floating point in c99

[Uno]

I seem to have an anger issue with englishmen.

Nothwithstanding, I know they would be great friends to me, as they have
always been. My favorite englishmen are the violin-daughters I remember
in New Koeln, where americans listened to music as opposed to shooting
people, indicating a brand of non-violence on the last thirty years in
central europe, where *real* american interest lies.

I have 2 questions for my english friends, among other respodents from
tennessee or a place that bush's coke-addled brain could quoth.

q1) why does bp still exist? I pay special attention to this issue.

q2) why is floating point fucked in C99?

http://gcc.gnu.org/c99status.html

Scroll down and see what is broken.

Please inform.

I have theories why C can't deal with a floating point, but my struggle
ends with a choice of a better syntax.

Hoping that the "special relationship" just doesn't mean we're fucking
retarded,

 

[Nick Keighley]

I seem to have an anger issue with englishmen.

Nothwithstanding, [..]  My favorite englishmen are the violin-daughters I remember
in New Koeln, where americans listened to music as opposed to shooting
people,

Americans aren't englishmen

I have 2 questions [...]

q1)  why does bp still exist?  I pay special attention to this issue.

what is bp? Is is a floating point variable of some sort?

q2) why is floating point [expletive indicating non-functional] in C99?

http://gcc.gnu.org/c99status.html

Scroll down and see what is broken.

I scrolled, I found nothing to indicate flaoting point was broken in
C99. Why don't you just tell us what you think the problem is?

Please inform.

please be clear and don't waffle.

I have theories why C can't deal with a floating point, but my struggle
ends with a choice of a better syntax.

in what sense can't C "deal with floating point"? There's a lot of
floating point software written in C. I suspect C is in the top 3
languages used for FP (I'm guessing Fortran and C++ for the other
two).

Hoping that the "special relationship" just doesn't mean we're <expletive>
retarded,

are you trolling? Have you been drinking?

 

[Seebs]

what is bp? Is is a floating point variable of some sort?

BP, as in the people with the oil drilling fiasco, which Our Hero believes
to be clearly a result of them being BRITISH Petroleum (remember when that
was true?) and also believes to be somehow topical.

I scrolled, I found nothing to indicate flaoting point was broken in
C99. Why don't you just tell us what you think the problem is?

Maybe he means that not all of C99 floating point is implemented in gcc?
Who knows?

are you trolling? Have you been drinking?

He's been like this through his last couple of changes of posting name,
if there's a pattern to indicate drinking vs. not-drinking, I couldn't name
it, but my experience has been that there's not much point trying to
discuss things unless by sheer coincidence he provides a topic you can
discuss without further input.

-s

 

[Rui Maciel]

I seem to have an anger issue with englishmen.

<snip/>

You are a poor troll. Ignoring your poorly framed rant against "englishmen" that you shoehorned
into your post, what you claim to be a floating point problem in C99 is at best a implementation
problem that affects a single compiler.


Rui Maciel

 

[Uno]

<snip/>

You are a poor troll. Ignoring your poorly framed rant against "englishmen" that you shoehorned
into your post, what you claim to be a floating point problem in C99 is at best a implementation
problem that affects a single compiler.


Rui Maciel

A person doesn't "shoehorn in" his first sentence.

I'm not poor at all, bitch. You, however, are a repressed catholic
homosexual with a daddy complex that makes you like authority and
therewith, The Topic.

What I don't know is why gcc99 developers are walking away from it in
disgust.

 

[Uno]

BP, as in the people with the oil drilling fiasco, which Our Hero believes
to be clearly a result of them being BRITISH Petroleum (remember when that
was true?) and also believes to be somehow topical.

The problem is that things are broken where they are not currently
acknowledged.

Maybe he means that not all of C99 floating point is implemented in gcc?
Who knows?


He's been like this through his last couple of changes of posting name,
if there's a pattern to indicate drinking vs. not-drinking, I couldn't name
it, but my experience has been that there's not much point trying to
discuss things unless by sheer coincidence he provides a topic you can
discuss without further input.

Alright, you win. I'm drinking.

Proper Aristoteleans begin with something less than germane and warm up
to the topic.

I would like that topic to be why the only compiler I could afford last
month can't do floating point. I hope I can catch Lawrence's or Louis'
eye on this, as they don't seem to care about the ideological bullshit.

