Yahoo Groups archive
Thread
2005-01-28 by gene_klein2000
Hi, I hope someone will help me with the next problem. When I'm using vsprintf the code generated by th compiler grows with about 30kB. Can anyone explain what I can to keep the code small and still be using th vsprintf. I'm using the gnuarm compiler. Thanks, Gene
2005-01-28 by Leon Heller
----- Original Message -----
From: "gene_klein2000" <gene_klein@...> To: <lpc2000@yahoogroups.com> Sent: Friday, January 28, 2005 11:48 AM Subject: [lpc2000] large code generation with vsprintf > > > Hi, > > I hope someone will help me with the next problem. > > > When I'm using vsprintf the code generated by th compiler > grows with about 30kB. > Can anyone explain what I can to keep the code small and > still be using th vsprintf. > > I'm using the gnuarm compiler. Functions like printf have a lot of code in them. I'd write my own function. Leon -- Internal Virus Database is out-of-date. Checked by AVG Anti-Virus. Version: 7.0.300 / Virus Database: 265.7.1 - Release Date: 19/01/2005
2005-01-28 by Marko Pavlin
Use keil DKarm if you have budget for that. Optimization is better than using GCC. Leon Heller wrote:
> ----- Original Message ----- > From: "gene_klein2000" <gene_klein@...> > To: <lpc2000@yahoogroups.com> > Sent: Friday, January 28, 2005 11:48 AM > Subject: [lpc2000] large code generation with vsprintf > > > > > > > > Hi, > > > > I hope someone will help me with the next problem. > > > > > > When I'm using vsprintf the code generated by th compiler > > grows with about 30kB. > > Can anyone explain what I can to keep the code small and > > still be using th vsprintf. > > > > I'm using the gnuarm compiler. > > Functions like printf have a lot of code in them. I'd write my own function. > > Leon > > > > -- > Internal Virus Database is out-of-date. > Checked by AVG Anti-Virus. > Version: 7.0.300 / Virus Database: 265.7.1 - Release Date: 19/01/2005 > > > ------------------------------------------------------------------------ > *Yahoo! Groups Links* > > * To visit your group on the web, go to: > http://groups.yahoo.com/group/lpc2000/ > > * To unsubscribe from this group, send an email to: > lpc2000-unsubscribe@yahoogroups.com > <mailto:lpc2000-unsubscribe@yahoogroups.com?subject=Unsubscribe> > > * Your use of Yahoo! Groups is subject to the Yahoo! Terms of > Service <http://docs.yahoo.com/info/terms/>. > >
2005-01-28 by gene_klein2000
Leon, thanks for your answer. Can you point me in a direction for finding such routines which are useable for. Gene. --- In lpc2000@yahoogroups.com, "Leon Heller" <leon.heller@d...> wrote: > ----- Original Message ----- > From: "gene_klein2000" <gene_klein@h...> > To: <lpc2000@yahoogroups.com> > Sent: Friday, January 28, 2005 11:48 AM > Subject: [lpc2000] large code generation with vsprintf > > > > > > > > Hi, > > > > I hope someone will help me with the next problem. > > > > > > When I'm using vsprintf the code generated by th compiler > > grows with about 30kB. > > Can anyone explain what I can to keep the code small and > > still be using th vsprintf. > > > > I'm using the gnuarm compiler. > > Functions like printf have a lot of code in them. I'd write my own function.
