Mail Archives: djgpp/1997/09/16/12:57:06
Yeah, whatever, if you dislike DJGPP so much, get the hell out of the
newsgroup.Jeez.
Brian
At 10:25 AM 9/16/97 GMT, Russ Williams wrote:
>Herman Schoenfeld <you AT somehost DOT somedomain> wrote in article
><341e2691 DOT 0 AT 139 DOT 134 DOT 5 DOT 33>...
>> In article <01bcc1b3$ccb39840$2b40cbc2 AT russnt>,
>russ AT algorithm DOT demon DOT co DOT uk
>[...]
>> >The code it produces being worse than anything else? DJGPP doesn't
>> >support Pentium optimising, yet VC5 supports the PPro (Watcom and
>> >Borland probably do too). It isn't in the same league as commercial
>> >compilers. GCC doesn't even support C++ as well as everything else -
>> >templates are a good way to see 'internal compiler error', and I've
>> >personally had it fuck up on function calling (pushing the wrong
>> >parameters, ignoring return values). GCC is great for doing C programs,
>> >and allows code to be remarkably portable, but it isn't the best.
>>
>> DJGPP isn't that bad. GCC on unix/linux is standard. There are no MSVC
>> compilers for them so any comment comparing GCC to MSVC is pretty
>> much a waste of bandwidth.
>
>Is the x86 code for Linux apps different to x86 code for Windows, then?
>MSVC could (with a little difficulty) be used to code for any Intel
>platform.
>
>> If you're comparinh MSVC with DJGPP, you're wrong in all instances.
>
>Sorry, not today.
>
>> DJGPP does support c++.
>
>Yes, but DOES IT WORK? I know DJGPP claims to support C++
>(.cc files), but for anything but the simplest code, I certainly wouldn't
>trust it.
>
>> DJGPP does support pentium optimising. (PGCC).
>
>i) PGCC != DJGPP.
>ii) Pentium MMX? PPro? Pentium 2?
>
>> DJGPP produces fast optimized code.
>
>Compared to Turbo C++, yes. Compared to a real compiler, no.
>
>> Sure, DJGPP doesn't have nice little point-and-click features but most
>people
>> don't need them.
>
>Gee, I'm such a bad coder that I like source-level debugging and a single
>key/icon to build the project. Just because DJGPP isn't easy to use,
>doesn't mean it's the best.
>
>> >GCC is a very nice, capable, free compiler, but FFS, the best people in
>> >compiler optimisation earn lots of money working for MS, Intel, Borland,
>> >Watcom, Symantec, SGI, Sun, DEC, HP etc. They don't work for free.
>>
>> With flawed logic like that its no wonder you have such trouble with
>> programming.
>
>I have trouble programming? Since when? Why hasn't someone told me
>about this? (Hrmph. Heads will roll...)
>
>> Just because somebody charges $250 p/hour to produce a compiler
>> like Turbo C++ it certainatly doesn't make it better than something
>> produced by hundreds of people who already make enough money
>> and contribute to a compiler such as DJGPP.
>
>No, but Watcom 11, Borland 5, MSVC 5 are all better than DJGPP.
>Not in terms of gcc-induced portability (portable code isn't something
>gcc invented, BTW), but in terms of code generation, debugging, IDE,
>online help, support.
>
>> You can put all your compilers together and you won't get even half the
>support
>> DJGPP has.
>
>'Support' doesn't mean the compiler is any good. DJGPP and GCC in
>general are very standard, but are not the best. The binary-only
>compilers that are available on most systems are usually better. Just
>because the PC doesn't have a compiler shipped with the OS doesn't
>alter the fact that GCC isn't up to commercial standards.
>
>BTW - Most of the support I've seen for DJGPP involves "don't worry,
>that's fixed in 2.8.0". If any commercial compiler manufacturers find
>a critical bug, they'll patch it within days - their livelihood depends on
>it. That's a damn sight bigger incentive than anything DJGPP/GCC
>can come up with.
>
>---
>Russ
>
- Raw text -