I consider Plauger's implementation of float.h a touchstone:


/* values used by <float.h> macros -- IEEE 754 version */
#include <float.h>

/* macros */
#define DFRAC (49+_DOFF)
#define DMAXE ((1U<<(15-_DOFF))-1)
#define FFRAC (17+_FOFF)
#define FMAXE ((1U<<(15-_FOFF))-1)
#define LFRAC (49+_LOFF)
#define LMAXE 0x7fff
#define LOG2 0.30103
#if _D0 != 0 /* low to high words */
#define DINIT(w0, wx) wx, wx, wx, w0
#define FINIT(w0, wx) wx, w0
#define LINIT(w0, w1, wx) wx, wx, wx, w1, w0
#else /* high to low words */
#define DINIT(w0, wx) w0, wx, wx, wx
#define FINIT(w0, wx) w0, wx
#define LINIT(w0, w1, wx) w0, w1, wx, wx, wx
#endif
/* static data */
_Dvals _Dbl = {
(int)((DFRAC-1)*LOG2), /* DBL_DIG */
(int)DFRAC, /* DBL_MANT_DIG */
(int)((DMAXE-_DBIAS-1)*LOG2), /* DBL_MAX_10_EXP */
(int)(DMAXE-_DBIAS-1), /* DBL_MAX_EXP */
(int)(-_DBIAS*LOG2), /* DBL_MIN_10_EXP */
(int)(1-_DBIAS), /* DBL_MIN_EXP */
{DINIT(_DBIAS-DFRAC+2<<_DOFF, 0)}, /* DBL_EPSILON */
{DINIT((DMAXE<<_DOFF)-1, ~0)}, /* DBL_MAX */
{DINIT(1<<_DOFF, 0)}, /* DBL_MIN */
};
_Dvals _Flt = {
(int)((FFRAC-1)*LOG2), /* FLT_DIG */
(int)FFRAC, /* FLT_MANT_DIG */
(int)((FMAXE-_FBIAS-1)*LOG2), /* FLT_MAX_10_EXP */
(int)(FMAXE-_FBIAS-1), /* FLT_MAX_EXP */
(int)(-_FBIAS*LOG2), /* FLT_MIN_10_EXP */
(int)(1-_FBIAS), /* FLT_MIN_EXP */
{FINIT(_FBIAS-FFRAC+2<<_FOFF, 0)}, /* FLT_EPSILON */
{FINIT((FMAXE<<_FOFF)-1, ~0)}, /* FLT_MAX */
{FINIT(1<<_FOFF, 0)}, /* FLT_MIN */
};
#if _DLONG
_Dvals _Ldbl = {
(int)((LFRAC-1)*LOG2), /* LDBL_DIG */
(int)LFRAC, /* LDBL_MANT_DIG */
(int)((LMAXE-_LBIAS-1)*LOG2), /* LDBL_MAX_10_EXP */
(int)(LMAXE-_LBIAS-1), /* LDBL_MAX_EXP */
(int)(-_LBIAS*LOG2), /* LDBL_MIN_10_EXP */
(int)(1-_LBIAS), /* LDBL_MIN_EXP */
{LINIT(_LBIAS-LFRAC+2, 0x8000, 0)}, /* LDBL_EPSILON */
{LINIT(LMAXE-1, ~0, ~0)}, /* LDBL_MAX */
{LINIT(1, 0x8000, 0)}, /* LDBL_MIN */
};
#else
_Dvals _Ldbl = {
(int)(DFRAC*LOG2), /* LDBL_DIG */
(int)DFRAC, /* LDBL_MANT_DIG */
(int)((DMAXE-_DBIAS-1)*LOG2), /* LDBL_MAX_10_EXP */
(int)(DMAXE-_DBIAS-1), /* LDBL_MAX_EXP */
(int)(-_DBIAS*LOG2), /* LDBL_MIN_10_EXP */
(int)(1-_DBIAS), /* LDBL_MIN_EXP */
{DINIT(_DBIAS-DFRAC+2<<_DOFF, 0)}, /* LDBL_EPSILON */
{DINIT((DMAXE<<_DOFF)-1, ~0)}, /* LDBL_MAX */
{DINIT(1<<_DOFF, 0)}, /* LDBL_MIN */
};
#endif

q1) How much different is the implementation of float.h in gcc99?

 

[Nick Keighley]

did you not understand taht bit?

<snip drooling nonsense>

 

[Wayne C. Morris]

The problem is that things are broken where they are not currently
acknowledged.

The page at <http://gcc.gnu.org/c99status.html> does not say C99 is broken. It
says GCC does not comply with the C99 standard.

That page is for old versions of GCC. If you are using GCC 4.5, you should look
at <http://gcc.gnu.org/gcc-4.5/c99status.html>.

The problems ARE currently acknowledged. They are acknowledged on those status
pages. That is the official GCC web site. The C99 Status pages tell how close
GCC is to being C99-compliant. There are a lot of things marked "Done", so they
are very close to making GCC completely C99-compliant. The items that have not
been fixed yet are marked as "Library issue", "Missing", or "Broken".
Eventually those items will be fixed.

Read the "Further notes" on those pages to find out exactly what's missing in
GCC. It says IEEE 754 floating point works on hardware that supports it. If
your computer has an Intel or AMD x86 processor and is less than 10 years old,
it should work. There are a few things missing, like __STDC_IEC_559__, but it
you can avoid using those features.