[Coco] making nitros9 on MinGW

[Robert Gault]

Bill Pierce wrote:
>
> I completed my nitros9 build and looked through the log. It seems there were some problems with Sierra disks and running out of space. This shouldn't be happening. All these disks were built before with the same size disks and barely no changes have been made to their code... only to the code that builds them.
> I haven't had a chance to do an indepth check but hope to this weekend to see how many disks built correctly.
>
> Bill Pierce

Bill,

I'm having exactly the same problem on a WinXP 32-bit system. The 248 errors on 
the disks are random, divided among floppies, dw, and becker disks. All of these 
disks have sufficient room and the files can be added by emulators or MESS 
utilities.

I've tried one test that generated interesting results. I split the SUPPORTFILES 
entry into several groups about half the size. Now instead of getting just the 
248 (no room) error, the first group got 248 but the second os9copy got 
'multiple copies require a directory as a destination.' The makefile commands 
were identical.

Regards the os9copy command, there is no real difference between the current 
makefiles and an older version of NitrOS-9 where all disks build correctly. 
MINGW is the same for both. The only difference I can point to is that the 
versions of msys are different. The one (msys) that works is older and seems to 
be 1.0.10. I'm not sure of the version for the newer one.

Robert