gnustep/libs-base
0
0
Fork 0
mirror of https://github.com/gnustep/libs-base.git synced 2025-04-22 16:33:29 +00:00
Code Issues Projects Releases Packages Wiki Activity

Add some optimisation for converting from unicode to latin1 or ascii ... move

Browse source 
as much as possible outside the loop iterating over the characters.


git-svn-id: svn+ssh://svn.gna.org/svn/gnustep/libs/base/trunk@28332 72102866-910b-0410-8b05-ffd578937521
This commit is contained in:
Richard Frith-MacDonald 2009-06-06 15:52:11 +00:00
parent 9c86cf78f2
commit 39839cb227
2 changed files with 104 additions and 31 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
ChangeLog
View file

@ -1,3 +1,8 @@
2009-06-06 Richard Frith-Macdonald <rfm@gnu.org>
* Source/Additions/Unicode.m: Optimise case where we are converting
from unicode to latin1 or ascii.
2009-06-06 Wolfgang Lux <wolfgang.lux@gmail.com>
* Source/GSFFIInvocation.m

130
Source/Additions/Unicode.m
View file

@ -1922,53 +1922,121 @@ GSFromUnicode(unsigned char **dst, unsigned int *size, const unichar *src,
goto bases;
bases:
if (strict == NO)
if (dst == 0)
{
while (spos < slen)
/* Just counting bytes, and we know there is exactly one
* unicode codepoint needed for each character.
*/
dpos = slen;
}
else
{
/* Because we know that each ascii chartacter is exactly
* one unicode character, we can check the destination
* buffer size and allocate more space in one go, before
* entering the loop where we deal with each character.
*/
if (slen > bsize)
{
unichar u = src[spos++];
if (swapped == YES)
if (zone == 0)
{
u = (((u & 0xff00) >> 8) + ((u & 0x00ff) << 8));
}
if (dpos >= bsize)
{
GROW();
}
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
result = NO; /* No buffer growth possible ... fail. */
goto done;
}
else
{
ptr[dpos++] = '?';
uint8_t *tmp;
#if GS_WITH_GC
tmp = NSAllocateCollectable(slen, 0);
#else
tmp = NSZoneMalloc(zone, slen);
if (ptr != buf)
{
NSZoneFree(zone, ptr);
}
#endif
ptr = tmp;
if (ptr == 0)
{
result = NO; /* Not enough memory */
break;
}
bsize = slen;
}
}
}
if (strict == NO)
{
if (swapped == YES)
{
while (spos < slen)
{
unichar u = src[spos++];
u = (((u & 0xff00) >> 8) + ((u & 0x00ff) << 8));
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
}
else
{
ptr[dpos++] = '?';
}
}
}
else
{
while (spos < slen)
{
unichar u = src[spos++];
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
}
else
{
ptr[dpos++] = '?';
}
}
}
}
else
{
while (spos < slen)
if (swapped == YES)
{
unichar u = src[spos++];
while (spos < slen)
{
unichar u = src[spos++];
if (swapped == YES)
{
u = (((u & 0xff00) >> 8) + ((u & 0x00ff) << 8));
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
}
else
{
result = NO;
goto done;
}
}
if (dpos >= bsize)
}
else
{
while (spos < slen)
{
GROW();
}
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
}
else
{
result = NO;
goto done;
unichar u = src[spos++];
if (u < base)
{
ptr[dpos++] = (unsigned char)u;
}
else
{
result = NO;
goto done;
}
}
}
}