Anyone know a good newsgroup for Linux

[Dawn Minnis]

Hi

I know you're not supposed to post OS specific questions on the newsgroup so
does anyone know a good newsgroup - where there are actually people talking
to eachother and not full of spam - where I can get advice on developing C
for a command line driven interface in Linux?

Any advice greatly appreciated.

Kind regards
Dawn

[Chris Croughton]

On Fri, 11 Feb 2005 05:26:25 GMT, CBFalconer
<cb********@yah oo.com> wrote:

Chris Croughton wrote:

... snip ...

Since from your email address you're in the UK, you could look at
<news:uk.comp.o s.linux>. While a lot of the discussion there is
about non-programming stuff (installing, using etc.) there are a
number of knowledgable Linux programmers there and it is relatively
low-traffic (compared to comp.lang.c for instance).

(Non-UK people could also look at that newsgroup, but it is less
likely to be on non-UK servers, I know that it is on the BTInternet
one...)

Out of perversity I looked, and att does carry it (and 600 odd
other in the uk group) here in left-pondia.

Oh yes, many non-UK servers do carry uk.* newsgroups (some of them, like
the Berlin university public server news.individual .net, carry them on
auto-create), but they are less likely to do so than UK servers in
general, I've had (as UK Usenet Control) a number of queries "Where can
I get uk.<x> newsgroup, my ISP doesn't have it?" from people using
non-UK servers. I usually direct them to the free (for individuals)
server above, or for companies to gradwell.net.

Chris C

[Dawn Minnis]

Ok, well any chance someone could give me a quick explanation of what this
single line means?

Based on I, K and lda being integers
#define A(I,J) a[(I)-1 + ((J)-1)* ( *lda)]

If someone could just explain that one line to me i would be very grateful

The rest of the top of the program looks like this:



/* Subroutine */ int dgemm_(char *transa, char *transb, integer *m, integer
*
n, integer *k, doublereal *alpha, doublereal *a, integer *lda,
doublereal *b, integer *ldb, doublereal *beta, doublereal *c, integer
*ldc)
{
/* System generated locals */
integer a_dim1, a_offset, b_dim1, b_offset, c_dim1, c_offset, i__1,
i__2,
i__3;

/* Local variables */
static integer info;
static logical nota, notb;
static doublereal temp;
static integer i, j, l, ncola;
extern logical lsame_(char *, char *);
static integer nrowa, nrowb;
extern /* Subroutine */ int xerbla_(char *, integer *);
/*
Parameter adjustments
Function Body */

#define A(I,J) a[(I)-1 + ((J)-1)* ( *lda)]
#define B(I,J) b[(I)-1 + ((J)-1)* ( *ldb)]
#define C(I,J) c[(I)-1 + ((J)-1)* ( *ldc)]

[David Resnick]

Dawn Minnis wrote:

Ok, well any chance someone could give me a quick explanation of what this
single line means?

Based on I, K and lda being integers
#define A(I,J) a[(I)-1 + ((J)-1)* ( *lda)]

If someone could just explain that one line to me i would be very grateful

The rest of the top of the program looks like this:

It is doing treating an one dimentional array as a 1 based two
dimensional array. In C arrays are 0 based (i.e a[0] is the first
element). They want this array to be one based.

lda is not an integer, it is a pointer to an integer, which presumably
has the number of elements in the row.

e.g. if there are 10 elements in a row

A(1,1) -> a[0]; /* first element in first row */
A(2,1) -> a[1]; / second element in first row */
A(1,2) -> a[10]; /* first element in second row */

-David

[Mark McIntyre]

On Fri, 11 Feb 2005 18:55:42 -0000, in comp.lang.c , "Dawn Minnis"
<da*********@bt internet.com> wrote:

Ok, well any chance someone could give me a quick explanation of what this
single line means?

Based on I, K and lda being integers
#define A(I,J) a[(I)-1 + ((J)-1)* ( *lda)]

looks to me like an extract from Numerical Recipes in C, trying to use a
C-style one-dimensional array "a" as if it were a fortran-style 1-based
2-dimensional array "A". I and J are the row/column indices, and Ida is the
row width. They subtract one because they can't concieve of arrays starting
from zero.

Throw away that book, and buy the new edition, which was rewritten by
people who actually programmed in C.

--
Mark McIntyre
CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
CLC readme: <http://www.ungerhu.com/jxh/clc.welcome.txt >

[CBFalconer]

Dawn Minnis wrote:

Ok, well any chance someone could give me a quick explanation of
what this single line means?

Based on I, K and lda being integers
#define A(I,J) a[(I)-1 + ((J)-1)* ( *lda)]

If lda is an integer it will generate syntax errors. lda needs to
be a pointer to an integer. It looks like it is trying to map a 1
based indexing scheme into Cs zero based indexing. It will blow up
badly for I or J values of zero.

--
"If you want to post a followup via groups.google.c om, don't use
the broken "Reply" link at the bottom of the article. Click on
"show options" at the top of the article, then click on the
"Reply" at the bottom of the article headers." - Keith Thompson

[Dawn Minnis]

Thanks guys.

Its a line from the resulting C file after putting dgemm.f through the f2c
converter. The result it genuine C code but it looks horrid. But thanks
for explaining that to me. I wasnt aware that you could do that in C.

Regards
Dawn

[Mark McIntyre]

On Fri, 11 Feb 2005 23:33:04 -0000, in comp.lang.c , "Dawn Minnis"
<da*********@bt internet.com> wrote:

Thanks guys.

Its a line from the resulting C file after putting dgemm.f through the f2c
converter. The result it genuine C code but it looks horrid. But thanks
for explaining that to me. I wasnt aware that you could do that in C.

You can, but its dangerous, as it could invoke undefined behaviour for some
values of I and J. Better to understand the algo and write your own C.
--
Mark McIntyre
CLC FAQ <http://www.eskimo.com/~scs/C-faq/top.html>
CLC readme: <http://www.ungerhu.com/jxh/clc.welcome.txt >