aboutsummaryrefslogtreecommitdiffstats
path: root/erts/emulator/beam
diff options
context:
space:
mode:
authorPatrik Nyblom <[email protected]>2013-08-16 11:37:54 +0200
committerPatrik Nyblom <[email protected]>2013-08-16 11:37:54 +0200
commit5d9a587a8fcc164e02f043959338edec2ff69381 (patch)
tree554f05a944777622b30031724010f6f31707565b /erts/emulator/beam
parent23610dbfc1c409f83349e9e293dd3cfc1f74d497 (diff)
parent52cb62b7930d9c7b9e04a210ff6b02946f27ae79 (diff)
downloadotp-5d9a587a8fcc164e02f043959338edec2ff69381.tar.gz
otp-5d9a587a8fcc164e02f043959338edec2ff69381.tar.bz2
otp-5d9a587a8fcc164e02f043959338edec2ff69381.zip
Merge branch 'pan/update_pcre_8.33'
* pan/update_pcre_8.33: Workaround TR gnu/181328, GCC 4.2.1 20070831 on FreeBSD 9.1 Clarify relation between erts_iolist_{size|to_buf} Fix backslash in titles of manpages Correct UTF-8 in stdlib's notes.xml Add more tests for corner error cases in erl_bif_re.c Add documentation of report_errors and match_limit(_recursion) Add match_limit and match_limit_recursion options Add return_errors option to re:run/3 Add README for updating PCRE Add documentation of extensions to re module Add new options to Erlang re interface and mend dupnames Update PCRE doc part of re.xml to PCRE 8.33 state Integrate new PCRE test suites Integrate patch for PCRE bug id 1370 Handle CRLF correctly in global regexp Add erts_prefix to pcre_library and update erl_bif_re Update to PCRE 8.33, w/o the erts_ prefix added OTP-11204 OTP-11205 OTP-10285
Diffstat (limited to 'erts/emulator/beam')
-rw-r--r--erts/emulator/beam/atom.names9
-rw-r--r--erts/emulator/beam/bif.tab6
-rw-r--r--erts/emulator/beam/erl_bif_re.c469
-rw-r--r--erts/emulator/beam/utils.c13
4 files changed, 443 insertions, 54 deletions
diff --git a/erts/emulator/beam/atom.names b/erts/emulator/beam/atom.names
index c2f32ba089..e108e706c3 100644
--- a/erts/emulator/beam/atom.names
+++ b/erts/emulator/beam/atom.names
@@ -71,6 +71,7 @@ atom ac
atom active
atom all
atom all_but_first
+atom all_names
atom alloc_info
atom alloc_sizes
atom allocated
@@ -322,6 +323,8 @@ atom low
atom Lt='<'
atom machine
atom match
+atom match_limit
+atom match_limit_recursion
atom match_spec
atom max
atom maximum
@@ -351,11 +354,13 @@ atom multi_scheduling
atom multiline
atom name
atom named_table
+atom namelist
atom native_addresses
atom Neq='=/='
atom Neqeq='/='
atom net_kernel
atom net_kernel_terminated
+atom never_utf
atom new
atom new_index
atom new_uniq
@@ -381,6 +386,7 @@ atom nosuspend
atom no_float
atom no_integer
atom no_network
+atom no_start_optimize
atom not
atom not_a_list
atom not_loaded
@@ -391,6 +397,7 @@ atom notalive
atom notbol
atom noteol
atom notempty
+atom notempty_atstart
atom notify
atom notsup
atom nouse_stdio
@@ -475,6 +482,7 @@ atom register
atom registered_name
atom reload
atom rem
+atom report_errors
atom reset
atom restart
atom return_from
@@ -557,6 +565,7 @@ atom true
atom tuple
atom type
atom ucompile
+atom ucp
atom undef
atom ungreedy
atom unicode
diff --git a/erts/emulator/beam/bif.tab b/erts/emulator/beam/bif.tab
index dc8e9101de..7c8e4b31cf 100644
--- a/erts/emulator/beam/bif.tab
+++ b/erts/emulator/beam/bif.tab
@@ -574,6 +574,12 @@ bif erlang:binary_to_float/1
bif io:printable_range/0
#
+# New in R17A
+#
+
+bif re:inspect/2
+
+#
# Obsolete
#
diff --git a/erts/emulator/beam/erl_bif_re.c b/erts/emulator/beam/erl_bif_re.c
index 3d34c2a77f..99c31738a5 100644
--- a/erts/emulator/beam/erl_bif_re.c
+++ b/erts/emulator/beam/erl_bif_re.c
@@ -180,6 +180,9 @@ static Eterm make_signed_integer(int x, Process *p)
#define PARSE_FLAG_STARTOFFSET 8
#define PARSE_FLAG_CAPTURE_OPT 16
#define PARSE_FLAG_GLOBAL 32
+#define PARSE_FLAG_REPORT_ERRORS 64
+#define PARSE_FLAG_MATCH_LIMIT 128
+#define PARSE_FLAG_MATCH_LIMIT_RECURSION 256
#define CAPSPEC_VALUES 0
#define CAPSPEC_TYPE 1
