Message-Id: <199812171359.IAA15907@pop02.globecomm.net> MIME-Version: 1.0 X-Mailer: Denshin 8 Go V1.9b1 Date: Thu, 17 Dec 1998 22:59:23 +0900 X-My-Real-Login-Name: adamtk AT altavista DOT net; mail.iname.com From: "Toshio 'ADAM' Kudo" <adamtk AT altavista DOT net> To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> Cc: djgpp-workers AT delorie DOT com, Charles Sandmann <sandmann AT clio DOT rice DOT edu> Subject: Re: patches to 2.02 Reply-To: djgpp-workers AT delorie DOT com Hi and KON-NICHIWA. Eli Zaretskii wrote: >The changes I refered to cited CWSDPMI-specific behavior. I was under >the impression that these changes assume something about what the DPMI >host does when the call to __dpmi_get_coprocessor_status returns a >failure indication. I am worried that other DPMI hosts (which also don't >support __dpmi_get_coprocessor_status) might behave differently, and that >setting DPMIfpustate to 1 is not the right thing to do with these other >DPMI hosts. Thanks. I understand. DPMI 0.9 specification says, must returns only carry flag set if error. # Wow! DPMI 1.0 says return carry set and ax must contains negative # number as error code. But, __dpmi_get_coprocessor_status must return -1 if error occures. Please see src/libc/dpmi/api/dpmidefs.h, src/libc/dpmi/api/d0e00.S. So, that patch will work well on all DPMI hosts, I think. *If*, there is a DPMI host that not support 0E00h and never returns carry flag set, ... mmm ... I don't think about that :-). >Thanks for the rest of information. My pleasure :). -------- ADAM, as Toshio KUDO GCG02632 AT nifty DOT ne DOT jp / adamtk AT altavista DOT net