/* Convert Unicode/PdfDocEncoding string into utf-8 */
char *
pdf_to_utf8(fz_obj *src)
{
unsigned char *srcptr = (unsigned char *) fz_to_str_buf(src);
char *dstptr, *dst;
int srclen = fz_to_str_len(src);
int dstlen = 0;
int ucs;
int i;
if (srclen >= 2 && srcptr[0] == 254 && srcptr[1] == 255)
{
for (i = 2; i + 1 < srclen; i += 2)
{
ucs = srcptr[i] << 8 | srcptr[i+1];
dstlen += runelen(ucs);
}
dstptr = dst = fz_malloc(dstlen + 1);
for (i = 2; i + 1 < srclen; i += 2)
{
ucs = srcptr[i] << 8 | srcptr[i+1];
dstptr += runetochar(dstptr, &ucs);
}
}
else if (srclen >= 2 && srcptr[0] == 255 && srcptr[1] == 254)
{
for (i = 2; i + 1 < srclen; i += 2)
{
ucs = srcptr[i] | srcptr[i+1] << 8;
dstlen += runelen(ucs);
}
dstptr = dst = fz_malloc(dstlen + 1);
for (i = 2; i + 1 < srclen; i += 2)
{
ucs = srcptr[i] | srcptr[i+1] << 8;
dstptr += runetochar(dstptr, &ucs);
}
}
else
{
for (i = 0; i < srclen; i++)
dstlen += runelen(pdf_doc_encoding[srcptr[i]]);
dstptr = dst = fz_malloc(dstlen + 1);
for (i = 0; i < srclen; i++)
{
ucs = pdf_doc_encoding[srcptr[i]];
dstptr += runetochar(dstptr, &ucs);
}
}
*dstptr = '\0';
return dst;
}
fz_error *
pdf_toutf8(char **dstp, fz_obj *src)
{
unsigned char *srcptr = fz_tostrbuf(src);
char *dstptr;
int srclen = fz_tostrlen(src);
int dstlen = 0;
int ucs;
int i;
if (srclen > 2 && srcptr[0] == 254 && srcptr[1] == 255)
{
for (i = 2; i < srclen; i += 2)
{
ucs = (srcptr[i] << 8) | srcptr[i+1];
dstlen += runelen(ucs);
}
dstptr = *dstp = fz_malloc(dstlen + 1);
if (!dstptr)
return fz_outofmem;
for (i = 2; i < srclen; i += 2)
{
ucs = (srcptr[i] << 8) | srcptr[i+1];
dstptr += runetochar(dstptr, &ucs);
}
}
else
{
for (i = 0; i < srclen; i++)
dstlen += runelen(pdf_docencoding[srcptr[i]]);
dstptr = *dstp = fz_malloc(dstlen + 1);
if (!dstptr)
return fz_outofmem;
for (i = 0; i < srclen; i++)
{
ucs = pdf_docencoding[srcptr[i]];
dstptr += runetochar(dstptr, &ucs);
}
}
*dstptr = '\0';
return nil;
}
char *
smbstringdup(SmbHeader *h, uchar *base, uchar **bdatap, uchar *edata)
{
char *p;
if (h && h->flags2 & SMB_FLAGS2_UNICODE) {
uchar *bdata = *bdatap;
uchar *savebdata;
Rune r;
int l;
char *q;
l = 0;
if ((bdata - base) & 1)
bdata++;
savebdata = bdata;
do {
if (bdata + 2 > edata)
return nil;
r = smbnhgets(bdata); bdata += 2;
l += runelen(r);
} while (r != 0);
p = smbemalloc(l);
bdata = savebdata;
q = p;
do {
r = smbnhgets(bdata); bdata += 2;
q += runetochar(q, &r);
} while (r != 0);
*bdatap = bdata;
return p;
}
return smbstrdup(bdatap, edata);
}
static void textpush(js_State *J, Rune c)
{
int n = runelen(c);
if (J->lexbuf.len + n > J->lexbuf.cap) {
J->lexbuf.cap = J->lexbuf.cap * 2;
J->lexbuf.text = js_realloc(J, J->lexbuf.text, J->lexbuf.cap);
}
J->lexbuf.len += runetochar(J->lexbuf.text + J->lexbuf.len, &c);
}
int
widebytes(wchar_t *ws)
{
int n = 0;
while (*ws)
n += runelen(*ws++);
return n+1;
}
int
wstrutflen(Rune *s)
{
