Date: Wed, 3 Jan 2001 10:49:05 +0200 (IST) From: Eli Zaretskii X-Sender: eliz AT is To: Martin Str|mberg cc: djgpp-workers AT delorie DOT com Subject: Re: v1: floating point on 386, v2 FAQ 11.1 In-Reply-To: <200101021934.UAA29017@father.ludd.luth.se> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Reply-To: djgpp-workers AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk On Tue, 2 Jan 2001, Martin Str|mberg wrote: > > I think in v1.x, you couldn't force emulation. I don't remember why, > > though. Perhaps because v1.x didn't call the DPMI functions for the > > emulation? (In fact, FP emulation didn't work at all in v1.x when > > running in DPMI mode.) > > But on a 386, using v1, I need the setting "387=YES" otherwise it > crash with a message ~"No coprocessor present" (I don't remember the > exact words). I guess the 387 detection code fails in your case. What happens if you set 387=QUERY ? On what OS is that? Plain DOS? what version? > > > Perhaps this setting was inverted in v1? > > > > Hmm? What did you meant to say by that? > > In v1, I need "387=YES" for it to work. > > In v2, I don't need any setting for it work. But if I would need one > the FAQ says I should use "387=N". I don't think there can be any inversion here, since these solutions are for two very different problems: in v1, setting 387=YES tells the go32 not to use emulation. By contrast, in v2, setting 387=N forces a DPMI function call (that wasn't there in v1.x) which turns on the FP emulation via the DPMI host.