After Hours BBS Program

[+arcadeshopper]

fctn-w also does not work for me..

 

the other fctn keys work fine.. fctn-= will log off a user/force hangup  fctn-z will initiate chat 

 

you can never telnet to the tipi from external machines, it's not allowed except for localhost(port 23) only ssh works from external.. ssh works for me when the bbs is up

 

signup works fine for me. perhaps you have a bad file? 

 

Greg

[wolhess]

Hi Greg,

 

yes, I think too I have some faulty files. Now I can use my own user #2 very well. 
e.g the caller list is not correct at the end.
 

your hint with ssh is helpful. I used right now a ssh client from my iPad and it works fine and fast.

 

can you please test to create a new user on my BBS?

 

the hostname is: mytbox99.ddns.net

the port is: 61643

 

I used this morning only the Telnet client from my Windows PC. Maybe this was my problem.

[+arcadeshopper]

1 hour ago, wolhess said:

Hi Greg,

 

yes, I think too I have some faulty files. Now I can use my own user #2 very well. 
e.g the caller list is not correct at the end.
 

your hint with ssh is helpful. I used right now a ssh client from my iPad and it works fine and fast.

 

can you please test to create a new user on my BBS?

 

the hostname is: mytbox99.ddns.net

the port is: 61643

 

I used this morning only the Telnet client from my Windows PC. Maybe this was my problem.

worked fine for me! logged on, created an account and left you a message

 

[+9640News]

1 hour ago, wolhess said:

Hi Greg,

 

yes, I think too I have some faulty files. Now I can use my own user #2 very well. 
e.g the caller list is not correct at the end.
 

your hint with ssh is helpful. I used right now a ssh client from my iPad and it works fine and fast.

 

can you please test to create a new user on my BBS?

 

the hostname is: mytbox99.ddns.net

the port is: 61643

 

I used this morning only the Telnet client from my Windows PC. Maybe this was my problem.

The last caller list was removed from the menu, but not from hitting the "L" key to get the list.  This was intentional.

 

The list is maintained in memory.  If the BBS times out from an idle caller, or the BBS thinks there is a caller because someone did not close their telnet connection out quick enough (Windows Telnet does this), then it goes through the same timeout reset and corrupts the in-memory list.  I think it is due to a register value, but haven't found the solution yet to correct.  I'd be more than happy for someone to chase that down for me.  Basically, in the LIB file, look for the label B @BBS.  There is a test to identify if the interrupt timer has allowed 3 minutes to elapse without character input.  If it has, then the code branches back to the beginning of the program.

 

The other solution is to create another module for the caller history list and scroll that list.  That would require from a programming aspect one of two things.  One, deleting the existing call history list to free enough upper memory for the module, or, using 8K ram in the >6000 memory space to move the LIB code for the TIPI outside of the upper 24K to free up another 1-2K of memory.  As is, there are only a few hundred bytes of free memory in the upper 24K.

 

I'll send the update MENUMB and MENUMM files tonight when I get home.  Those two files are just 40 column DV/80 files that are displayed.

 

Beery

 

 

[+9640News]

OK, here are the all the files in my working copy of AfterHours that belong in the AFTER folder.  I apparently copied the files from tipi#2 instead of tipi#3.

 

 

 

Menus.zip

[+arcadeshopper]

btw I have a test board up as well, but I'm not about to put the PI on the internet without a filter so you can telnet or dial into my main BBS thekeep.net

and then type "/go dialout" at the main menu (without the quotes).. once there typing "D" to dial and then system 18 is Beery's, 19 is mine and 20 is wolhees's.. 6 is heatwave and 1 is hidden reef  

 

Greg

[+9640News]

OK, Greg identified a bug in the new user registration process.  Only the AFTERHOURS, AFTERHOURT, and AFTERHOURU files in this zip file need to be used to replace an existing installations.  This version was compiled for The Keep, so you may need to sector edit the name to your BBS name in the file and change the port number contained towards the end of the AFTERHOURU file where *.9918<space> needs to be replaced with *.<5 chars max with any unused chars as the space >20 char).  As previously mentioned, Port 23 is not useable.

 

 

 

AFTERHOURS.zip

[+InsaneMultitasker]

You need to add an "AfterHours Update" module so that you can force SysOps to accept and execute updates, preferably while a user is connected to the system.  <big grin>

 

 

[+9640News]

23 minutes ago, InsaneMultitasker said:

You need to add an "AfterHours Update" module so that you can force SysOps to accept and execute updates, preferably while a user is connected to the system.  <big grin>

 

 

Can you share the source you are using  for that feature also built into the new ROS for the HRD 4000+??