int n;
for(n=0; *s; n+=runelen(*s),s++)
;
return n;
}
size_t
runenlen(const Rune *p, size_t len)
{
size_t i, n = 0;
for(i = 0; i < len; i++)
n += runelen(p[i]);
return n;
}
void
rinsert(Bufblock *buf, Rune r)
{
int n;
n = runelen(r);
if (buf->current+n > buf->end)
growbuf(buf);
runetochar(buf->current, &r);
buf->current += n;
}
int
smbbuffergetucs2(SmbBuffer *b, ulong flags, char **sp)
{
uchar *bdata = b->buf + b->rn;
uchar *edata = b->buf + b->wn;
Rune r;
int l;
char *p, *q;
uchar *savebdata;
int first;
l = 0;
if ((flags & SMB_STRING_UNALIGNED) == 0 && (bdata - b->buf) & 1)
bdata++;
savebdata = bdata;
first = 1;
do {
if (bdata + 2 > edata) {
l++;
break;
}
r = smbnhgets(bdata); bdata += 2;
if (first && (flags & SMB_STRING_PATH) && r != '\\')
l++;
first = 0;
if (flags & SMB_STRING_CONVERT_MASK)
r = smbruneconvert(r, flags);
l += runelen(r);
} while (r != 0);
p = smbemalloc(l);
bdata = savebdata;
q = p;
first = 1;
do {
if (bdata + 2 > edata) {
*q = 0;
break;
}
r = smbnhgets(bdata); bdata += 2;
if (first && (flags & SMB_STRING_PATH) && r != '\\')
*q++ = '/';
first = 0;
if (flags & SMB_STRING_CONVERT_MASK)
r = smbruneconvert(r, flags);
q += runetochar(q, &r);
} while (r != 0);
b->rn = bdata - b->buf;
*sp = p;
return 1;
}
int
charntorune(Rune *p, const char *s, size_t len)
{
unsigned int n, i = 1;
Rune r;
if(len == 0) /* can't even look at s[0] */
return 0;
r = (unsigned char)s[0];
n = lookup[r/2];
if(n == 1)
goto done;
if(n == 0) {
r = Runeerror;
goto done;
}
r &= 0xFF >> n;
if(len > n)
len = n;
/* add values from continuation bytes */
for(; i < len; i++) {
if((s[i] & 0xC0) != 0x80) { /* not a continuation byte */
r = Runeerror;
goto done;
}
/* add bits from continuation byte to rune value
* cannot overflow: 6 byte sequences contain 31 bits */
r = (r << 6) | (s[i] & 0x3F); /* 10xxxxxx */
}
if(i < n) /* must have reached len limit */
return 0;
/* reject invalid or overlong sequences */
if(runelen(r) < (int)n) {
r = Runeerror;
goto done;
}
done:
*p = r;
return i;
}
static void
pchar(Rune c, Fconv *fp)
{
int n;
n = fp->eout - fp->out;
if(n > 0) {
if(c < Runeself) {
*fp->out++ = c;
return;
}
if(n >= UTFmax || n >= runelen(c)) {
n = runetochar(fp->out, &c);
fp->out += n;
return;
}
fp->eout = fp->out;
}
}
int
wstrtoutf(char *s, Rune *t, int n)
{
int i;
char *s0;
s0 = s;
if(n <= 0)
return wstrutflen(t)+1;
while(*t) {
if(n < UTFmax+1 && n < runelen(*t)+1) {
*s = 0;
return s-s0+wstrutflen(t)+1;
}
i = runetochar(s, t);
s += i;
n -= i;
t++;
}
*s = 0;
return s-s0;
}
char*
runes16toutf(char *p, Rune16 *r, int nc)
{
char *op, *ep;
int n, c;
Rune rc;
op = p;
ep = p + nc;
while(c = *r++) {
n = 1;
if(c >= Runeself)
n = runelen(c);
if(p + n >= ep)
break;
rc = c;
if(c < Runeself)
*p++ = c;
else
p += runetochar(p, &rc);
}
*p = '\0';
return op;
}
/*
* How many bytes of output UTF will be produced by quoting (if necessary) this string?
