fmt: improved version using less memory
functional identical, but uses different data structure. instead of reading all words into memory, we read and tokenize on demand keeping a simple linked list of words to conume by further calls of getword(). once the list is empty we read the next line, tokenize it and fill the list up again. so we only need as mutch memory to hold all the words of the longest line.
This commit is contained in:
parent
22c7584780
commit
4fdc7dbe9c
|
@ -12,14 +12,19 @@ int indent = 0; /* current value of indent, before extra indent */
|
|||
int length = 70; /* how many columns per output line */
|
||||
int join = 1; /* can lines be joined? */
|
||||
int maxtab = 8;
|
||||
|
||||
Biobuf bin;
|
||||
Biobuf bout;
|
||||
|
||||
typedef struct Word Word;
|
||||
struct Word{
|
||||
int bol;
|
||||
struct Word
|
||||
{
|
||||
Word *next;
|
||||
|
||||
int indent;
|
||||
char text[1];
|
||||
int length;
|
||||
char bol;
|
||||
char text[];
|
||||
};
|
||||
|
||||
void fmt(void);
|
||||
|
@ -84,17 +89,14 @@ main(int argc, char **argv)
|
|||
}
|
||||
|
||||
int
|
||||
indentof(char **linep)
|
||||
indentof(char *s)
|
||||
{
|
||||
int i, ind;
|
||||
char *line;
|
||||
int ind;
|
||||
|
||||
ind = 0;
|
||||
line = *linep;
|
||||
for(i=0; line[i]; i++)
|
||||
switch(line[i]){
|
||||
for(; *s != '\0'; s++)
|
||||
switch(*s){
|
||||
default:
|
||||
*linep = line;
|
||||
return ind;
|
||||
case ' ':
|
||||
ind++;
|
||||
|
@ -104,53 +106,65 @@ indentof(char **linep)
|
|||
ind -= ind%maxtab;
|
||||
break;
|
||||
}
|
||||
|
||||
|
||||
/* plain white space doesn't change the indent */
|
||||
*linep = "";
|
||||
return indent;
|
||||
}
|
||||
|
||||
Word**
|
||||
addword(Word **words, int *nwordp, char *s, int l, int indent, int bol)
|
||||
Word*
|
||||
newword(char *s, int n, int ind, int bol)
|
||||
{
|
||||
Word *w;
|
||||
|
||||
w = malloc(sizeof(Word)+l+1);
|
||||
memmove(w->text, s, l);
|
||||
w->text[l] = '\0';
|
||||
w->indent = indent;
|
||||
w = malloc(sizeof(Word) + n+1);
|
||||
w->next = nil;
|
||||
w->indent = ind;
|
||||
w->bol = bol;
|
||||
words = realloc(words, (*nwordp+1)*sizeof(Word*));
|
||||
words[(*nwordp)++] = w;
|
||||
return words;
|
||||
memmove(w->text, s, n);
|
||||
w->text[n] = 0;
|
||||
w->length = utflen(w->text);
|
||||
return w;
|
||||
}
|
||||
|
||||
Word**
|
||||
parseline(char *line, Word **words, int *nwordp)
|
||||
Word*
|
||||
getword(void)
|
||||
{
|
||||
int ind, l, bol;
|
||||
|
||||
ind = indentof(&line);
|
||||
indent = ind;
|
||||
bol = 1;
|
||||
static Word *head, *tail;
|
||||
char *line, *s;
|
||||
Word *w;
|
||||
|
||||
w = head;
|
||||
if(w != nil){
|
||||
head = w->next;
|
||||
return w;
|
||||
}
|
||||
line = Brdstr(&bin, '\n', 1);
|
||||
if(line == nil)
|
||||
return nil;
|
||||
tail = nil;
|
||||
indent = indentof(line);
|
||||
for(;;){
|
||||
/* find next word */
|
||||
while(*line==' ' || *line=='\t')
|
||||
while(*line == ' ' || *line == '\t')
|
||||
line++;
|
||||
if(*line == '\0'){
|
||||
if(bol)
|
||||
return addword(words, nwordp, "", 0, -1, bol);
|
||||
if(head == nil)
|
