PET->ANSI

Here some pictures what it look like
if i try login  my Amiga it's not even that clear...

Im dealing with this on an Amiga BBS converting doors from one bbs package to another. It comes down to these display screens need to be cleaned through the ANSI (or native) editor to clean them up.  I have found for my issues just like this but on Amiga BBS it requires multiple editors to clean them up.  with that said not sure what if anything will happen on the image screens being the focus is on the 95% of callers are on Commodore CG mode not ansi. But I get where you are coming from, im there with you.  

I would clean them up for ansi but just overloaded with other projects to make that happen.

If you clean them up you can submit them to me via PRIVATE MESSAGE and we will incorperate them in the the next Image realease.

on a side note im still looking through my OG disks from 1992-1993 when I have ANSI screens for Image 1.2a that with one simple ansi code would lock the pc calling in.  We used this back then for troubling wanna be hackers.

S.

Posted by: @codetsu

↑

let's tackle the following problem
For some reason, when logging in with an ANSI compatible Terminal, the problem only occurs if you select the ANSI color setting
usually this is most visible in the IM menu and in the list of last logged in users
menus are more obviously filled with Telnet code or ANSI commands, often several lines are missing from the printout.
this is only if you select ANSI colors as active
(standard in the menus that came with IM3)
it seems like some error occurred in the Pet -> ANSI translation, any ideas what could be the cause?

I have seen this before, a couple of things to look at. I can login with 2 PC Programs Syncterm and Netrunner with no issues. I can login on an Amiga Emulation using Terminus and get what you are getting. I am far from a fan of the ANSI emulation on Image, back in the day it was slow as anything at 2400bps. Today it is tolerable at 38.4k. All I can say is try to use a term program that can handle IBM graphics better. Like I said both that I have used work fine, there are some Petscii characters that do not translate and you will get some off looking characters, but unfortunately there really isn't much I can tell you to do..

Sorry..

If I understand correctly Image doesn't actually accept graphics made to the ANSI standard,

but it converts PETscii character graphics into something that looks like ANSI.
the fact that I would use something better for an ANSI-compatible terminal is a bit strange

for two reasons

1) the errors are not consistent so that I could locate the exact code causing the error

2) I think SyncTerm follows the ANSI standard really well, and secondly, no other BBS gives me a similar error on any terminal or platform I use. Even BBS use IM3.0

The strangest thing here is that the PETscii side works perfectly. Similarly removing colors from IBM graphics removes the error I'm not saying that the IM3.0 program is the problem, but I'm interested in why this happens, is there a bug that only appears on my configuration? and is there a problem with the Wifi modem I'm using or with the M3.0 serial or PETscii translation routines?

One thing I strongly suspect is that my modem doesn't always stay in Binary transfer mode like it should or it's confusing some hex combination with the device's own or Telnet commands.

To be completely sure I have to ask again whether IM .seq menu files are converted to ANSI format on the fly or does it print them directly to serial?
I think I have tried putting ANSI in the .seq file and that experiment did not work, of course it has been a while so I can't remember exactly.

I personally don't like messing with ANSI graphics with PETsci BBS, but non-C= users should be offered at least some kind of working experience if such a possibility is available from the BBS program side
if no solution is found, the only option for me seems to be to remove the possibility of using ANSI color coding altogether

Btw. using StrikeTerm program with ANSI terminal, similar errors do not appear in the menu?!?

Posted by: @codetsu

↑

If I understand correctly Image doesn't actually accept graphics made to the ANSI standard,

but it converts PETscii character graphics into something that looks like ANSI.
the fact that I would use something better for an ANSI-compatible terminal is a bit strange

for two reasons

1) the errors are not consistent so that I could locate the exact code causing the error

2) I think SyncTerm follows the ANSI standard really well, and secondly, no other BBS gives me a similar error on any terminal or platform I use. Even BBS use IM3.0

The strangest thing here is that the PETscii side works perfectly. Similarly removing colors from IBM graphics removes the error I'm not saying that the IM3.0 program is the problem, but I'm interested in why this happens, is there a bug that only appears on my configuration? and is there a problem with the Wifi modem I'm using or with the M3.0 serial or PETscii translation routines?

One thing I strongly suspect is that my modem doesn't always stay in Binary transfer mode like it should or it's confusing some hex combination with the device's own or Telnet commands.

To be completely sure I have to ask again whether IM .seq menu files are converted to ANSI format on the fly or does it print them directly to serial?
I think I have tried putting ANSI in the .seq file and that experiment did not work, of course it has been a while so I can't remember exactly.

I personally don't like messing with ANSI graphics with PETsci BBS, but non-C= users should be offered at least some kind of working experience if such a possibility is available from the BBS program side
if no solution is found, the only option for me seems to be to remove the possibility of using ANSI color coding altogether

Btw. using StrikeTerm program with ANSI terminal, similar errors do not appear in the menu?!?

Unfortunately, the PETSCII to ANSI conversion is done in the ML. Something you can try create files with the suffix 2, ASCII is 0, Petscii is 1 and ANSI is 2, try creating screens with an ANSI creator name them with the 2 s.login 80 2 s.start 80 2 (most ANSI terms are 80 columns). See how that goes, it might work out, I haven't did a lot of testing creating ANSI screens for Image. Back in the day I did 0 testing of it.

You might be right about the modem, when an ANSI user logs on to Image 3.0 running TCPSER most of the time TCPSER will spit out NVT I believe it is, not sure what it stands for. You can look at that..

Image sends the file directly to the modem, there is no conversion then spitting it out, In ANSI mode it reads the petscii file and sends it to the modem converting it in the ML to the modem.

Completely agree about at least offering something. Thus the reason it sends the petscii files as ANSI, what term program are you using that is giving you these issues?? I just logged in fully from beginning to end with Netrunner on my BBS and had no breakdowns like you showed in the pictures.  There were a few characters that did not translate but that was it. I only use Striketerm when I need to do file transfers, I stick with Syncterm when calling Petscii boards.

I did a little test
and maybe I figured out what's wrong here...
I first loaded the s.login 2 80 file into the WF editor, it seemed to work fine with the M command
I also loaded the 0 80 file and that also seemed to work and the command codes were the same in both
then I loaded s.login 0, its commands were of course PETscii codes and they didn't work
all sorts of number codes appeared that didn't belong there
so my assumption is that if the BBS pushes PETscii code into the ANSI terminal, the graphics break down
this is not taken into account at all points, even if ANSI is selected at the beginning

Now the question is as follows if I make such a radical change in IM that I remove ANSI support completely, how do I do that?
(in turn, I have the WiFI modem do the translation from PET to ANSI at the hardware level)

Posted by: @codetsu

↑

I did a little test
and maybe I figured out what's wrong here...
I first loaded the s.login 2 80 file into the WF editor, it seemed to work fine with the M command
I also loaded the 0 80 file and that also seemed to work and the command codes were the same in both
then I loaded s.login 0, its commands were of course PETscii codes and they didn't work
all sorts of number codes appeared that didn't belong there
so my assumption is that if the BBS pushes PETscii code into the ANSI terminal, the graphics break down
this is not taken into account at all points, even if ANSI is selected at the beginning

Now the question is as follows if I make such a radical change in IM that I remove ANSI support completely, how do I do that?
(in turn, I have the WiFI modem do the translation from PET to ANSI at the hardware level)

Yes, the files with the 2 at the end are for ANSI screens, so you can create files with minimal petscii graphics if you wish and they should be fine.. I would try creating some ANSI screens and name them with the 2 at the end to see if they translate fine. I never really went crazy for ANSI users on my systems, so I have never tried it..