* How many runes? How much of the input will be consumed?
* The parameter q is filled in by __quotesetup.
* The string may be UTF or Runes (s or r).
* Return count does not include NUL.
* Terminate the scan at the first of:
* NUL in input
* count exceeded in input
* count exceeded on output
* *ninp is set to number of input bytes accepted.
* nin may be <0 initially, to avoid checking input by count.
*/
void
__quotesetup(char *s, Rune *r, int nin, int nout, Quoteinfo *q, int sharp, int runesout)
{
int w;
Rune c;
q->quoted = 0;
q->nbytesout = 0;
q->nrunesout = 0;
q->nbytesin = 0;
q->nrunesin = 0;
if(sharp || nin==0 || (s && *s=='\0') || (r && *r=='\0')){
if(nout < 2)
return;
q->quoted = 1;
q->nbytesout = 2;
q->nrunesout = 2;
}
for(; nin!=0; nin--){
if(s)
w = chartorune(&c, s);
else{
c = *r;
w = runelen(c);
}
if(c == '\0')
break;
if(runesout){
if(q->nrunesout+1 > nout)
break;
}else{
if(q->nbytesout+w > nout)
break;
}
if((c <= L' ') || (c == L'\'') || (fmtdoquote!=nil && fmtdoquote((int)c))){
if(!q->quoted){
if(runesout){
if(1+q->nrunesout+1+1 > nout) /* no room for quotes */
break;
}else{
if(1+q->nbytesout+w+1 > nout) /* no room for quotes */
break;
}
q->nrunesout += 2; /* include quotes */
q->nbytesout += 2; /* include quotes */
q->quoted = 1;
}
if(c == '\'') {
if(runesout){
if(1+q->nrunesout+1 > nout) /* no room for quotes */
break;
}else{
if(1+q->nbytesout+w > nout) /* no room for quotes */
break;
}
q->nbytesout++;
q->nrunesout++; /* quotes reproduce as two characters */
}
}
/* advance input */
if(s)
s += w;
else
r++;
q->nbytesin += w;
q->nrunesin++;
/* advance output */
q->nbytesout += w;
q->nrunesout++;
#ifndef PLAN9PORT
/* ANSI requires precision in bytes, not Runes. */
nin-= w-1; /* and then n-- in the loop */
#endif
}
}
char *
tcs(char *cs, char *s, long *np)
{
Channel *sync;
Exec *e;
Rune r;
long i, n;
void **a;
uchar *us;
char buf[BUFSIZE], cmd[50];
char *t, *u;
int p[2], q[2];
if(s==nil || *s=='\0' || *np==0){
werrstr("tcs failed: no data");
return s;
}
if(cs == nil){
werrstr("tcs failed: no charset");
return s;
}
if(cistrncmp(cs, "utf-8", 5)==0 || cistrncmp(cs, "utf8", 4)==0)
return s;
for(i=0; tcstab[i].mime!=nil; i++)
if(cistrncmp(cs, tcstab[i].mime, strlen(tcstab[i].mime)) == 0)
break;
if(tcstab[i].mime == nil){
fprint(2, "abaco: charset: %s not supported\n", cs);
goto latin1;
}
if(cistrcmp(tcstab[i].tcs, "8859-1")==0 || cistrcmp(tcstab[i].tcs, "ascii")==0){
latin1:
n = 0;
for(us=(uchar*)s; *us; us++)
n += runelen(*us);
n++;
t = emalloc(n);
for(us=(uchar*)s, u=t; *us; us++){
if(*us>=Winstart && *us<=Winend)
*u++ = winchars[*us-Winstart];
else{
r = *us;
u += runetochar(u, &r);
}
}
*u = 0;
free(s);
return t;
}
if(pipe(p)<0 || pipe(q)<0)
error("can't create pipe");
sync = chancreate(sizeof(ulong), 0);