||||
return newword("", 0, -1, 1);
|
||||
break;
|
||||
}
|
||||
/* how long is this word? */
|
||||
for(l=0; line[l]; l++)
|
||||
if(line[l]==' ' || line[l]=='\t')
|
||||
for(s=line++; *line != '\0'; line++)
|
||||
if(*line==' ' || *line=='\t')
|
||||
break;
|
||||
words = addword(words, nwordp, line, l, indent, bol);
|
||||
bol = 0;
|
||||
line += l;
|
||||
w = newword(s, line-s, indent, head==nil);
|
||||
if(head == nil)
|
||||
head = w;
|
||||
else
|
||||
tail->next = w;
|
||||
tail = w;
|
||||
}
|
||||
return words;
|
||||
w = head;
|
||||
head = w->next;
|
||||
return w;
|
||||
}
|
||||
|
||||
void
|
||||
|
@ -181,61 +195,46 @@ nspaceafter(char *s)
|
|||
return 2;
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
void
|
||||
printwords(Word **w, int nw)
|
||||
{
|
||||
int i, j, n, col, nsp;
|
||||
|
||||
/* one output line per loop */
|
||||
for(i=0; i<nw; ){
|
||||
/* if it's a blank line, print it */
|
||||
if(w[i]->indent == -1){
|
||||
Bputc(&bout, '\n');
|
||||
if(++i == nw) /* out of words */
|
||||
break;
|
||||
}
|
||||
/* emit leading indent */
|
||||
col = extraindent+w[i]->indent;
|
||||
printindent(col);
|
||||
/* emit words until overflow; always emit at least one word */
|
||||
for(n=0;; n++){
|
||||
Bprint(&bout, "%s", w[i]->text);
|
||||
col += utflen(w[i]->text);
|
||||
if(++i == nw)
|
||||
break; /* out of words */
|
||||
if(w[i]->indent != w[i-1]->indent)
|
||||
break; /* indent change */
|
||||
nsp = nspaceafter(w[i-1]->text);
|
||||
if(col+nsp+utflen(w[i]->text) > extraindent+length)
|
||||
break; /* fold line */
|
||||
if(!join && w[i]->bol)
|
||||
break;
|
||||
for(j=0; j<nsp; j++)
|
||||
Bputc(&bout, ' '); /* emit space; another word will follow */
|
||||
col += nsp;
|
||||
}
|
||||
/* emit newline */
|
||||
Bputc(&bout, '\n');
|
||||
}
|
||||
}
|
||||
|
||||
void
|
||||
fmt(void)
|
||||
{
|
||||
char *s;
|
||||
int i, nw;
|
||||
Word **w;
|
||||
Word *w, *o;
|
||||
int col, nsp;
|
||||
|
||||
nw = 0;
|
||||
w = nil;
|
||||
while((s = Brdstr(&bin, '\n', 1)) != nil){
|
||||
w = parseline(s, w, &nw);
|
||||
free(s);
|
||||
w = getword();
|
||||
while(w != nil){
|
||||
if(w->indent == -1){
|
||||
Bputc(&bout, '\n');
|
||||
free(w);
|
||||
w = getword();
|
||||
if(w == nil)
|
||||
break;
|
||||
}
|
||||
col = w->indent;
|
||||
printindent(extraindent+col);
|
||||
/* emit words until overflow; always emit at least one word */
|
||||
for(;;){
|
||||
Bprint(&bout, "%s", w->text);
|
||||
col += w->length;
|
||||
o = w;
|
||||
w = getword();
|
||||
if(w == nil)
|
||||
break;
|
||||
if(w->indent != o->indent)
|
||||
break; /* indent change */
|
||||
nsp = nspaceafter(o->text);
|
||||
if(col+nsp+w->length > length)
|
||||
break; /* fold line */
|
||||
if(!join && w->bol)
|
||||
break;
|
||||
while(--nsp >= 0){
|
||||
Bputc(&bout, ' '); /* emit space; another word will follow */
|
||||
col++;
|
||||
}
|
||||
free(o);
|
||||
}
|
||||
free(o);
|
||||
Bputc(&bout, '\n');
|
||||
}
|
||||
printwords(w, nw);
|
||||
for(i=0; i<nw; i++)
|
||||
free(w[i]);
|
||||
free(w);
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue