emo/busybox
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

awk: tighten rules in action parsing

Browse Source 
Disallow:
    BEGIN
	{ action }  - must start on the same line
Disallow:
    func f()
	print "hello" - must be in {...}

function                                             old     new   delta
chain_until_rbrace                                     -      41     +41
parse_program                                        307     336     +29
chain_group                                          649     616     -33
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 1/1 up/down: 70/-33)             Total: 37 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This commit is contained in:
Denys Vlasenko 2021-07-03 01:16:48 +02:00
parent 717200eb43
commit 2b65e73db3
1 changed files with 66 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

108
editors/awk.c
View File

@ -1549,29 +1549,35 @@ static node *chain_loop(node *nn)
return n; return n;
} }
static void chain_until_rbrace(void)
{
uint32_t tc;
while ((tc = next_token(TS_GRPSEQ | TC_RBRACE)) != TC_RBRACE) {
debug_printf_parse("%s: !TC_RBRACE\n", __func__);
if (tc == TC_NEWLINE)
continue;
rollback_token();
chain_group();
}
debug_printf_parse("%s: TC_RBRACE\n", __func__);
}
/* parse group and attach it to chain */ /* parse group and attach it to chain */
static void chain_group(void) static void chain_group(void)
{ {
uint32_t c; uint32_t tc;
node *n, *n2, *n3; node *n, *n2, *n3;
do { do {
c = next_token(TS_GRPSEQ); tc = next_token(TS_GRPSEQ);
} while (c & TC_NEWLINE); } while (tc == TC_NEWLINE);
if (c & TC_LBRACE) { if (tc == TC_LBRACE) {
debug_printf_parse("%s: TC_LBRACE\n", __func__); debug_printf_parse("%s: TC_LBRACE\n", __func__);
while ((c = next_token(TS_GRPSEQ | TC_RBRACE)) != TC_RBRACE) { chain_until_rbrace();
debug_printf_parse("%s: !TC_RBRACE\n", __func__);
if (c & TC_NEWLINE)
continue;
rollback_token();
chain_group();
}
debug_printf_parse("%s: TC_RBRACE\n", __func__);
return; return;
} }
if (c & (TS_OPSEQ | TS_OPTERM)) { if (tc & (TS_OPSEQ | TS_OPTERM)) {
debug_printf_parse("%s: TS_OPSEQ | TS_OPTERM\n", __func__); debug_printf_parse("%s: TS_OPSEQ | TS_OPTERM\n", __func__);
rollback_token(); rollback_token();
chain_expr(OC_EXEC | Vx); chain_expr(OC_EXEC | Vx);
@ -1675,37 +1681,48 @@ static void chain_group(void)
static void parse_program(char *p) static void parse_program(char *p)
{ {
uint32_t tclass;
node *cn;
func *f;
var *v;
debug_printf_parse("%s()\n", __func__); debug_printf_parse("%s()\n", __func__);
g_pos = p; g_pos = p;
t_lineno = 1; t_lineno = 1;
while ((tclass = next_token(TC_EOF | TS_OPSEQ | TC_LBRACE | for (;;) {
TS_OPTERM | TC_BEGIN | TC_END | TC_FUNCDECL)) != TC_EOF) { uint32_t tclass;
if (tclass & TS_OPTERM) { tclass = next_token(TC_EOF | TS_OPSEQ | TC_LBRACE |
TS_OPTERM | TC_BEGIN | TC_END | TC_FUNCDECL);
if (tclass == TC_EOF) {
debug_printf_parse("%s: TC_EOF\n", __func__);
break;
}
if (tclass & TS_OPTERM) { /* ; or <newline> */
debug_printf_parse("%s: TS_OPTERM\n", __func__); debug_printf_parse("%s: TS_OPTERM\n", __func__);
//NB: gawk allows many newlines, but does not allow more than one semicolon:
// BEGIN {...}<newline>;<newline>;
//would complain "each rule must have a pattern or an action part".
//Same message for