if(sync == nil)
error("can't create channel");
snprint(cmd, sizeof cmd, "tcs -f %s", tcstab[i].tcs);
e = emalloc(sizeof(Exec));
e->p[0] = p[0];
e->p[1] = p[1];
e->q[0] = q[0];
e->q[1] = q[1];
e->cmd = cmd;
e->sync = sync;
proccreate(execproc, e, STACK);
recvul(sync);
chanfree(sync);
close(p[0]);
close(q[1]);
/* in case tcs fails */
t = s;
sync = chancreate(sizeof(ulong), 0);
if(sync == nil)
error("can't create channel");
a = emalloc(4*sizeof(void *));
a[0] = sync;
a[1] = (void *)p[1];
a[2] = s;
a[3] = (void *)*np;
proccreate(writeproc, a, STACK);
s = nil;
while((n = read(q[0], buf, sizeof(buf))) > 0){
s = erealloc(s, i+n+1);
memmove(s+i, buf, n);
i += n;
s[i] = '\0';
}
n = recvul(sync);
if(n != *np)
fprint(2, "tcs: did not write %ld; wrote %uld\n", *np, n);
*np = i;
chanfree(sync);
close(q[0]);
if(s == nil){
fprint(2, "tcs failed: can't convert charset=%s to %s\n", cs, tcstab[i].tcs);
return t;
}
free(t);
return s;
}
int
eenter(char *ask, char *buf, int len, Mouse *m)
{
int done, down, tick, n, h, w, l, i;
Image *b, *save, *backcol, *bordcol;
Point p, o, t;
Rectangle r, sc;
Event ev;
Rune k;
o = screen->r.min;
backcol = allocimagemix(display, DPurpleblue, DWhite);
bordcol = allocimage(display, Rect(0,0,1,1), screen->chan, 1, DPurpleblue);
if(backcol == nil || bordcol == nil)
return -1;
while(ecankbd())
ekbd();
if(m) o = m->xy;
if(buf && len > 0)
n = strlen(buf);
else {
buf = nil;
len = 0;
n = 0;
}
k = -1;
tick = n;
save = nil;
done = down = 0;
p = stringsize(font, " ");
h = p.y;
w = p.x;
b = screen;
sc = b->clipr;
replclipr(b, 0, b->r);
while(!done){
p = stringsize(font, buf ? buf : "");
if(ask && ask[0]){
if(buf) p.x += w;
p.x += stringwidth(font, ask);
}
r = rectaddpt(insetrect(Rpt(ZP, p), -4), o);
p.x = 0;
r = rectsubpt(r, p);
p = ZP;
if(r.min.x < screen->r.min.x)
p.x = screen->r.min.x - r.min.x;
if(r.min.y < screen->r.min.y)
p.y = screen->r.min.y - r.min.y;
r = rectaddpt(r, p);
p = ZP;
if(r.max.x > screen->r.max.x)
p.x = r.max.x - screen->r.max.x;
if(r.max.y > screen->r.max.y)
p.y = r.max.y - screen->r.max.y;
r = rectsubpt(r, p);
r = insetrect(r, -2);
if(save == nil){
save = allocimage(display, r, b->chan, 0, DNofill);
if(save == nil){
n = -1;
break;
}
draw(save, r, b, nil, r.min);
}
draw(b, r, backcol, nil, ZP);
border(b, r, 2, bordcol, ZP);
p = addpt(r.min, Pt(6, 6));
if(ask && ask[0]){
p = string(b, p, bordcol, ZP, font, ask);
if(buf) p.x += w;
}
if(buf){
t = p;
p = stringn(b, p, display->black, ZP, font, buf, utfnlen(buf, tick));
draw(b, Rect(p.x-1, p.y, p.x+2, p.y+3), display->black, nil, ZP);
draw(b, Rect(p.x, p.y, p.x+1, p.y+h), display->black, nil, ZP);
draw(b, Rect(p.x-1, p.y+h-3, p.x+2, p.y+h), display->black, nil, ZP);
p = string(b, p, display->black, ZP, font, buf+tick);