[+arcadeshopper]

You need to add an "AfterHours Update" module so that you can force SysOps to accept and execute updates, preferably while a user is connected to the system. 
 
 

Shutting down for update bye!

Reminds me of the days when BBS was so busy that I would have to stay up super late or get up super early and still have to kick people off to do updates or maintenance.

Sent from my LM-G820 using Tapatalk

[Omega-TI]

Remember when the autodial had to call 100+ times to get past the busy signals so you could connect?

[+InsaneMultitasker]

3 hours ago, BeeryMiller said:

Can you share the source you are using  for that feature also built into the new ROS for the HRD 4000+??

Are you really sure you want me to answer this question?    

 

I was thinking along the lines of using the TIPI PI.HTTP or similar to read a file in a repository (if publicly available) to check the current version number, which would then alert the user to an upgrade if present. For example, when you load CFG, you could poll for the version manually or automatically.

[+arcadeshopper]

Are you really sure you want me to answer this question?   [emoji6] 
 
I was thinking along the lines of using the TIPI PI.HTTP or similar to read a file in a repository (if publicly available) to check the current version number, which would then alert the user to an upgrade if present. For example, when you load CFG, you could poll for the version manually or automatically.

Doit

Sent from my LM-G820 using Tapatalk

[+9640News]

20 minutes ago, InsaneMultitasker said:

Are you really sure you want me to answer this question?    

Actually, as I think about it, it wouldn't be all that hard since one can already access a file from a website URL.  It would be not much more than LOAD the file then save it to a path.  One could check the version number at the site, load the individual fragments, save the individual fragments, then either flag on the screen an update is available, or if saved to the default path, load and run the program.

 

I would have to have a few individual changes in the program so the person would read a config file to set the port and the BBS name in the code.

 

Then, perhaps if the time matches a particular time/timeframe, it checks for the latest version.

 

Thoughts anyone?

 

Beery

 

[+9640News]

6 minutes ago, arcadeshopper said:

Doit

Sent from my LM-G820 using Tapatalk
 

OK, if it is just a flag to alert of an update, then I could probably just barely fit that into the upper 24K.  To automate the download, one would need an EA Supercart.

 

Thoughts?

Beery

 

[+arcadeshopper]

OK, if it is just a flag to alert of an update, then I could probably just barely fit that into the upper 24K.  To automate the download, one would need an EA Supercart.
 
Thoughts?

Beery
 

Sams?

Sent from my LM-G820 using Tapatalk

[+InsaneMultitasker]

1 hour ago, arcadeshopper said:

Doit

Sent from my LM-G820 using Tapatalk
 

I added a load option to CFG to grab the current ROS from GitHub. However, it seems the file isn't accessible directly when it is a binary.    I guess I would need to host the binary outside of the repository. 

 

2020-05-20 19:49:46,543 Pab         : INFO     opcode: Load, fileType: Sequential, mode: Update, dataType: Display, recordType: Fixed, recordLength: 0, recordNumber: 8192
2020-05-20 19:49:46,544 CurlFile    : INFO     load devname - PI.https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN
2020-05-20 19:49:46,545 CurlFile    : INFO     cmd: wget -O /tmp/CF https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN
2020-05-20 19:49:47,652 CurlFile    : ERROR    failed to load file - PI.https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN

 

Beery, if you pursue with standard equipment you could just check for a new version at startup.  Run that code from a buffer address since it is a once-and-done.  Just a thought.
 

[+9640News]

7 hours ago, InsaneMultitasker said:

I added a load option to CFG to grab the current ROS from GitHub. However, it seems the file isn't accessible directly when it is a binary.    I guess I would need to host the binary outside of the repository. 

 

2020-05-20 19:49:46,543 Pab         : INFO     opcode: Load, fileType: Sequential, mode: Update, dataType: Display, recordType: Fixed, recordLength: 0, recordNumber: 8192
2020-05-20 19:49:46,544 CurlFile    : INFO     load devname - PI.https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN
2020-05-20 19:49:46,545 CurlFile    : INFO     cmd: wget -O /tmp/CF https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN
2020-05-20 19:49:47,652 CurlFile    : ERROR    failed to load file - PI.https://github.com/horizonramdisk/Horizon-Ramdisk-ti994a/blob/master/ROS/ROSBIN

 

Beery, if you pursue with standard equipment you could just check for a new version at startup.  Run that code from a buffer address since it is a once-and-done.  Just a thought.
 

I tried something last night before it got too late.  I used MyWord to load a document from URI3.CONFIG , a DV80 file with a TIFILES header.  If the filename at the location matches the filename in the TIFILES header, file loads no problem.  However, if I rename the file so the Windows filename does not match the filename in the TIFILES header, no load.  I eventually get an error.

