Accessing ESP32/RunCPM via Network WiFi using telnet about runcpm HOT 18 CLOSED

from runcpm.

At the command prompt I got a double-CR (when pressing Enter/Return) so I do get

A0>
A0>

also after a DIR command and at the start of opening a file in Wordstar I seem to get another char which prevent the initial edit mode :(

because I didnt got this problem with a telnet-connect from linux/armbian I found the problematic setting in puTTY / KiTTY - which has to be disabled to get it working right ;)

from runcpm.

If you get distorted output from the ESP32-Telnet-Server on a Telnet-Client
(like the FabGL Network Terminal) then change in the ESP32-Telnet-Server
ESP32_telnetd.h

from

// start listening on the WiFi IP
server.begin();
server.setNoDelay(true);

to

// start listening on the WiFi IP
server.begin();
// server.setNoDelay(true);

this is because of the Nagle algorithm:
The purpose of the Nagle algorithm is to reduce the tcp/ip traffic of small data packets sent over the network by combining some small sending messages and then sending all the messages at once. The disadvantage of this method is that it delays the sending of a single message until a sufficiently large TCP-packet is assembled

from runcpm.

Yayyyy... telnetting into CP/M... \o/

I created a ESP32_telnetd.h and with a small change to the arduino_abstraction, I am now able to telnet into the ESP32/RunCPM board.

Can you make a PR for this?

Is there a blog somewhere that we can post code and make comments?
Github seems to be the wrong tool for that type of activity.

I don't think fragmentation of information is a good idea. So looking for a way how to keep blogging and documentation here would be my preferred choice.

ForthHub is using the issues of an extra repository as a kind of forum.
—▷ https://github.com/ForthHub/discussion
Let's have a look how they do it.

PRs have discussions too.
Can issues and PRs be cloned like the Wikis which just are git-repos too?
If not, that'd would be a showstopper and a GitHub based Wiki would be the better choice.

Doing it with Git and keeping it clone-able is better for the survival of the information.

Summary: Probably we can discuss code and have blogging here in many ways. We just need to define how to do it.

from runcpm.

from runcpm.

from runcpm.

I’ll give the code a check tomorrow evening to make all is good. I’ll have to figure out how to properly do github without embarrassing myself too badly.

Don't worry! I periodically screw up my own repositories and then need to pull them from GitHub/GitLab again to restart from a safe state. Shit happens! Life has too many other tasks that are far more important than becoming a never failing git expert...

If you would like to play with the code right away: http://web1.foxhollow.ca/RunCPM/

I'll look at it but maybe your polished version will be there faster than my results.
We'll see...

Stay ommmmmmPtimistic! \o/

from runcpm.

I imported the ZIP as new branch into my RunCPM clone.

from runcpm.

(yeti@kumari:15)~$ telnet doit-esp32-devkit-v1-1

After clearing the screen:

CP/M 2.2 Emulator v3.7 by Marcelo Dantas
Arduino read/write support by Krzysztof Klis
      Build Jan  7 2019 - 07:13:50
--------------------------------------------
CCP: INTERNAL v1.5    CCP Address: 0xe400
BOARD: DOIT ESP32 DEVKIT V1
Initializing SD card.

RunCPM Version 3.7 (CP/M 2.2 60K)

A0>info
RunCPM System Information v1.0

RunCPM Version 3.7 running on ESP32
CCP is Internal Version 1.5
CCP loads at address E400h
BDOS is at address EC00h
BIOS is at address FE00h
60160 bytes available for applications


RunCPM Version 3.7 (CP/M 2.2 60K)

A0>user 1
A1>c:
C1>do myhi
SuperSUB V1.1
Directory full error on line number: 5


RunCPM Version 3.7 (CP/M 2.2 60K)

C1>█

Ok... so far, so good.
I get some directory full errors when trying to compile stuff but I had this yesterday too without the telnetd changes. Maybe a waggly connection in the jumper wires or a bitchy µSD card or it just needs a filesystem check. Definitely not a telnetd related problem because I get this on the other ESP32 with the non-telnetd-RunCPM too.

I need to narrow that down before anything else.

Both DOIT boards have filesystem problems. One runs RunCPM with Telnetd, the other one runs the "normal" RunCPM. The filesystem problem shows up with a 2GB and a 4GB card in both boards.

So far I'm out of ideas what to look at next.

from runcpm.

Changes in my branch for experimenting (local+telnetd):

I've put my WiFi config in an own header which is added to .gitignore:
https://github.com/drawkula/RunCPM/blob/local%2Btelnetd/.gitignore#L62

There is an empty example for this file instead:
https://github.com/drawkula/RunCPM/blob/local%2Btelnetd/RunCPM/esp32_wifi_config.h.example

It is included here:
https://github.com/drawkula/RunCPM/blob/local%2Btelnetd/RunCPM/ESP32_telnetd.h#L12-L20

Maybe somewhen we'll have a better idea how to set SSID and PASSWORD without fearing to publish it via git.

So far for a summary of the stuff I did before dropping the pen because of the SD card problems. I need to wait for ideas what to examine next to track that down.

Update @ 20190109-1353-UTC:

Maybe the old SD library activated internal pull-up resistors or might have been slow enough not to cause errors with missing pull-ups and therefore slower signal rising times. The now faster new library may need the external pull-ups. Adding them looks ugly on the breadboard. I'll look for Wroom32 boards with an onboard SD card slot to have a mechanically robust solution instead.

from runcpm.

Can you please retitle this issue to include ESP32 and TELNETD?

from runcpm.

from runcpm.

Sorry guys I have been disconnected for some time.
You know, sometimes life gets in the way.
What did I miss?

from runcpm.

You know, sometimes life gets in the way.

Yes... confirmed!

What did I miss?

Telnetting to RunCPM.
It works!

I threw it in the corner because of SD card access errors. Meanwhile I have some more SD breakouts (still untested) but I'm not expecting them to behave better because I had write failures on the Wrover32 board with onboard SD slot too which should have those pullup resistors. I'm waiting for the right moment and mood to retry it but that's not a telnet related issue.

from runcpm.

Cool ... I will go read about it.

from runcpm.

This one is not really an issue anymore right? Can I close it?

(we need that forum)

from runcpm.

Cool. Closing this one then.

from runcpm.

@mbramwel @drawkula @MockbaTheBorg
Today I found this announcement and tested it with the normal v5.7 of RunCPM and didnt got problems with File-access by now :)

At the command prompt I got a double-CR (when pressing Enter/Return) so I do get

A0>
A0>

also after a DIR command
and at the start of opening a file in Wordstar I seem to get another char which prevent the initial edit mode :(

@mbramwel
Did you ever create another network-mode? ;)

from runcpm.