}
flushimage(display, 1);
nodraw:
i = Ekeyboard;
if(m != nil)
i |= Emouse;
replclipr(b, 0, sc);
i = eread(i, &ev);
/* screen might have been resized */
if(b != screen || !eqrect(screen->clipr, sc)){
freeimage(save);
save = nil;
}
b = screen;
sc = b->clipr;
replclipr(b, 0, b->r);
switch(i){
default:
done = 1;
n = -1;
break;
case Ekeyboard:
k = ev.kbdc;
if(buf == nil || k == Keof || k == '\n'){
done = 1;
break;
}
if(k == Knack || k == Kesc){
done = !n;
buf[n = tick = 0] = 0;
break;
}
if(k == Ksoh || k == Khome){
tick = 0;
continue;
}
if(k == Kenq || k == Kend){
tick = n;
continue;
}
if(k == Kright){
if(tick < n)
tick += chartorune(&k, buf+tick);
continue;
}
if(k == Kleft){
for(i = 0; i < n; i += l){
l = chartorune(&k, buf+tick);
if(i+l >= tick){
tick = i;
break;
}
}
continue;
}
if(k == Ketb){
while(tick > 0){
tick--;
if(tick == 0 ||
strchr(" !\"#$%&'()*+,-./:;<=>?@`[\\]^{|}~", buf[tick-1]))
break;
}
buf[n = tick] = 0;
break;
}
if(k == Kbs){
if(tick <= 0)
continue;
for(i = 0; i < n; i += l){
l = chartorune(&k, buf+i);
if(i+l >= tick){
memmove(buf+i, buf+i+l, n - (i+l));
buf[n -= l] = 0;
tick -= l;
break;
}
}
break;
}
if(k < 0x20 || k == Kdel || (k & 0xFF00) == KF || (k & 0xFF00) == Spec)
continue;
if((len-n) <= (l = runelen(k)))
continue;
memmove(buf+tick+l, buf+tick, n - tick);
runetochar(buf+tick, &k);
buf[n += l] = 0;
tick += l;
break;
case Emouse:
*m = ev.mouse;
if(!ptinrect(m->xy, r)){
down = 0;
goto nodraw;
}
if(m->buttons & 7){
down = 1;
if(buf && m->xy.x >= (t.x - w)){
down = 0;
for(i = 0; i < n; i += l){
l = chartorune(&k, buf+i);
t.x += stringnwidth(font, buf+i, 1);
if(t.x > m->xy.x)
break;
}
tick = i;
}
continue;
}
done = down;
break;
}
if(save){
draw(b, save->r, save, nil, save->r.min);
freeimage(save);
save = nil;
}
}
replclipr(b, 0, sc);
freeimage(backcol);
freeimage(bordcol);
flushimage(display, 1);
return n;
}
char*
tcs(char *charset, char *s)
{
char *buf;
int i, n, nbuf;
int fd[3], p[2], pp[2];
uchar *us;
char *t, *u;
char *argv[4];
Rune r;
Writeargs *w;
if(s == nil || charset == nil || *s == 0)
return s;
if(cistrcmp(charset, "utf-8") == 0)
return s;
if(cistrcmp(charset, "iso-8859-1") == 0 || cistrcmp(charset, "us-ascii") == 0){
latin1:
n = 0;
for(us=(uchar*)s; *us; us++)
n += runelen(*us);
n++;
t = emalloc(n);
for(us=(uchar*)s, u=t; *us; us++){
r = *us;
u += runetochar(u, &r);
}
*u = 0;
free(s);
return t;
}
for(i=0; i<nelem(tcstab); i++)
if(cistrcmp(charset, tcstab[i].mime) == 0)
goto tcs;
goto latin1;
tcs:
argv[0] = "tcs";
argv[1] = "-f";
argv[2] = charset;
argv[3] = nil;
if(pipe(p) < 0 || pipe(pp) < 0)
sysfatal("pipe: %r");
fd[0] = p[0];
fd[1] = pp[0];
fd[2] = dup(2, -1);
if(threadspawnl(fd, "tcs", "tcs", "-f", tcstab[i].tcs, nil) < 0){
close(p[0]);
close(p[1]);
close(pp[0]);
