Hi

Today I pushed a fix for Talisman squish message bases to work better
with other squish software.

I did not realize that control lines are seperated by carraige returns,
and so have added that.

This means if you update to this latest commit, you will NEED to
recompile postie, qwkie, toolbelt and falcon in addition to talisman.

I've tested postie, and it seems to work, I have not tested qwkie or
falcon as I don't use them presently (but they should work)

Andrew


--- Talisman v0.49-dev (FreeBSD/amd64)
* Origin: Smuggler's Cove - scove.talismanbbs.com:2323 (21:1/182)

apam wrote (2023-12-10):

Hi

Today I pushed a fix for Talisman squish message bases to work better
with other squish software.

I did not realize that control lines are seperated by carraige returns,
and so have added that.

IIRC control (kludge) lines are part of the message body / text and should be stored unmodified. Is the fix only related to control lines that are created by Talisman or did the bug also affect control lines from imported messages? I'm just curious.

⁂

---
* Origin: No REPLY kludge - no reply (21:3/102)

IIRC control (kludge) lines are part of the message body / text and
should be stored unmodified. Is the fix only related to control lines
that are created by Talisman or did the bug also affect control lines
from imported messages? I'm just curious.

Squish stores most kludges seperately from the message text (all but
seenby and path).

The fix only changes the way talisman stores them in the squish base.
Prior to the fix messages exported from talisman would have carraige
returns added, and imported messages would have carriage returns removed
for storage.

The issue occured when using a particular third party on bases created by talisman, it was expecting a carriage return to signify the end of the
kludge line and did stupid things when it didn't get one.

Although now that you mention it, I suppose it could be seen as all in
the message body / text as the the control body is right before the
message body in the squish structure - and that makes sense, hmm.

looks like it's also not supposed to have a carriage return after the
last kludge :(

Andrew


--- Talisman v0.49-dev (FreeBSD/amd64)
* Origin: Smuggler's Cove - scove.talismanbbs.com:2323 (21:1/182)

apam wrote (2023-12-10):

IIRC control (kludge) lines are part of the message body / text and
should be stored unmodified. Is the fix only related to control lines
that are created by Talisman or did the bug also affect control lines
from imported messages? I'm just curious.

Squish stores most kludges seperately from the message text (all but
seenby and path).

The fix only changes the way talisman stores them in the squish base.
Prior to the fix messages exported from talisman would have carraige returns added, and imported messages would have carriage returns removed for storage.

Looks like I didn't remember it correctly. According to the "Squish Developers Kit (Version 2.00)" documentation there is no CR after a control item. They are separated by SOH (0x01) only. Example from the spec:

<SOH>CHARSET: LATIN1<SOH>REALNAME: Mark Twain<NUL>

It seems the previous behavior was correct. I also cannot find any CR between control items in sqd files that were created by Squish (the tosser) on Linux.

The issue occured when using a particular third party on bases created by talisman, it was expecting a carriage return to signify the end of the kludge line and did stupid things when it didn't get one.

Which software is that?

Although now that you mention it, I suppose it could be seen as all in
the message body / text as the the control body is right before the
message body in the squish structure - and that makes sense, hmm.

looks like it's also not supposed to have a carriage return after the
last kludge :(

⁂

---
* Origin: No REPLY kludge - no reply (21:3/102)

Looks like I didn't remember it correctly. According to the "Squish Developers Kit (Version 2.00)" documentation there is no CR after a
control item. They are separated by SOH (0x01) only. Example from the
spec:

Yeah, that's what I thought - I was following a different spec and
misread it when I "fixed" it.

It's unfixed now, so should be correct again.

Andrew


--- Talisman v0.50-dev (FreeBSD/amd64)
* Origin: Smuggler's Cove - scove.talismanbbs.com:2323 (21:1/182)