cr88192
5/17/2011 1:19:00 AM
On 5/16/2011 4:34 PM, Seebs wrote:
> On 2011-05-16, Mel Smith<med_cutout_syntel@aol.com> wrote:
>> Are there any common and current 'standard' C implementations that are
>> more acceptable than others (to those people and experts posting here) ?
>
> Not that I know of. I don't see an implementation as "acceptable" or
> "less acceptable". About all I can think of is that Visual C++ used to have
> a really dismal C implementation, but that was years ago.
>
MSVC is still a bit weak.
it works though...
>> My 'Hi-C' language (both Harbour and xHarbour versions ) use (among
>> several others) both Borland's C/C++ Version 5.5.1 and MinGW Version 4.5.2
>> for compilation.
>
I have several of my own languages, just they are mostly OT here...
pardon the OT-ness:
I have one I call BGBScript, which as-is mostly looks like JavaScript
and ActionScript (similar syntax and core language features). sadly,
this is my only really "usable" languages at the moment. it has existed
in my projects since about 2004.
I had tried to implement a new language and VM I had called
"BGBScript2", which would differ mostly in using stronger type-checking,
as well as a more Java-like declaration syntax. however, this project
has since been put on hold (I have more important stuff to do).
also a considered language I called "C-Aux" which would mostly look and
behave like C99 (and would likely be source-compatible with most C
code), but would have many subtle differences mostly for implementation
reasons (mostly to make it better suited for a VM environment), however
if/when I will do much with the idea is uncertain.
I also have my own C implementation (most of C99, minus VLAs and a few
legacy features), but it is not presently usable (I couldn't really ever
get it "usably good" and it has since fallen into bit-rot...).
(C makes a poor script language, and for everything else I can
static-compile it...).
also, a custom mini-JVM implementation (but it is based on J2ME and
depends on 'javac' for compiling, so lacked much real point).
BGBScript2 was meant as a compromise between BGBScript and Java (and
partly so I could have something "sort of like Java" but unchained from
legal and trademark issues).
more recent effort though has gone into continued expansion and
maintenance of BGBScript, me mostly having to keep a more narrow focus
to keep productivity and code-reliability up.
most of these generally work via bytecode interpretation and/or JIT
compilation.
>> How are these implementations just noted rated among the many available.
>
> I don't think any such rating exists. Different implementations are good
> at different sorts of things, or have different attractions. Some people
> will like one or another more, but it's basically up to personal taste or
> the technical demands of a specific project.
>
> Me, I don't really *care* about any non-Unix-like implementations, because I
> don't have much reason to deal with them or worry about them, but that doesn't
> mean I "rate" them more poorly or anything, I just don't use them. For all
> I know MS's compiler is actually good these days.
>
as of VS 2010, MS added "stdint.h"...
hence, one has "long long" and "stdint.h"...
what is still missing?
_Complex;
VLAs;
inline;
mixing statements and expressions;
....
I would personally like extra to see the ability to give errors over
implicit function declaration (say, an analogue of
"-Werror-implicit-declarations" or similar...).
but, no huge loss though...
or such...