is there a way to have the strNPrintF example to get a length of the args
added into the formated string before writing it? i ask this because when i
run the example in pose it complains that the buffer is to small and writes
to data structures or something (some error). or is it possible to have it
start writing to the buffer but stop at the specified length?
heres what i have so far....
/***********************************************************************/
Boolean StrNPrintF(CharPtr s, UInt size, CharPtr formatStr, ...) {
// Char text[size];
Int i, j, k, l, m;
Boolean failed;
Char *text;
va_list args;
va_start(args, formatStr);
i=sizeof(formatStr); ///
j=sizeof(args); ///
k=StrLen(formatStr);
l=StrLen(args);
text=MemPtrNew(size);
StrVPrintF(text, formatStr, args);
m=StrLen(text);
va_end(args);
StrNCopy(s, text, size);
// if(i>size) {
// FrmCustomAlert(TestAlert, "Too long for variable", (Char*)"test", NULL);
// failed=true;
// }else{
failed=false;
// }
MemPtrFree(text);
return(failed);
}
pardon the extra redundant code, i've been trying a few differant things to
try to get what i need
-----Original Message-----
From: Chris Antos <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Thursday, January 06, 2000 1:17 PM
Subject: Re: StrPrintF
>true, although that's a very inefficient way to implement it, with respect
>to CPU cycles (and battery usage).
>
>i've written a generic formatting engine before, on the Win32 platform. it
>is extendable, can work against fixed length buffers, or dynamic length
>buffers (via a callback), and is localization friendly (eg, "{s0}{s1}"
>instead of "%s%s" - this means if some language needs the two strings
>swapped, you don't have to change any code to do so, since the format tag
>includes the index of the argument). eventually i'll work on porting it to
>the Palm platform (eventually i'll need it for my own freeware app anyway).
>OTOH, it did require that all arguments were 32-bit, and used C++ in
>validate the supported types (without adding any code bytes in the
>process!).
>
>sorry, i can't give any timeframe on it, though.
>
>
>----- Original Message -----
>From: "William F. Weiher III" <[EMAIL PROTECTED]>
>To: <[EMAIL PROTECTED]>
>Sent: Thursday, January 06, 2000 7:58 AM
>Subject: RE: StrPrintF
>
>
>> If StrPrintF would accept a NULL destination but still return the correct
>> count of characters generated, this would be easy to implement.
>> Unfutunately, the current code does a StrLen on the string generated to
>get
>> the length). Just a suggestion for development
>>
>> -----Original Message-----
>> From: Mark Nudelman [mailto:[EMAIL PROTECTED]]
>> Sent: Wednesday, January 05, 2000 14:56
>> To: [EMAIL PROTECTED]
>> Subject: Re: StrPrintF
>>
>> Um, but this code doesn't do anything. Since "text" is allocated as the
>> same size as the input buffer (s), this doesn't provide any protection.
>> It just does an unnecessary StrNCopy. You want to call MemPtrNew with a
>> LARGER size, a size that's large enough to handle any formatStr.
>> Unfortunately, there is no such size, so the best you can do is estimate
>> the largest size your app will need and use that. It's unfortunate that
>> Palm didn't provide a printf-like variant that calls a callback function
>> to output each char into the buffer. Then you could do a real
>> StrNPrintF without an extra buffer. Of course, you could write your own
>> from scratch by duplicating all the logic in StrNPrintF.
>> --Mark
>>
>>
>> more - deleted
>>
>>
>
>