// ; BEGIN {...}
continue; continue;
} }
if (tclass == TC_BEGIN) {
seq = &mainseq;
if (tclass & TC_BEGIN) {
debug_printf_parse("%s: TC_BEGIN\n", __func__); debug_printf_parse("%s: TC_BEGIN\n", __func__);
seq = &beginseq; seq = &beginseq;
//TODO: ensure there is no newline between BEGIN and { /* ensure there is no newline between BEGIN and { */
//next_token(TC_LBRACE); rollback_token(); next_token(TC_LBRACE);
chain_group(); chain_until_rbrace();
} else if (tclass & TC_END) { continue;
}
if (tclass == TC_END) {
debug_printf_parse("%s: TC_END\n", __func__); debug_printf_parse("%s: TC_END\n", __func__);
seq = &endseq; seq = &endseq;
//TODO: ensure there is no newline between END and { /* ensure there is no newline between END and { */
//next_token(TC_LBRACE); rollback_token(); next_token(TC_LBRACE);
chain_group(); chain_until_rbrace();
} else if (tclass & TC_FUNCDECL) { continue;
}
if (tclass == TC_FUNCDECL) {
func *f;
debug_printf_parse("%s: TC_FUNCDECL\n", __func__); debug_printf_parse("%s: TC_FUNCDECL\n", __func__);
next_token(TC_FUNCTION); next_token(TC_FUNCTION);
f = newfunc(t_string); f = newfunc(t_string);
@ -1716,6 +1733,7 @@ static void parse_program(char *p)
//f->nargs = 0; - already is //f->nargs = 0; - already is
/* func arg list: comma sep list of args, and a close paren */ /* func arg list: comma sep list of args, and a close paren */
for (;;) { for (;;) {
var *v;
if (next_token(TC_VARIABLE | TC_RPAREN) == TC_RPAREN) { if (next_token(TC_VARIABLE | TC_RPAREN) == TC_RPAREN) {
if (f->nargs == 0) if (f->nargs == 0)
break; /* func() is ok */ break; /* func() is ok */
@ -1730,31 +1748,37 @@ static void parse_program(char *p)
/* it was a comma, we ate it */ /* it was a comma, we ate it */
} }
seq = &f->body; seq = &f->body;
//TODO: ensure there is { after "func F(...)" - but newlines are allowed /* ensure there is { after "func F(...)" - but newlines are allowed */
//while (next_token(TC_LBRACE | TC_NEWLINE) == TC_NEWLINE) continue; rollback_token(); while (next_token(TC_LBRACE | TC_NEWLINE) == TC_NEWLINE)
chain_group(); continue;
chain_until_rbrace();
hash_clear(ahash); hash_clear(ahash);
} else if (tclass & TS_OPSEQ) { continue;
}
seq = &mainseq;
if (tclass & TS_OPSEQ) {
node *cn;
debug_printf_parse("%s: TS_OPSEQ\n", __func__); debug_printf_parse("%s: TS_OPSEQ\n", __func__);
rollback_token(); rollback_token();
cn = chain_node(OC_TEST); cn = chain_node(OC_TEST);
cn->l.n = parse_expr(TS_OPTERM | TC_EOF | TC_LBRACE); cn->l.n = parse_expr(TS_OPTERM | TC_EOF | TC_LBRACE);
if (t_tclass & TC_LBRACE) { if (t_tclass == TC_LBRACE) {
debug_printf_parse("%s: TC_LBRACE\n", __func__); debug_printf_parse("%s: TC_LBRACE\n", __func__);
rollback_token(); rollback_token();
chain_group(); chain_group();
} else { } else {
/* no action, assume default "{ print }" */
debug_printf_parse("%s: !TC_LBRACE\n", __func__); debug_printf_parse("%s: !TC_LBRACE\n", __func__);
chain_node(OC_PRINT); chain_node(OC_PRINT);
} }
cn->r.n = mainseq.last; cn->r.n = mainseq.last;
} else /* if (tclass & TC_LBRACE) */ { continue;
debug_printf_parse("%s: TC_LBRACE(?)\n", __func__);
rollback_token();
chain_group();
} }
/* tclass == TC_LBRACE */
debug_printf_parse("%s: TC_LBRACE(?)\n", __func__);
chain_until_rbrace();
} }
debug_printf_parse("%s: TC_EOF\n", __func__);
} }