Mail Archives: djgpp/1997/02/01/00:07:21
cigna AT helios DOT phy DOT OhioU DOT Edu (Dave Cigna) wrote:
><afn03257 AT afn DOT org> wrote:
>>Nope. GCC, has not always been 100% fully ANSI complaint, there were
>>"features" that made it act differently. Read the docs on the GCC
>>updates line through the times, Often and I mean very often, bug XX
>>fixed to aquire more ANSI conformance, etc.. was written.
>Do you have any real experience using these compilers, or have you
>just read the docs thoroughly?
Does that really matter? Whether I've read about the bugs or
experienced them? To answer your question both, I like to read about
bugs before I experience them if possible. Now, judging from the rest
of your post may I ask you a question? Were you born in an primarily
English speaking community or is it a second langauge? Your writting
is fine, but your interpretation of my statements is alarming.
>You've made two broad claims:
Are you sure?
> 1) Borland C is comparatively bug free, while DJGPP (GCC in
> particular) is infested.
Bullshit. I said Borland has always released bug fixes fairly quickly
on their www/ftp site. Does that look like bug free, comparatively or
otherwise, to you? I don't think so. I said that GNU has as many bugs
as Borland, I never said either were bug free in any way. That clears
that up I hope.
> 2) Borland C is virtually ANSI compliant, while GCC has been
> continually struggling to achieve ANSI compliance.
You've mis-read the statements again. I never said anything about
Borland's compliance. I asked how long it took to get enough bugs out
of GNU C for it to be considered complaint. Borland had bugs, has bugs,
and always will. However, when we side step back to DOS only Borland
products, the bug count reduces drastically, and in fact, is comparable
to GCC w/ DJGPP's record. Bugs are normal, not bad, but normal in large
software applications. What makes them bad is when the company refuses
to fix them. While GNU C w/ DJGPP is not exactly a company, they do fix
them, and so did Borland as long as DOS was still being supported.
Windows is another story. Some bugs are actually in the windows API and
can not be fixed by Borland. Now, have you even seen a bug fix from MS
under windows? Under Dos, other than DOS patches for $10? Me either.
Watcom? Nope. Have you ever seen any compiler patches distributed with
Simtel, night Owl or other shareware/freeware CD-ROMS? Nope? Borland's.
are!
>The thing is, you've offered no evidence whatsoever except for
>some references to GCC bug reports and updates. (You might want
>to consider the fact that Gnu's policy on such things is *ENTIRELY*
>different from Borlands.)
The thing is, you're talking about things I never said and therefore
I'm not going to try and justify actions I never made. nor would
you expect me to offer evidence about things I did not say.
>I have used both compilers, as well as a variety of others, and
>my own entirely anecdotal (but entirely real) experience is that
>GCC is by far the most bug free compiler on the planet. (I removed
>Borland from my HD and gave away the diskettes and books; I couldn't
>stand it locking up or rebooting my machine anymore.) Exactly the
>same goes for it's ANSI compliance. I have found that my code
And Borland ran under windows right? DJGPP runs under? Ooh what's
that, DOS? The more stable OS for home PC's? If your Borland C++ was
DOS based then you are a rare one, kind of like me and Linux, I'm told.
I've used Borland and Turbo C++ versions from 1.0 and never experienced
anything like you claim unless I was using risky code, in which case I
was the one who caused the crash and not the compiler, And I've already
crashed under RHIDE 1.1 [forget 1.0] about 4 times, so ....
Everyone raves about Linux, yet, is trashed my system 3 times and
never installed correctly. I tried all 3 distributions, BTW. I gave up,
Linux is obviously not for me. Maybe I'll try FreeBSD, or the Que
Linux CD-ROM one day. Que book+products always seem to work well
for me. However, I'm not going to claim Linux is an absolute piece
of junk. Now, are you going to claim Borland's is because you
couldn't get it working right? Only if your 3 years old maybe.
>behaves as predicted far more often with GCC than any other compiler.
>(Borland is king of added 'features' that make their products
>non-standard. Just look at Turbo Pascal!)
Huh? We are not talking about Pascal and BTW QPascal wasn't any more
complaint, but C(++). Like it or not, anyone in c.l.c++ will
tell you that between Symantec, Watcom, MSVC++, GNU C++, and Borland
C++, that Borland has been following the Standard proposal the
closest. Each release has supported any additions or alterations the
proposal has made. Perhaps you better read what I say before you accuse
me of saying things I didn't, and perhaps you better learn to
distinguish between Borland C(++) and Turbo Pascal. Borland NEVER
claimed TP was ANSI/ISO complaint did they? However, they did claim
ANSI C compliance and they delivered it. In fact, if memory serves they
were the first commercial vendor for the PC to offer it and to get the
bugs out. MS is still working on it. If Borland is so bad, why is RHIDE
like Borland's IDE? Why is TVision ported to GCC? Why are Turbo Pascal
extensions attempted to be supported by GPC? I think you may have
jumped on a band-wagon without knowing any facts first. GCC is a great
product, but being hyped up like BatMan, doesn't make it superior.
>>This is true, however, technically you could patch the commercial
>>software yourself with a debugger.
>What planet are you from?
You might want to answer this one yourself after your two false
assumptions there, Mac. I have patched commercial software using debug
before, it is POSSIBLE! In fact, I patched EDIT using ASCII coding
techniques for machine code in EDIT. It is utterly amazing what one can
do when one is determined. Granted I would have preferred HLL src code
to patch, I got the task accomplished just the same.
- Raw text -