5a442066714e5

From: Pat Thoyts on
mghembru(a)harshrealm.uwaterloo.ca (MH) writes:

>In article <1145582043.927848.312420(a)v46g2000cwv.googlegroups.com>,
>Robert Hicks <sigzero(a)gmail.com> wrote:
>>Anything is possible. What is wierd is that the script worked once in
>>production (and numerous times in testing) and then stopped.
>>
>>How would I set a message-id field??? I have no clue about smtp except
>>to write the script.
>>
>>Robert
>
>Well, if you're not using authentication, you could try something like:
>telnet smtp-server 25
>mail from: robert(a)abc.com
>rcpt to: robert(a)def.com
>data
>This is a test email

A simpler option is to add -debug 1 to the smtp::sendmessage
options. This will cause it to output to stderr all the SMTP traffic.

--
Pat Thoyts http://www.patthoyts.tk/
To reply, rot13 the return address or read the X-Address header.
PGP fingerprint 2C 6E 98 07 2C 59 C8 97 10 CE 11 E6 04 E0 B9 DD
From: Robert Hicks on
I am using ActiveTcl so I am pretty sure I have all the required stuff
AND it did work when I tested and once when I set it up in cron (all on
the same machine). "Somehow" it just stopped. I am using the same
ActiveTcl build across 3 servers and only this one is having the issue.

I will investigate some more though.

Robert