 

Does your filename on GitHub match the filename in the TIFILES header?

 

Beery

 

[wolhess]

Hi Beery,

 

I'm still testing the BBS. Now with your program files from yesterday.

 

I'm using the Telnet client from my Windows 10.

 

If I make a call to the BBS and it asks for the user # then I do nothing, the BBS timeouts after about 3 minutes. --> This is OK

The Telnet client lost the connection. --> Also OK!

 

Then after about 15 seconds the BBS screen shows the login screen again and ask for the user #. --> This is not OK I think!

Then it timeouts again after 3 minutes and the BBS is then waiting for a next caller.

 

The same thing happens when I login with my user #2 wolhess. The system shows the main menu and I do nothing.

After 3 minutes the client shows " lost the connection" and the BBS is waiting for the next caller, only about 15 seconds.

Then it shows the loging screen again.

 

So I tested this with the tipi Telnet client on my main TI system and get the same results.

 

I seems that the BBS has the caller request in a buffer and works from the buffer input.

 

Maybe you can fix this behavior.

 

You can test this on a faster way too:

1. Call the BBS

2. Press FCTN=  from the sysop screen

   --> the system makes the timeout and shows the unbind, bind status and then the login screen again.

3. Press FCTN= again and it shows again the login screen.

:

4. After more than 3 minutes or after three times the sysop was pressing FCTN= the BBS is waiting in a normal way of the next caller!

[wolhess]

Using the BBS:

 

If somebody logged in and the sysop press CTRL+Z to open a chat with the caller it will work.

 

At this time I see nothing to cancel the chat. Not from the user side and not from the sysop side.

The only thing I find out was pressing FCTN+= to cancel the connection.

 

Is this so or do we have a command to go back to the main menu from user side or from sysop side?

 

 

[+9640News]

17 minutes ago, wolhess said:

 

 

Then after about 15 seconds the BBS screen shows the login screen again and ask for the user #. --> This is not OK I think!

Then it timeouts again after 3 minutes and the BBS is then waiting for a next caller.

 

The same thing happens when I login with my user #2 wolhess. The system shows the main menu and I do nothing.

After 3 minutes the client shows " lost the connection" and the BBS is waiting for the next caller, only about 15 seconds.

Then it shows the loging screen again.

 

So I tested this with the tipi Telnet client on my main TI system and get the same results.

 

I seems that the BBS has the caller request in a buffer and works from the buffer input.

 

 

This is a behavior that I believe is not "fixable" within AfterHours without possibly increasing the "pause" time before the BBS goes through a fresh reset. I too have seen the same behavior as you and I believe it is something related to the Windows to PI connection.  It seems Windows is holding onto the telnet connection for a few seconds even when the Windows Telnet connection client has indicated it is closed.  I have not tested with Matt's telnet client, but with MyTerm for the TIPI, when I pressed G for goodbye, it was permanent and there was no re-login screen presented.  That leads me to believe it is more of a Windows issue.

 

I will confirm this tonight with MyTerm again to make sure.  I do know that if you hit <ENTER> as soon as you see the connection has been closed on the Telnet client, you don't get that screen.  That issue with the Telnet client is one of the things that forced me to put that three minute timer in place.

 

I may just go ahead and increase that "pause" time  between the fresh resets even more so the Windows Telnet issue disappears (hopefully).

Beery

 

[+9640News]

16 minutes ago, wolhess said:

Using the BBS:

 

If somebody logged in and the sysop press CTRL+Z to open a chat with the caller it will work.

 

At this time I see nothing to cancel the chat. Not from the user side and not from the sysop side.

The only thing I find out was pressing FCTN+= to cancel the connection.

 

Is this so or do we have a command to go back to the main menu from user side or from sysop side?

 

I believe if you hit the character "&" it closes the chat session.  I tested when I was chatting with myself some time ago, and I know I was able to exit the chat.  I believe that information was in the original AfterHours docs.

 

Beery

[wolhess]

4 minutes ago, BeeryMiller said:

I believe if you hit the character "&" it closes the chat session.  I tested when I was chatting with myself some time ago, and I know I was able to exit the chat.  I believe that information was in the original AfterHours docs.

 

Beery

Yes, a "&" from the sysop side cancels the chat. That's fine.

[wolhess]

8 minutes ago, BeeryMiller said:

I believe it is something related to the Windows to PI connection

I'm use the tipi Telnet program on my TI and the same behavior occure. So I think it has nothing to do with windows.

 

 

[+arcadeshopper]

Yes, a "&" from the sysop side cancels the chat. That's fine.

Yeah this is in the docs

Sent from my LM-G820 using Tapatalk