Extended attrstr.c. Much more complete API now.
This commit is contained in:
parent
4f8f94b85a
commit
b45e5f4de2
150
common/attrstr.c
150
common/attrstr.c
|
@ -39,6 +39,7 @@ uiAttributedString *uiNewAttributedString(const char *initialString)
|
||||||
return s;
|
return s;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
// TODO make sure that all implementations of graphemes() work fine with empty strings; in particular, the Windows one might not
|
||||||
static void recomputeGraphemes(uiAttributedString *s)
|
static void recomputeGraphemes(uiAttributedString *s)
|
||||||
{
|
{
|
||||||
if (s->graphemes != NULL)
|
if (s->graphemes != NULL)
|
||||||
|
@ -75,31 +76,95 @@ const char *uiAttributedStringString(uiAttributedString *s)
|
||||||
return s->s;
|
return s->s;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
size_t uiAttributedStringLen(uiAttributedString *s)
|
||||||
|
{
|
||||||
|
return s->len;
|
||||||
|
}
|
||||||
|
|
||||||
|
static void u8u16len(const char *str, size_t *n8, size_t *n16)
|
||||||
|
{
|
||||||
|
uint32_t rune;
|
||||||
|
char buf[4];
|
||||||
|
uint16_t buf16[2];
|
||||||
|
|
||||||
|
*n8 = 0;
|
||||||
|
*n16 = 0;
|
||||||
|
while (*str) {
|
||||||
|
str = utf8DecodeRune(str, 0, &rune);
|
||||||
|
*n8 += utf8EncodeRune(rune, buf);
|
||||||
|
*n16 += utf16EncodeRune(rune, buf16);
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
void uiAttributedStringAppendUnattributed(uiAttributedString *s, const char *str)
|
void uiAttributedStringAppendUnattributed(uiAttributedString *s, const char *str)
|
||||||
{
|
{
|
||||||
const char *t;
|
uiAttributedStringInsertAtUnattributed(s, str, s->len);
|
||||||
|
}
|
||||||
|
|
||||||
|
// this works (and returns true, which is what we want) at s->len too because s->s[s->len] is always going to be 0 due to us allocating s->len + 1 bytes and because uiRealloc() always zero-fills allocated memory
|
||||||
|
static int onCodepointBoundary(uiAttributedString *s, size_t at)
|
||||||
|
{
|
||||||
|
uint8_t c;
|
||||||
|
|
||||||
|
// for uiNewAttributedString()
|
||||||
|
if (s->s == NULL && at == 0)
|
||||||
|
return 1;
|
||||||
|
c = (uint8_t) (s->s[at]);
|
||||||
|
return c < 0x80 || c >= 0xC0;
|
||||||
|
}
|
||||||
|
|
||||||
|
// TODO note that at must be on a codeoint boundary
|
||||||
|
void uiAttributedStringInsertAtUnattributed(uiAttributedString *s, const char *str, size_t at)
|
||||||
|
{
|
||||||
uint32_t rune;
|
uint32_t rune;
|
||||||
char buf[4];
|
char buf[4];
|
||||||
uint16_t u16buf[2];
|
uint16_t u16buf[2];
|
||||||
size_t n, n16;
|
size_t n, n16;
|
||||||
size_t old, old;
|
size_t old, old16;
|
||||||
|
size_t oldlen, old16len;
|
||||||
|
size_t at16;
|
||||||
|
size_t i;
|
||||||
|
|
||||||
|
if (!onCodepointBoundary(s, at)) {
|
||||||
|
// TODO
|
||||||
|
}
|
||||||
|
|
||||||
|
at16 = s->u8tou16[at];
|
||||||
|
|
||||||
|
// do this first to reclaim memory
|
||||||
|
invalidateGraphemes(s);
|
||||||
|
|
||||||
// first figure out how much we need to grow by
|
// first figure out how much we need to grow by
|
||||||
// this includes post-validated UTF-8
|
// this includes post-validated UTF-8
|
||||||
t = str;
|
u8u16len(str, &n, &n16);
|
||||||
n = 0;
|
|
||||||
n16 = 0;
|
|
||||||
while (*t) {
|
|
||||||
t = utf8DecodeRune(t, 0, &rune);
|
|
||||||
n += utf8EncodeRune(rune, buf);
|
|
||||||
n16 += utf16EncodeRune(rune, buf16);
|
|
||||||
}
|
|
||||||
|
|
||||||
// and resize
|
// and resize
|
||||||
old = s->len;
|
old = at;
|
||||||
old16 = s->len16;
|
old16 = at16;
|
||||||
|
oldlen = s->len;
|
||||||
|
old16len = s->u16len;
|
||||||
resize(s, s->len + n, s->u16len + n16);
|
resize(s, s->len + n, s->u16len + n16);
|
||||||
|
|
||||||
|
