problems

["Evan Langlois" <ekl@sdf.saomai.org>]

OK, a few problems and questions for everyone.

What FTP site is ttydev.tgz on?  I want to apply the recent patches
and compile for the new features (especially the thought of ANSI
style IBM-PC chars or the nice NVDI monoco font).

Can someone please send me the "test" binary or source file?  I can't
find it anywhere, and its no fun looking.  I know everyone has it.

What is the newest bash, and is it base-relative?  Mine doesn't parse
script files correctly - it almost always screws up the first two lines
or more of a file by dropping the first character of every line (I think
this is a Unixmode thing, but using "sh script" to let sh parse it instead
of bash works fine).

I'm using GCC 2.6.1, and while it produces notably smaller code than 2.5.x,
it does not support -mbaserel or -mpcrel.  I'm not sure if it supports
-mshort.  Does anyone know of a better version of GCC2.6.x?  I know - I want
my cake and eat it too!

Under GCC 2.6.1, compiling minixfs .60 pl 11, with the 2 other patches from
this list, gives the following errors :

minixfs/hdio.c:25: parse error before `('
minixfs/hdio.c:25: `trap_13_wwlwwwl' declared as function returning a function
minixfs/hdio.c:25: parse error before `short'
minixfs/hdio.c:31: `a' undeclared here (not in a function)
minixfs/hdio.c:32: `b' undeclared here (not in a function)
minixfs/hdio.c:33: `c' undeclared here (not in a function)
minixfs/hdio.c:34: `d' undeclared here (not in a function)
minixfs/hdio.c:35: `e' undeclared here (not in a function)
minixfs/hdio.c:36: `f' undeclared here (not in a function)
minixfs/hdio.c:38: parse error before `volatile'
make: *** [hdio.o] (null)

Can anyone tell me why?  The file looks fine to me!  
Can anyone confirm a compiler bug or recognize this as a common problem?
Should I post part of the source?

Anything I should check for?


Where are the MiNT MH3 sources/patches?  And does MiNT block the whole system
when the VDI is called?  If so, wouldn't this waste the CPU a bit when
doing a bit-blit, or worse, when printing!?!?

Does MiNT try and punish polling programs?  Can the priority of a program
be lowered every time it makes a polling-type call (either Bcon or Ccon
or related "check for input" call) that fails to find input.  And possibly
reset to some default once a poll DOES find input.  Also, a poll should
always give up the remainder of the time-slice - like a built-in Syield.
Is this a good idea or bad?

I've heard that MagiC gives MORE time to programs that use alot of CPU power,
and less to those that don't need it.  MiNT punishes programs that use
alot of CPU crunching.  Which way is correct, should programs that eat CPU
without calling the OS be punished or given what they need?

Any word on non-blocking disk drivers for ACSI?  (OK, bad question - can't
hurt to ask)..