wconsd

The libcli git repo is one dir deeper than expected, fixup all the references

The libcli project is now kept in a git repo, so we can point at the real one

CLEANUP: uncommitted note found

Document a future direction

Add an example dummy SCM for unix

Pass the args to the main function and thus allow the init function to be optional.

Use the SCM args if they are provided, otherwise just use the cmdline ones

Update win-scm service detection from the svctest results

Add missing definition

Update win-scm with new learnings, also add a service test helper program

Add more libcli commands to the main program

Quieten down the failed service error path

Fix wconsd_init to exit if the getopt processing indicates to do so

Fix the bugs with the new scm console code

Move the console mode handling inside the SCM for increased consistancy

Ooops, meant to hook the getopt processing in while cleaning up main

Clarify the meaning of one of the dprintf variables

Clean up the main function some more

Move the libcli initialisation to earlier in the init function

Move to getopt commandline processing

Housekeeping - ignore object files and reduce warnings from the scm

Hook the module management system into a startup routine

Add the beginnings of a module management system

Add libcli to the mix

Add another method to the com port enum test program

Add the libcli project

Remove the net thread. Simplify some of the control logic

Begin to add code allowing the merging of menu and net threads

Clean up scm helper

More work on the SCM helper. This looks like it is working

Move the Service Control functions to a separate file - first pass

Document which files to ignore

Dunno

Make telnet option processing cross packet boundaries. Implement binary mode.
These two changes allow xmodem uploading from client to serial port.

Show link to documentation in menu help

Handle WSAECONNRESET, add Telnet keepalives, rework the show_conn_table.
During long-term connection testing, I found that at least one of my
connections would have some kind of disconnect - perhaps an idle timeout,
so I have added support for sending IAC NOP packets on each net loop...

Begin testing in an attempt to enumerate serial ports

Update run_menu to get it to detect a closed network connection. Also message beautification

Clean up messages and provide better help information

Implement send break function (closes #23)

Implement the ability to kill existing connections
This is a new command to force quit any current serial connection

Uncook CRLF sequences into plan CR in net_to_com
My testing had been done with the linux telnet client and it sent CR NUL
as the end-of-line character sequence. Both putty and windows telnet
are sending CR LF as end-of-line sequence. This change pretty much...

Update version number

Add telnet processing to serial-data-pump. Add ability to exit back to menu.
There is still a lot of cleanup work remaining from here. We do not need the
menu thread, we should run the menu from the net_to_com thread, however it
appears to meet the requirements. closes #16, #20

Migrate threads to new connection model, allowing multiple connections.
This is the first part of migrating the two threads to use the new
connection table. This change also rips out a lot of the event waiting
system. All the unused variables are going, but I have not yet looked...

Minor build fixups

Update copywright and version

Convert run_menu to not busy wait
It took me a while to notice, but the run_menu is written using
blocking file metaphor, but the socket it is working on is a non
blocking socket. This manifests itself with the run_menu loop
consuming most of the available CPU time

Add telnet option processing, including ECHO support
The amount of effort here seems like overkill, but it seems
to be the practical minimum implementation for telnet clients
to do the right thing. I have added in more decoding than is
strictly required and I hacked in the status suboption so that...

Try using getnameinfo to show more information about connected users

Add connection table debugging, net byte counting, some more thread interlocks

Migrate most of the menu system into the connection structure

Clean up lots of compile warnings.
I have been getting lots of errors while testing under wine, so this
is the start of looking for problems. I have not cleaned up any
of the warnings in process_menu_line since that whole function is
an abortion and all the warnings are either send() signedness mismatch...

Start to implement a connection table
This change is only half baked as I still have not altered the rest of the
threads to use the new connection table. Nor have I updated the event
signalling model to take into account the existance of multiple connections.

Add winelib support and the ability to set the listen port

More cmdline processing fiddling

More cmdline processing hacks

Fix mode autodetection, change default tcp port back to 9600
It turns out that my telnet client doesnt work at all when I default
to port 23. Instead of debugging this problem at the moment, I am setting
the default port back to 9600

Detect if we are running as a service or console app. Change listen port to 23

Add a quit option to the menu and update help texts

Use strtok for commandline tokenisation

Factor the menu handling code and slightly clean it up.
closes 18

Add a makefile target for testing the application

More debug message changes

More debug print statement cleanups

Dont copy as much stuff to the windows compile system

Sanitise the excessive number of lines used by all output messages

Create a dprintf function and convert all log messages to use it

Update version to mark this one as compiling

First version that compiles with mingw. Tested and working

Update makefile to stop using lcc
It turns out that my previous testing was all invalid as lcc could
not even correctly compile a hello world program.

Start makeing some minor changes

Move the include statements around - this version compiles

Remove CR's from every line
I noticed that not every line had a CRLF and that this was screwing up viewing
diffs and confusing various editors. So, I removed all the CR's

Add compile line to makefile

Add enough CR's to make vi realise that this is a dos format file

Add initial makefile

Newer version from ftp://ftp.h07.org/pub/h07.org/projects/system/wconsd-0.1.zip

Original version from http://www.greenend.org.uk/steve/software/wconsd.c