> > Leon > > > > -- > Internal Virus Database is out-of-date. > Checked by AVG Anti-Virus. > Version: 7.0.300 / Virus Database: 265.7.1 - Release Date: 19/01/2005
2005-01-28 by Leon Heller
----- Original Message -----
From: "gene_klein2000" <gene_klein@...> To: <lpc2000@yahoogroups.com> Sent: Friday, January 28, 2005 1:15 PM Subject: [lpc2000] Re: large code generation with vsprintf > > > Leon, > > thanks for your answer. > > > Can you point me in a direction for finding such routines > which are useable for. I'd simply overwrite the string by finding its address in memory, and poking the characters directly into it, then placing the null byte in the appropriate place. Leon Leon -- Internal Virus Database is out-of-date. Checked by AVG Anti-Virus. Version: 7.0.300 / Virus Database: 265.7.1 - Release Date: 19/01/2005
2005-01-28 by 42Bastian Schick
> > When I'm using vsprintf the code generated by th compiler > grows with about 30kB. > Can anyone explain what I can to keep the code small and > still be using th vsprintf. > > I'm using the gnuarm compiler. give vsiprintf a chance, it does not have floating point support. Also newlib can be tuned regarding these things. -- 42Bastian Schick
2005-01-28 by tah2k
Hi Gene-
I ran into the same problem a few years ago when I was trying to use
sprintf in my boot block. It turns out that most of the code is
related to floating point conversions. I wrote my own 'tprintf' ('t'
for Tim) that does everything except floating point and it is very
tight and small. I can email it to you if youlike.
-Tim
--- In lpc2000@yahoogroups.com, "gene_klein2000" <gene_klein@h...>
wrote:> > Hi, > > I hope someone will help me with the next problem. > > > When I'm using vsprintf the code generated by th compiler > grows with about 30kB. > Can anyone explain what I can to keep the code small and > still be using th vsprintf. > > I'm using the gnuarm compiler. > > > Thanks, > > Gene
2005-01-28 by gene_klein2000
Hoi Tim, It would be great if you wanna mail it to me. THANKS, Gene. --- In lpc2000@yahoogroups.com, "tah2k" <tah2k@y...> wrote:
>
> Hi Gene-
>
> I ran into the same problem a few years ago when I was trying to use
> sprintf in my boot block. It turns out that most of the code is
> related to floating point conversions. I wrote my own 'tprintf' ('t'
> for Tim) that does everything except floating point and it is very
> tight and small. I can email it to you if youlike.
>
> -Tim
>
> --- In lpc2000@yahoogroups.com, "gene_klein2000" <gene_klein@h...>
> wrote:
> >
> > Hi,
> >
> > I hope someone will help me with the next problem.
> >
> >
> > When I'm using vsprintf the code generated by th compiler
> > grows with about 30kB.
> > Can anyone explain what I can to keep the code small and
> > still be using th vsprintf.
> >
> > I'm using the gnuarm compiler.
> >
> >
> > Thanks,
> >
> > Gene2005-01-28 by tah2k
I couldn't figure out how to get your email, so I will just attach
it here....Now that I pasted it, its a mess. If you can't resolve
the linewrap mess, just send me your email address.
/********************************************************************
**********
Filename: tprintf.h
Description: The following function is a limited version of
sprintf(). It is
useful for applications where memory
is limited or code size
is an issue.
*********************************************************************
**********/
#ifndef TPRINTF_H
#define TPRINTF_H
extern "C" void tprintf(INT8S* buf, const INT8S* format, ...);
#endif
/******************************* End of
file***********************************/
/********************************************************************
**********
Filename: tprintf.cpp
Description: The following function is a limited version of
sprintf(). It is
useful for applications where memory
is limited or code size
is an issue.
*********************************************************************
**********/
/////////////////////////////////////////////////////////////////////
//////////
// includes
#include <string.h>
#include <stdarg.h>
#include <stdlib.h>
#include "Typedefs.h"
/////////////////////////////////////////////////////////////////////
//////////
// defines
/////////////////////////////////////////////////////////////////////
//////////
// Static's
/********************************************************************
**********
Name: numToAscii()
Description: The following function converts an integer to its
ASCII
representation in a given numeric
base.
Input: char* buf: Preallocated storage to place the
ASCII representation
unsigned long num: The number to be
converted to ASCII
int base: The number base to
represent the number.
