Mail Archives: djgpp/1995/01/11/04:06:41
On Wed, 11 Jan 1995, Stephen Turnbull wrote:
> From: Bill Davidson <bdavidson AT ra DOT isisnet DOT com>
> Mr. Murphy demands, of course, that this was always the case when info
> choked. However, there generally seemed to be more than 2000K
>
> This is a serious "oops!", I should think. Below, GO32 reports *less*
> than 2000K with *no* program loaded. Presumably that's with something
> memory-using (RAMdrive or combi).
>
Oops indeed! I am glad you pointed that out. I just tried go32 and info
again (I have since switched to combi, enabled cache, and set size to
1536KB). I have been trying various configurations and scribbling the
results on scraps of paper that I always keep on hand or find lying on
the floor, so I wasn't sure I had reported numbers from the same config
to you (although was _pretty_ sure). Now go32 reports 1400Kb, and
topline under info reports 1768Kb!! (aside: surely KB not Kb; I always
thought KB meant KiloBytes and Kb meant Kilobits!).
> up. I do note that the return stack trace shows that you're dying in
> different routines at different levels of nesting every time :-(
> That's beyond my capability to debug. Somebody else said that they
> had a "-g" version of info, maybe they can symify it.
More bad news! Since I have trouble reading the docs [ ;( ] I don't know
how symify works (or what it is); I presume -g means "compile with debug
info" (-g == debug; Of course!). To get this do I have to get all the
sources, re-apply the maint releases, and compile?
QUERY: How do I read the top line under GO32=topline? What is that hex
number in the middle, and why does it write over everything while
everthing writes over the stuff I want to see? How do I know when go32
is paging, etc?
Also, I don't see any docs for symify, although symify is in my bin
directory. Is it buried in there somewhere? I have spent some time
squinting at the .inf files to glean what I can, but haven't seen symify
in there.
> Have you tried running in the QDPMI environment? This costs some
Yeah, no difference except QDPMI seems to steal some memory from go32
(although I don't think I tried QDPMI and GO32=nodpmi). Initially I
thought go32 would require a DPMI host like QDPMI to run so loaded it,
but it seemed to make things worse rather that better!
>
> combi is using "all available memory" but Info "ran great"? OK, we'll
Well, combi is using all *extended* memory, or claims to (actually, all
but 32K, according to all my mem utils). The reason I have it set now at
1536K is I have been trying to increase it to try to force info to VM
where I can see it. Still getting exceptions, segmentation violations,
and protection faults. That is what I find weird. The same sequence of
commands under the same environment produces different (or differently
reported) exceptions at different points in the program. I would have
thought that I would see the problem at *exactly* the same place after
*exactly* the same commands (and it would raise the same error message).
> Weird, weird, weird. You could try getting a new, clean copy of Info,
> maybe it got trashed.
probably best idea yet!
> You could compile Info yourself. Then you can have the "-g" flag
> and get debugging info with symify.
another good idea
> Have you checked for virii recently?
Yet another good idea, but if this is caused by a virus, it came in on
djgpp!
> I hope it doesn't take the *whole* new year to get DJGPP
> running on your system!
> --Steve
Thanx, and BTW, I know you must have a REAL job, not just answering my
questions, and I sure appreciate the trouble you have taken so far to
help me out! Have a good one yourself, and if you are ever in Halifax,
Nova Scotia, look me up and we will share a virtual ale!
Bill Davidson
bdavidson AT ra DOT isisnet DOT com
- Raw text -