Message-ID: <3EEFD834.45AD84AE@yahoo.com> Date: Tue, 17 Jun 2003 23:10:44 -0400 From: CBFalconer Organization: Ched Research X-Mailer: Mozilla 4.75 [en] (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: djgpp-workers AT delorie DOT com Subject: Re: nmalloc integration issues: memalign, names References: <3EECBA0F DOT 9D6D4B30 AT phekda DOT freeserve DOT co DOT uk> <3EECD63E DOT 19C52226 AT yahoo DOT com> <3EEE24E7 DOT E7B03AB0 AT phekda DOT freeserve DOT co DOT uk> <3EEE3821 DOT 2C04E9D4 AT yahoo DOT com> <3EEF6925 DOT 43CCBDBB AT phekda DOT freeserve DOT co DOT uk> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Reply-To: djgpp-workers AT delorie DOT com Richard Dawe wrote: > CBFalconer wrote: > > Richard Dawe wrote: > > > > > ... snip ... > > > > > > I don't think it would be hard to implement. I'd do it myself, > > > except that I can't contribute any code at the moment, because > > > I need a disclaimer from my new employer. > > > [snip] > > I don't remember seeing any mention of it here. I am very leery > > of saying I can do anything now, because I expect to be taking off > > for a couple of months shortly. I can see my way to implementing > > it within nmalloc, though, if it doesn't need a special realloc. > > It was committed to CVS on 20 Jan 2001, according to my archives of > djgpp-cvs. > > Now I'm getting slightly frustrated - I feel like I'm banging my > head against a wall. We've tried gently persuading you to look at > CVS or a snapshot of CVS from MartinS every time nmalloc has come > up recently. Now you say you didn't know about memalign. That's > not really surprising, since you haven't looked at CVS (or a > snapshot thereof). Now take a look at the frustrations from my viewpoint. As a user, about 1.5 years ago, I found a serious deficiency in malloc. I volunteered to build a replacement. I was told to get on this list, and did. The discussions were basically "why don't you patch the existing", and I said the structure was incompatible. So I built a replacement, over about a month, and posted it. (It solved the O(N*N) probems I was having, which I found in several other systems). I was fairly proud of it. Then somebody pipes up about the debuggery systems. I had already put in provisions for such, without the set of hooks. Eventually a set of specifications were posted/found (the info .txh file), and I built something to match. Then a new set of requirements appeared - make matching documentation for something I had never used, the info preparation system. I did so. The next was a demand for warning free compilation under a set of switches I had never used. I did so. Then the switches turn out to be wrong (I am getting more recent, so more detailed). Fixed. Now I hear about memalign! The malldbg system was built from documentation that somebody gave me detailed instructions on 'how to acquire', which instructions are long gone from my memory and system. They produced a flurry of exchanges here as to permissable extensions and modification. I think the result was a good package, in fact better than the model it was built on, and reasonably idiot luser proof. I have no urge to learn a new system, and install who knows what, for each baby step. I have never used CVS, have no idea how to access your archives, am short of disk space, and don't want to spend money on lengthy connect times. I did download the existing library source and look at some spots way back, and none of this was apparent. That was long past the 2001 date you mention. Somewhere along the line the CVS archive was brought up. I did take a look at the CVS documentation, and it is unclear to me. It may be perfectly simple (probably is, once you get past the start), but meanwhile "here there be monsters". I am certainly not going to move some unknown tens or hundreds of megabytes onto this system when I don't have any use for it. I already have too many squirrel instincts, just ask my wife. I haven't spoken (in person) to anyone with a real idea of programming and system programming for at least three years, maybe ten. This medium is conducive to misunderstanding and misconception, since you can't grab a piece of chalk and scribble a diagram. I am still willing to install something that allows the memalign call to work, but this will get right into the warp and woof of the design. It is a potential source of evil insects, inasmuch as the possibility was not allowed for initially. I have grave doubts about the usefulness of such _for DJGPP_ in the first place. The nmalloc code does not have adequate regression testing, because I could never figure out a way of faking the sbrk calls in such a way as to return a preknown sequence of absolute addresses independent of system, compiler, whatever. I am not working with modern high speed machinery, wide band links, etc. I don't see any great need for them. I have a perfectly stable system, pretty well under control, and limited time and energy. I have saved uncounted thousands by resisting any urges to upgrade, and who knows how much frustration. The cost is that I am dependant on this system remaining functional, as I have no other available. As I have aged I have come to realize that I am not eventually going to learn everything about everything, and have put limits on my probing in self defence. -- Chuck F (cbfalconer AT yahoo DOT com) (cbfalconer AT worldnet DOT att DOT net) Available for consulting/temporary embedded and systems. USE worldnet address!