Output: None
*********************************************************************
**********/
extern "C" void numToAscii(INT8S* buf, INT32U num, INT32S base) {
static INT8S digits[] = "0123456789ABCDEF";
INT32U temp;
if(num == 0L){
// If the number is 0, just stuf a "0" into the
buffer
buf[0] = '0';
buf[1] = '\0';
}
else{
// Figure out the length necessary of the ASCII
string representing
// the number
for(temp = num; temp > 0; temp /= base)
buf++;
// Mark the spot that will be the end of the string
*buf = '\0';
// Generate ASCII representation
for(temp = num; temp > 0; temp /= base)
*--buf = digits[temp % base];
}
}
/********************************************************************
**********
Name: tprintf()
Description: The following function is a limited version of
sprintf(). It is
useful for applications where memory
is limited or code size
is an issue. The syntax implements
the sprintf() standard argument
passing. A conversion specification
looks like:
% {flags} {field_width}
{length_modifier} conversion
where flags=
'-': Left justify field
'0': Zero padding to
field width, for d, l, h, and X
where field_width=
Number of characters to be
printed in the field. Field width
will be padded with space if
needed. If given as '*', the
next argument should be an
integer holding the field width.
where length_modifier=
h: Used before d or X
conversions to denote a short int
l: Used before d or X
conversions to denote a long int
where conversion=
d: Write signed decimal
integer value
X: Write an unsigned
hexadecimal (0-9, ABC...) integer value
c: Write a single character
s: Write a string
Input: char* buf: Pointer to a preallocated buffer
to store formatted
output
const char* format: The printf style
format string reduced to
the above control fields
...: Variable length argument list
Output: None
*********************************************************************
**********/
extern "C" void tprintf(INT8S* buf, const INT8S* format, ...) {
// Points to each unnamed argument in turn
va_list argPtr;
// Value of the pad string
INT32S pad;
INT32S clip;
// Conversion of ASCII to long
INT32S longVal;
INT8S* src;
INT8S* dst;
// Character used to pad a field appropriately given a width
argument
INT8S padChar = ' ';
// Is the format requesting a zero pad?
BOOL reqZeroPad = FALSE;
// Make argPtr point to the first unnamed argument
va_start(argPtr, format);
while(*format) {
// Loop through the entire format string
if(*format != '%'){
// If not the start of a conversion
specification, copy the character
// literally
*buf++ = *format++;
}
else{
// Found the start of a conversion
specification, increment past the
// "%" and get the field width value
format++;
if(*format == '*'){
// A width specification, an 'int'
argument from the argument
// list supplies the value
pad = va_arg(argPtr, int);
// Increment past the '*'
format++;
}
else{
if(*format == '-' || (*format >= '0'
&& *format <= '9')){
// A width specification
controlling the minimum number of
// characters printed. The '-
' specifies left justification.
pad = atoi(format);
// A leading zero indicates
the pad character should be a
// '0' instead of a space
if(*format == '0')
reqZeroPad = TRUE;
do{
// Increment past
the width specification
format++;
}
while((*format >= '0')&&
(*format <= '9'));
}
else
// No width value was
specified
pad = 0;
}
// Handle the type
switch(*format++){
case 'c':
// Write a single character.
Note: `char's are always widened
// when they appear in
variable or unknown argument positions.
// Obtain a widened char and
narrow it.