// move existing characters out of the way
|
||||||
|
// note the use of memmove(): https://twitter.com/rob_pike/status/737797688217894912
|
||||||
|
memmove(
|
||||||
|
s->s + at + n8,
|
||||||
|
s->s + at,
|
||||||
|
(oldlen - at) * sizeof (char));
|
||||||
|
memmove(
|
||||||
|
s->u16 + at16 + n16,
|
||||||
|
s->u16 + at16,
|
||||||
|
(old16len - at16) * sizeof (uint16_t));
|
||||||
|
// note the + 1 for these; we want to copy the terminating null too
|
||||||
|
memmove(
|
||||||
|
s->u8tou16 + at + n8,
|
||||||
|
s->u8tou16 + at,
|
||||||
|
(oldlen - at + 1) * sizeof (size_t));
|
||||||
|
memmove(
|
||||||
|
s->u16tou8 + at16 + n16,
|
||||||
|
s->u16tou8 + at16,
|
||||||
|
(old16len - at16 + 1) * sizeof (size_t));
|
||||||
|
|
||||||
// and copy
|
// and copy
|
||||||
while (*str) {
|
while (*str) {
|
||||||
str = utf8DecodeRune(str, 0, &rune);
|
str = utf8DecodeRune(str, 0, &rune);
|
||||||
|
@ -129,10 +194,71 @@ void uiAttributedStringAppendUnattributed(uiAttributedString *s, const char *str
|
||||||
old16 += n16;
|
old16 += n16;
|
||||||
}
|
}
|
||||||
// and have an index for the end of the string
|
// and have an index for the end of the string
|
||||||
|
// TODO is this done by the below?
|
||||||
s->u8tou16[old] = old16;
|
s->u8tou16[old] = old16;
|
||||||
s->u16tou8[old16] = old;
|
s->u16tou8[old16] = old;
|
||||||
|
|
||||||
|
// and finally adjust the prior values in the conversion tables
|
||||||
|
// use <= so the terminating 0 gets updated too
|
||||||
|
for (i = 0; i <= oldlen - at; i++)
|
||||||
|
s->u8tou16[at + n + i] += n16;
|
||||||
|
for (i = 0; i <= old16len - at16; i++)
|
||||||
|
s->u16tou8[at16 + n16 + i] += n;
|
||||||
|
}
|
||||||
|
|
||||||
|
// TODO document that end is the first index that will be maintained
|
||||||
|
void uiAttributedStringDelete(uiAttributedString *s, size_t start, size_t end)
|
||||||
|
{
|
||||||
|
size_t start16, end16;
|
||||||
|
size_t count, count16;
|
||||||
|
size_t i;
|
||||||
|
|
||||||
|
if (!onCodepointBoundary(s, start)) {
|
||||||
|
// TODO
|
||||||
|
}
|
||||||
|
if (!onCodepointBoundary(s, end)) {
|
||||||
|
// TODO
|
||||||
|
}
|
||||||
|
|
||||||
|
count = end - start;
|
||||||
|
start16 = s->u8tou16[start];
|
||||||
|
end16 = s->u8tou16[end];
|
||||||
|
count16 = end16 - start16;
|
||||||
|
|
||||||
invalidateGraphemes(s);
|
invalidateGraphemes(s);
|
||||||
|
|
||||||
|
// overwrite old characters
|
||||||
|
memmove(
|
||||||
|
s->s + start,
|
||||||
|
s->s + end,
|
||||||
|
(oldlen - end) * sizeof (char));
|
||||||
|
memmove(
|
||||||
|
s->u16 + start16,
|
||||||
|
s->u16 + end16,
|
||||||
|
(old16len - end16) * sizeof (uint16_t));
|
||||||
|
// note the + 1 for these; we want to copy the terminating null too
|
||||||
|
memmove(
|
||||||
|
s->u8tou16 + start,
|
||||||
|
s->u8tou16 + end,
|
||||||
|
(oldlen - end + 1) * sizeof (size_t));
|
||||||
|
memmove(
|
||||||
|
s->u16tou8 + start16,
|
||||||
|
s->u16tou8 + end16,
|
||||||
|
(old16len - end16 + 1) * sizeof (size_t));
|
||||||
|
|
||||||
|
// update the conversion tables
|
||||||
|
// note the use of <= to include the null terminator
|
||||||
|
for (i = 0; i <= count; i++)
|
||||||
|
s->u8tou16[start + i] -= count16;
|
||||||
|
for (i = 0; i <= count16; i++)
|
||||||
|
s->u16tou8[start16 + i] -= count;
|
||||||
|
|
||||||
|
// null-terminate the string
|
||||||
|
s->s[start + count] = 0;
|
||||||
|
s->u16[start16 + count16] = 0;
|
||||||
|
|
||||||
|
// and finally resize
|
||||||
|
resize(s, start + count, start16 + count16);
|
||||||
}
|
}
|
||||||
|
|
||||||
// TODO figure out if we should count the grapheme past the end
|
// TODO figure out if we should count the grapheme past the end
|
||||||
|
|
Loading…
Reference in New Issue