X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f From: Radical NetSurfer Newsgroups: comp.os.msdos.djgpp Subject: Re: inport/outport BROKEN ? Date: Fri, 26 Aug 2005 05:08:40 -0400 Message-ID: References: <55mpg1h0hmo2nikgh2lb79vqht1puv25v2 AT 4ax DOT com> <200508242131 DOT j7OLVx75022085 AT envy DOT delorie DOT com> X-Newsreader: Forte Agent 1.93/32.576 English (American) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Organization: Velocity.Net Cache-Post-Path: web.velocity.net!unknown AT 66-211-206-127 DOT velocity DOT net X-Cache: nntpcache 3.0.1 (see http://www.nntpcache.org/) Lines: 47 X-Complaints-To: abuse AT newshosting DOT com To: djgpp AT delorie DOT com DJ-Gateway: from newsgroup comp.os.msdos.djgpp Reply-To: djgpp AT delorie DOT com In the DJGPP version the program does not seem to want to work; in the mingw version, that one fails after running. The ascii text is sent to a Text-to-Speech card via a outportb instruction, the status as to whether or not it is ok to send characters is checked by a inportw instruction. (yes the program has a way to escape incase it hangs) In the mingw application it appears to "stutter" at the 1st couple lines, works fine for about a couple minutes after that, than consistantly stutters to a point where its total gibberish. For whatever it is worth, a Borland 16-bit version of my application runs 100% of the time correctly and consistantly. Using the exact same code (at least in appearence). DJGPP has "issues" I get the stutter/scrambled speech instead of the proper characters coming together to form words. It is as if only certain acknowledgements from inportw() are "correct", and/or only certain writes by outportb() are correct, or even being performed. I honestly do not know. Someone suggested I examine the -S output to see what DJGPP is really putting in there for inport/outport; and do the same for Borland 16, and try to determine whats actually happening. I am aware of no real-time debugger for Borland 16 except through the IDE which I only rarely use. I am a command-line user. But probably should give that a try. I was thinking 32-bit mode vs. 16-bit mode somehow influences how inport/outport react/operate, and all that sort of line of reasoning. This would not be the first program under Borland that has inp* and outp* issues. Running it under a Windows DOS-Shell concerned me too, except the Borland one works fine (at least in 16-bit). Thanks! On Wed, 24 Aug 2005 17:31:59 -0400, DJ Delorie wrote: > >What do you mean by "fails miserably"? That's a very vague >description, and certainly not a useful way to get us to help you >figure out what's wrong. > >Those functions have been working fine for over a decade now, so I'm >pretty sure it's not actually broken.