8b4c42902
Denys Vlasenko
2021-07-01
Subject
awk: use static tmpvars instead of nvalloc(1)ed ones
Binary
No binary (build failed)
Architecture & Compiler Flags
—
Build Log
(log yok)
Commit Diff GitHub →
editors/awk.c+34 -15
| @@ -559,7 +559,9 @@ struct globals2 { | |
| unsigned evaluate__seed; | |
| regex_t evaluate__sreg; | |
| - | var ptest__v; |
| + | var ptest__tmpvar; |
| + | var awk_printf__tmpvar; |
| + | var as_regex__tmpvar; |
| tsplitter exec_builtin__tspl; | |
| @@ -1775,14 +1777,19 @@ static node *mk_splitter(const char *s, tsplitter *spl) | |
| static regex_t *as_regex(node *op, regex_t *preg) | |
| { | |
| int cflags; | |
| - | var *tmpvar; |
| const char *s; | |
| if ((op->info & OPCLSMASK) == OC_REGEXP) { | |
| return icase ? op->r.ire : op->l.re; | |
| } | |
| - | tmpvar = nvalloc(1); |
| - | s = getvar_s(evaluate(op, tmpvar)); |
| + | |
| + | #define TMPVAR (&G.as_regex__tmpvar) |
| + | //tmpvar = nvalloc(1); |
| + | // We use a single "static" tmpvar (instead of on-stack or malloced one) |
| + | // to decrease memory consumption in deeply-recursive awk programs. |
| + | // The rule to work safely is to never call evaluate() while our static |
| + | // TMPVAR's value is still needed. |
| + | s = getvar_s(evaluate(op, TMPVAR)); |
| cflags = icase ? REG_EXTENDED | REG_ICASE : REG_EXTENDED; | |
| /* Testcase where REG_EXTENDED fails (unpaired '{'): | |
| @@ -1794,7 +1801,8 @@ static regex_t *as_regex(node *op, regex_t *preg) | |
| cflags &= ~REG_EXTENDED; | |
| xregcomp(preg, s, cflags); | |
| } | |
| - | nvfree(tmpvar, 1); |
| + | //nvfree(tmpvar, 1); |
| + | #undef TMPVAR |
| return preg; | |
| } | |
| @@ -2105,8 +2113,11 @@ static int hashwalk_next(var *v) | |
| /* evaluate node, return 1 when result is true, 0 otherwise */ | |
| static int ptest(node *pattern) | |
| { | |
| - | /* ptest__v is "static": to save stack space? */ |
| - | return istrue(evaluate(pattern, &G.ptest__v)); |
| + | // We use a single "static" tmpvar (instead of on-stack or malloced one) |
| + | // to decrease memory consumption in deeply-recursive awk programs. |
| + | // The rule to work safely is to never call evaluate() while our static |
| + | // TMPVAR's value is still needed. |
| + | return istrue(evaluate(pattern, &G.ptest__tmpvar)); |
| } | |
| /* read next record from stream rsm into a variable v */ | |
| @@ -2243,12 +2254,18 @@ static char *awk_printf(node *n, int *len) | |
| const char *s1; | |
| int i, j, incr, bsize; | |
| char c, c1; | |
| - | var *tmpvar, *arg; |
| - | |
| - | tmpvar = nvalloc(1); |
| - | //TODO: above, to avoid allocating a single temporary var, take a pointer |
| - | //to a temporary that our caller (evaluate()) already has? |
| - | fmt = f = xstrdup(getvar_s(evaluate(nextarg(&n), tmpvar))); |
| + | var *arg; |
| + | |
| + | //tmpvar = nvalloc(1); |
| + | #define TMPVAR (&G.awk_printf__tmpvar) |
| + | // We use a single "static" tmpvar (instead of on-stack or malloced one) |
| + | // to decrease memory consumption in deeply-recursive awk programs. |
| + | // The rule to work safely is to never call evaluate() while our static |
| + | // TMPVAR's value is still needed. |
| + | fmt = f = xstrdup(getvar_s(evaluate(nextarg(&n), TMPVAR))); |
| + | // ^^^^^^^^^ here we immediately strdup() the value, so the later call |
| + | // to evaluate() potentially recursing into another awk_printf() can't |
| + | // mangle the value. |
| i = 0; | |
| while (*f) { | |
| @@ -2268,7 +2285,7 @@ static char *awk_printf(node *n, int *len) | |
| f++; | |
| c1 = *f; | |
| *f = '\0'; | |
| - | arg = evaluate(nextarg(&n), tmpvar); |
| + | arg = evaluate(nextarg(&n), TMPVAR); |
| j = i; | |
| if (c == 'c' || !c) { | |
| @@ -2289,7 +2306,9 @@ static char *awk_printf(node *n, int *len) | |
| } | |
| free(fmt); | |
| - | nvfree(tmpvar, 1); |
| + | // nvfree(tmpvar, 1); |
| + | #undef TMPVAR |
| + | |
| b = xrealloc(b, i + 1); | |
| b[i] = '\0'; | |
| #if ENABLE_FEATURE_AWK_GNU_EXTENSIONS | |