buf[0] = (char)(va_arg
(argPtr, INT32U));
buf[1] = '\0';
break;
case 's':
// Write a string
src = va_arg(argPtr, char*);
strcpy(buf, src);
break;
case 'l':
// Write a signed decimal
long integer value
// Skip the 'd' in '%ld'
format++;
// Pop the argument off teh
list
longVal = va_arg(argPtr,
long);
dst = buf;
if(longVal<0){
// Negative number
*dst++ = '-';
longVal = -longVal;
}
// Convert number to ASCII
representation
numToAscii(dst, longVal, 10);
// Determine pad character
if(reqZeroPad)
padChar = '0';
break;
case 'h':
// Write a signed short
integer value
// Skip the 'd' in '%ld'
format++;
// Pop the argument off teh
list
longVal = va_arg(argPtr,
INT32U);
dst = buf;
if(longVal<0){
// Negative number
*dst++ = '-';
longVal = -longVal;
}
// Convert number to ASCII
representation
numToAscii(dst, longVal, 10);
// Determine pad character
if(reqZeroPad)
padChar = '0';
break;
case 'X':
// Write an unsigned
hexidecimal (0-9, ABC...) integer value
longVal = (long)va_arg
(argPtr, int);
// Convert number to ASCII
representation
numToAscii(buf, longVal, 16);
// Determine pad character
if(reqZeroPad)
padChar = '0';
break;
case 'd':
// Write a signed decimal
integer value
longVal = (long)va_arg
(argPtr, int);
dst = buf;
if(longVal<0){
// Negative number
*dst++ = '-';
longVal = -longVal;
}
// Convert number to ASCII
representation
numToAscii(dst, longVal, 10);
break;
default:
buf[0] = format[-1];
buf[1] = '\0';
break;
};
// Fix the padding if the value is too short
clip = strlen(buf);
if(pad < 0){
// Add spaces after the value
pad = -pad - clip;
for(buf += clip; pad > 0; pad++)
*buf++ = padChar;
buf = '\0';
}
else{
// Add spaces before the value
pad -= clip;
if(pad > 0){
// Width specification is
greater than string length, therefore
// pad appropriatly
// Point to the end of the
destination string
src = buf + clip;
// Point to the end of the
unpadded destination string
dst = src + pad;
// Terminate this field
*dst = '\0';
while(src > buf)
// Right justify the
string within the width specification
*--dst = *--src;
while(dst > buf)
// Fill with spaces
to the left
*--dst = padChar;
}
buf += strlen(buf);
}
}
}
// Mark the end of the output strin
*buf = '\0';
// Clean up the variable argument list
va_end(argPtr);
}
/******************************* End of
file***********************************/
--- In lpc2000@yahoogroups.com, "gene_klein2000" <gene_klein@h...>
wrote:
>
> Hoi Tim,
>
> It would be great if you wanna mail it to me.
>
> THANKS,
>
> Gene.
>
> --- In lpc2000@yahoogroups.com, "tah2k" <tah2k@y...> wrote:
> >
> > Hi Gene-
> >
> > I ran into the same problem a few years ago when I was trying to
use
> > sprintf in my boot block. It turns out that most of the code is
> > related to floating point conversions. I wrote my own 'tprintf'
('t'
> > for Tim) that does everything except floating point and it is
very
> > tight and small. I can email it to you if youlike.
> >
> > -Tim
> >
> > --- In lpc2000@yahoogroups.com, "gene_klein2000"
<gene_klein@h...> > > wrote: > > > > > > Hi, > > > > > > I hope someone will help me with the next problem. > > > > > > > > > When I'm using vsprintf the code generated by th compiler > > > grows with about 30kB. > > > Can anyone explain what I can to keep the code small and > > > still be using th vsprintf. > > > > > > I'm using the gnuarm compiler. > > > > > > > > > Thanks, > > > > > > Gene
2005-01-28 by Curt Powell
Does anyone know if there is a yahoo group (or equivalent) for the STR71X mcus?
2005-01-28 by Robert Adsett
At 10:14 AM 1/28/05 -0800, you wrote:
>Does anyone know if there is a yahoo group (or equivalent) for the STR71X
>mcus?
http://groups.yahoo.com/group/str7
It's pretty close to dead at the moment though.
Robert
" 'Freedom' has no meaning of itself. There are always restrictions,
be they legal, genetic, or physical. If you don't believe me, try to
chew a radio signal. "
Kelvin Throop, III