Message-Id: <199801090827.KAA11233@ankara.duzen.com.tr> Comments: Authenticated sender is From: "S. M. Halloran" Organization: User RFC 822- and 1123-Compliant To: djgpp AT delorie DOT com Date: Fri, 9 Jan 1998 10:28:29 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: Re: problem with memory allocation (I think) CC: djgpp AT delorie DOT com In-reply-to: <199801090150.RAA16234@adit.ap.net> Precedence: bulk On 8 Jan 98, Nate Eldredge was found to have commented thusly: > At 11:07 1/8/1998 +0200, S. M. Halloran wrote: > >Well, yes...you are certainly challenging the system's memory > >protection system...it seems. Since it is W95, I am a little > >surprised that you were not immediately sent to Land of Warm Reboot > >:) > Windows 95's limitations would not affect you in this case. The problem it > has is with unmapping specific pages, which implements NULL pointer > protection. Pointer overrun exceeds the segment limit, which does get caught. > > But I agree in principle. :) > Actually, I would really like to learn more about how the Windows 95 "kernel" deals with memory protection...as often as it crashes , I rather think it is letting applications have the run of the mill. It appears to have no code to have all exceptions and interrupts go to MS code as last resort, to try to allow for graceful exits. I suspect a great many of the bugs are in MS or 3rd party network code used by PCs and does not happen so much for PCs that, if they network at all, use dialup. I say this because people who love W95 and claim it "never" crashes for them often have no network card installed; it may be coincidental. However, I can imagine Microsoft possibly rushing to ship out Windows 95 before it had to be called Windows 96, and not worrying about a lot of the "accessory" features of personal computing, such as networking. I only recently started using a PC in which installation of Windows 95 was an order of the boss, and so I am pretty much a beginner with respect to understanding how any Windows product of Microsoft pretends to run. There must be a web page or resource on the net that I can get access to without paying Microsoft $3000 (or whatever) for a peek at the developer kits (and maybe also for the memoranda passing between the programming development staff and MS' chairman regarding shipping dates and buggy code). Perhaps then I can understand why Microsoft has (possibly) failed to understand how to properly make use of features of Intel microprocessors that have been available for years. (I think I read somewhere that Intel engineers have said or expressed amazement that the programming geniuses at Microsoft have still not gotten it right.) Mitch Halloran Research (Bio)chemist Duzen Laboratories Group Ankara TURKEY mitch AT duzen DOT com DOT tr other job title: Sequoia's (dob 12-20-95) daddy