Subject:
|
Re: (Failed) Getting LegOS 0.2.2 working on Windows
|
Newsgroups:
|
lugnet.robotics.rcx.legos
|
Date:
|
Tue, 21 Dec 1999 20:11:35 GMT
|
Viewed:
|
4619 times
|
| |
|
|
In lugnet.robotics.rcx.legos, Peter Van der Beken writes:
>
> > chili : /legOS util/dll demo/helloworld.lx
> > sending keepalive: Invalid argument
> Haven't seen this. could you try with the original loader.c, lx.c and
> rcxtty.c and tell us what error it gives then. (They should be backed up as
> loader.c.orig, etc. Rename the patched files as loader.c.new to keep them
> around). Also, use the -v flag (dll -v demo/helloworld.lx).
>
> Peter
Hey --
More info about this odd bug. Installed 0.2.3, and the bug is still there.
A bit more tweaking revealed:
* Compiling with the original (.orig, pre-patch versions of loader, lx, and
rcxtty) yielded - after changing O_ASYNC to FASYNC - the following -v output:
loader hostaddr=80 hostmask=f0 portmask=0f
delete try 0: ack:0
try 1: ack:0
try 2: ack:0
try 3: ack:0
try 4: ack:0
error deleting program
Okay, so no surprises here. The tower LED lights up, and the RCX even claims
to be receiving something (the "IR receive" section of the LCD flashes) - but
the tower apparently never hears anything. SIGIO handler not catching it? I
believe that's what your patch was supposed to fix, yes? Definitely something
for the cygwin folks ...
* Compiled with the patched versions of loader, lx, and rcxtty, and then
running dll.exe immiediately produces "sending keepalive: Invalid argument."
This is from the very first keepalive sent from the very first call fom
keepaliveInit. So the failure is happening *before* any of the new, WINNT-
specific event handling stuff is wired up -- soooo my guess is that the
failure is up in the WINNT-specific sections of rcxInit(), in the Win32-handle-
to-fd mapping stuff.
* Just dinking around, I nuked the WINNT-specific stuff in rcxInit(), and sure
enough, dll went back to transmitting (same output as the non-patched sources
yielded, though must FASTER, for some reason). The RCX flashed that it was
receiving something, but of course the SIGIO/event stuff wasn't wired properly
since the WINNT-specific goo was still in loader.c.
So the rcxInit() stuff appears to be the problem. Any further ideas on what
the heck could be wrong?
Purdy please? ;-)
-Jeff
|
|
Message has 1 Reply:
Message is in Reply To:
13 Messages in This Thread:
 
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
