Hi All,

Apologies if this is answered elsewhere, but a quick search didn't reveal anything directly relevant.

Having installed smstools, it works great, and I'm very happy with it. However, I have encountered an issue when trying to create outgoing SMS messages using either an in-house script or even simply with a text file.

If I create the message using the included 'sendsms' it works fine.
If I copy/modify one of the included examples, it works fine.

If I (programmatically, or using just a text file) create a new SMS file from scratch and place it in the outgoing spool, the server reports:

smsd: No destination in file /var/spool/sms/outgoing/<filename>

However, the format of the file is exactly as described in the documentation, and even creating (manually) an exact copy of one of the examples (and comparing with 'diff' to be sure it's exactly the same), it still fails with this error.

Example format of output file (excluding the '-----' lines)
-------------------------------------
To: 336XXXXXXXX
Flash: yes

This is a test message.
-------------------------------------

Am I missing something obvious?

Thanks in advance

Leland

Very strange...

Can you show what hexdump -C < filename says?

keke wrote

Very strange...

Can you show what hexdump -C < filename says?

(obfuscating the actual destination telephone number)

00000000 54 6f 3a 20 33 33 XX XX XX XX XX XX XX XX XX 0a |To: 336********.| 00000010 46 6c 61 73 68 3a 20 79 65 73 0a 0a 54 68 69 73 |Flash: yes..This| 00000020 20 69 73 20 61 20 74 65 73 74 0a | is a test.| 0000002b

Looks okay to me .. hmm..

There was no any other messages in the log?

Is you hexdump surely from the failed file? The dump is ok and should work.

When you create a file, what are the permissions?

And you are not using any checkhandler which modifies message headers?

You could mkdir /var/spool/sms/failed and define failed = /var/spool/sms/failed in the smsd.conf, then restart smsd and try once more. Show the hexdump of unaltered failed message, use private tag to protect it.

Think I found the problem...

The daemon was slurping from the spool directory before the file was completely released during creation...

Writing to a temporary file elsewhere first, and then moving the temp file into the spool seems to have resolved the issue.

Will do some more tests and let you know to be sure.

Thanks,

Leland