close(pp[1]);
close(fd[2]);
goto latin1;
}
close(p[0]);
close(pp[0]);
nbuf = UTFmax*strlen(s)+100; /* just a guess at worst case */
buf = emalloc(nbuf);
w = emalloc(sizeof *w);
w->fd = p[1];
w->s = estrdup(s);
proccreate(twriter, w, STACK);
n = readn(pp[1], buf, nbuf-1);
close(pp[1]);
if(n <= 0){
free(buf);
goto latin1;
}
buf[n] = 0;
free(s);
s = estrdup(buf);
free(buf);
return s;
}
static int
__sendinput(Window *w, uint32_t q0, uint32_t q1)
{
char *s, *t;
int n, nb, eofchar;
static int partial;
static char tmp[UTFmax];
Req *r;
Rune rune;
if(!q)
return 0;
r = q;
n = 0;
if(partial){
Partial:
nb = partial;
if(nb > r->ifcall.count)
nb = r->ifcall.count;
memmove(r->ofcall.data, tmp, nb);
if(nb!=partial)
memmove(tmp, tmp+nb, partial-nb);
partial -= nb;
q = r->aux;
if(q == nil)
eq = &q;
r->aux = nil;
r->ofcall.count = nb;
if(debug)
fprint(2, "satisfy read with partial\n");
respond(r, nil);
return n;
}
if(q0==q1)
return 0;
s = emalloc((q1-q0)*UTFmax+1);
n = winread(w, q0, q1, s);
s[n] = '\0';
t = strpbrk(s, "\n\004");
if(t == nil){
free(s);
return 0;
}
r = q;
eofchar = 0;
if(*t == '\004'){
eofchar = 1;
*t = '\0';
}else
*++t = '\0';
nb = utfncpy((char*)r->ofcall.data, s, r->ifcall.count);
if(nb==0 && s<t && r->ifcall.count > 0){
partial = utfncpy(tmp, s, UTFmax);
assert(partial > 0);
chartorune(&rune, tmp);
partial = runelen(rune);
free(s);
n = 1;
goto Partial;
}
n = utfnlen(r->ofcall.data, nb);
if(nb==strlen(s) && eofchar)
n++;
r->ofcall.count = nb;
q = r->aux;
if(q == nil)
eq = &q;
r->aux = nil;
if(debug)
fprint(2, "read returns %lud-%lud: %.*q\n", q0, q0+n, n, r->ofcall.data);
respond(r, nil);
return n;
}
int
charntorune(Rune *p, const char *s, size_t len)
{
unsigned int i, n;
Rune r;
if(len == 0) /* can't even look at s[0] */
return 0;
if((s[0] & 0x80) == 0x00) { /* 0xxxxxxx */
*p = s[0];
return 1;
}
else if((s[0] & 0xE0) == 0xC0) { /* 110xxxxx */
r = s[0] & 0x1F;
n = 2;
}
else if((s[0] & 0xF0) == 0xE0) { /* 1110xxxx */
r = s[0] & 0x0F;
n = 3;
}
else if((s[0] & 0xF8) == 0xF0) { /* 11110xxx */
r = s[0] & 0x07;
n = 4;
}
else if((s[0] & 0xFC) == 0xF8) { /* 111110xx */
r = s[0] & 0x03;
n = 5;
}
else if((s[0] & 0xFE) == 0xFC) { /* 1111110x */
r = s[0] & 0x01;
n = 6;
}
else { /* invalid leading byte */
*p = Runeerror;
return 1;
}
if(len > n)
len = n;
/* add values from continuation bytes */
for(i = 1; i < len; i++)
if((s[i] & 0xC0) == 0x80) {
/* add bits from continuation byte to rune value
* cannot overflow: 6 byte sequences contain 31 bits */
r = (r << 6) | (s[i] & 0x3F); /* 10xxxxxx */
}
else { /* expected continuation */
*p = Runeerror;
return i;
}
if(i < n) /* must have reached len limit */
return 0;
/* reject invalid or overlong sequences */
if(runelen(r) < (int)n)
r = Runeerror;
*p = r;
return n;
}
