xfnw/plan9fox
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

mothra: fix unicode rendering for plaintext

Browse source 
we did the utf-8 to unicode conversion in pl_nextc(),
but the plaintext handler uses pl_readc() which only
translates newlines but otherwise returns bytes.

move unicode conversion in pl_readc() fixes it.
This commit is contained in:
cinap_lenrek 2014-02-16 18:50:30 +01:00
parent c54271a588
commit 218e1b4b3e
1 changed files with 21 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

34
sys/src/cmd/mothra/rdhtml.c
View file

@ -172,10 +172,14 @@ int pl_bread(Hglob *g){
} }
/* /*
* Read a character, translating \r\n, \n\r, \r and \n into \n * Read a character, translating \r\n, \n\r, \r and \n into \n
* convert to runes.
*/ */
int pl_readc(Hglob *g){ int pl_readc(Hglob *g){
int c;
static int peek=-1; static int peek=-1;
char crune[UTFmax+1];
int c, n;
Rune r;
if(peek!=-1){ if(peek!=-1){
c=peek; c=peek;
peek=-1; peek=-1;
@ -192,6 +196,21 @@ int pl_readc(Hglob *g){
if(c!='\r') peek=c; if(c!='\r') peek=c;
return '\n'; return '\n';
} }
if(c < Runeself)
return c;
crune[0]=c;
for (n=1; n<=sizeof(crune); n++){
if(fullrune(crune, n)){
chartorune(&r, crune);
return r;
}
c=pl_bread(g);
if(c==EOF)
return EOF;
crune[n]=c;
}
return c; return c;
} }
void pl_putback(Hglob *g, int c){ void pl_putback(Hglob *g, int c){
@ -202,7 +221,6 @@ int pl_nextc(Hglob *g){
int c; int c;
int n; int n;
Rune r; Rune r;
char crune[UTFmax+1];
if(g->heof) return EOF; if(g->heof) return EOF;
if(g->npeekc!=0) return g->peekc[--g->npeekc]; if(g->npeekc!=0) return g->peekc[--g->npeekc];
c=pl_readc(g); c=pl_readc(g);
@ -220,19 +238,9 @@ int pl_nextc(Hglob *g){
return '<'; return '<';
} }
if(c=='>') return ETAG; if(c=='>') return ETAG;
if(c==EOF) return c;
for (n=1; n<=sizeof(crune); n++){
crune[n-1]=c;
if(fullrune(crune, n)){
chartorune(&r, crune);
return r;
}
c=pl_readc(g);
if(c==EOF)
return EOF;
}
return c; return c;
} }
char *unquot(char *dst, char *src, int len){ char *unquot(char *dst, char *src, int len){
char *e; char *e;