• k.inaba (17/18) Jan 05 2006 but decode("%41%42%43") returns "ABC". ('A', 'B', and 'C' seems to be va...
• Walter Bright (3/11) Jan 08 2006 I think the doc is wrong here. I'll fix it.

k.inaba <k.inaba_member pathlink.com> writes:

1.
The document comment for the std.uri.decode() function states
 Escape sequences that resolve to valid URI characters are not replaced.

but decode("%41%42%43") returns "ABC". ('A', 'B', and 'C' seems to be valid URI
characters, but replaced.) Maybe this behaviour is intended, and the document
should be corrected?

2.
The std.uri module internally allocates temporary buffers by alloca() on stack.
But this causes a problem when one deals with large strings:

import std.uri;
void main() {
char[] str = new char[10000000]; // about 10MB
str[] = 'A';
encodeComponent(str);
}

This example halts on "Error: Stack Overflow" on my WinXP+dmd 1.42 environment.
All allocas in this module should be rewritten to heap allocation, i think. Or
at least, this limitation should be noted in the document.

"Walter Bright" <newshound digitalmars.com> writes:

"k.inaba" <k.inaba_member pathlink.com> wrote in message 
news:dpjmhp$2hcp$1 digitaldaemon.com...
 1.
 The document comment for the std.uri.decode() function states
 Escape sequences that resolve to valid URI characters are not replaced.

 but decode("%41%42%43") returns "ABC". ('A', 'B', and 'C' seems to be 
 valid URI
 characters, but replaced.) Maybe this behaviour is intended, and the 
 document
 should be corrected?

I think the doc is wrong here. I'll fix it.