@@ -192,7 +195,9 @@ parse_options(Eterm listp, /* in */
int *exec_options, /* out */
int *flags,/* out */
int *startoffset, /* out */
- Eterm *capture_spec) /* capture_spec[CAPSPEC_SIZE] */ /* out */
+ Eterm *capture_spec, /* capture_spec[CAPSPEC_SIZE] */ /* out */
+ int *match_limit, /* out */
+ int *match_limit_recursion) /* out */
{
int copt,eopt,fl;
Eterm item;
@@ -234,7 +239,7 @@ parse_options(Eterm listp, /* in */
case am_offset:
{
int tmp;
- if (!term_to_int(tp[2],&tmp)) {
+ if (!term_to_int(tp[2],&tmp) || tmp < 0) {
return -1;
}
if (startoffset != NULL) {
@@ -243,6 +248,31 @@ parse_options(Eterm listp, /* in */
}
fl |= (PARSE_FLAG_UNIQUE_EXEC_OPT|PARSE_FLAG_STARTOFFSET);
break;
+ case am_match_limit:
+ {
+ int tmp;
+ if (!term_to_int(tp[2],&tmp) || tmp < 0) {
+ return -1;
+ }
+ if (match_limit != NULL) {
+ *match_limit = tmp;
+ }
+ }
+ fl |= (PARSE_FLAG_UNIQUE_EXEC_OPT|PARSE_FLAG_MATCH_LIMIT);
+ break;
+ case am_match_limit_recursion:
+ {
+ int tmp;
+ if (!term_to_int(tp[2],&tmp) || tmp < 0) {
+ return -1;
+ }
+ if (match_limit_recursion != NULL) {
+ *match_limit_recursion = tmp;
+ }
+ }
+ fl |= (PARSE_FLAG_UNIQUE_EXEC_OPT|
+ PARSE_FLAG_MATCH_LIMIT_RECURSION);
+ break;
case am_newline:
if (!is_atom(tp[2])) {
return -1;
@@ -276,7 +306,7 @@ parse_options(Eterm listp, /* in */
default:
return -1;
}
- }else if (is_not_atom(item)) {
+ } else if (is_not_atom(item)) {
return -1;
} else {
switch(item) {
@@ -288,6 +318,10 @@ parse_options(Eterm listp, /* in */
eopt |= PCRE_NOTEMPTY;
fl |= PARSE_FLAG_UNIQUE_EXEC_OPT;
break;
+ case am_notempty_atstart:
+ eopt |= PCRE_NOTEMPTY_ATSTART;
+ fl |= PARSE_FLAG_UNIQUE_EXEC_OPT;
+ break;
case am_notbol:
eopt |= PCRE_NOTBOL;
fl |= PARSE_FLAG_UNIQUE_EXEC_OPT;
@@ -296,6 +330,10 @@ parse_options(Eterm listp, /* in */
eopt |= PCRE_NOTEOL;
fl |= PARSE_FLAG_UNIQUE_EXEC_OPT;
break;
+ case am_no_start_optimize:
+ copt |= PCRE_NO_START_OPTIMIZE;
+ fl |= PARSE_FLAG_UNIQUE_COMPILE_OPT;
+ break;
case am_caseless:
copt |= PCRE_CASELESS;
fl |= PARSE_FLAG_UNIQUE_COMPILE_OPT;
@@ -332,6 +370,18 @@ parse_options(Eterm listp, /* in */
copt |= PCRE_UNGREEDY;
fl |= PARSE_FLAG_UNIQUE_COMPILE_OPT;
break;
+ case am_ucp:
+ copt |= PCRE_UCP;
+ fl |= PARSE_FLAG_UNIQUE_COMPILE_OPT;
+ break;
+ case am_never_utf:
+ copt |= PCRE_NEVER_UTF;
+ fl |= PARSE_FLAG_UNIQUE_COMPILE_OPT;
+ break;
+ case am_report_errors:
+ fl |= (PARSE_FLAG_UNIQUE_EXEC_OPT |
+ PARSE_FLAG_REPORT_ERRORS);
+ break;
case am_unicode:
copt |= PCRE_UTF8;
fl |= (PARSE_FLAG_UNIQUE_COMPILE_OPT | PARSE_FLAG_UNICODE);
@@ -359,7 +409,7 @@ parse_options(Eterm listp, /* in */
if (compile_options != NULL) {
*compile_options = copt;
}
- if (exec_options != NULL) {
+ if (exec_options != NULL) {
*exec_options = eopt;
}
if (flags != NULL) {
@@ -373,34 +423,49 @@ parse_options(Eterm listp, /* in */
*/
static Eterm
-build_compile_result(Process *p, Eterm error_tag, pcre *result, int errcode, const char *errstr, int errofset, int unicode, int with_ok)
+build_compile_result(Process *p, Eterm error_tag, pcre *result, int errcode, const char *errstr, int errofset, int unicode, int with_ok, Eterm extra_err_tag)
{
Eterm *hp;
Eterm ret;
size_t pattern_size;
int capture_count;
+ int use_crlf;
+ unsigned long options;
if (!result) {
/* Return {error_tag, {Code, String, Offset}} */
int elen = sys_strlen(errstr);
int need = 3 /* tuple of 2 */ +
3 /* tuple of 2 */ +
- (2 * elen) /* The error string list */;
+ (2 * elen) /* The error string list */ +
+ ((extra_err_tag != NIL) ? 3 : 0);
hp = HAlloc(p, need);
ret = buf_to_intlist(&hp, (char *) errstr, elen, NIL);
ret = TUPLE2(hp, ret, make_small(errofset));
hp += 3;
+ if (extra_err_tag != NIL) {
+ /* Return {error_tag, {extra_tag,
+ {Code, String, Offset}}} instead */
+ ret = TUPLE2(hp, extra_err_tag, ret);
+ hp += 3;
+ }
ret = TUPLE2(hp, error_tag, ret);
} else {
erts_pcre_fullinfo(result, NULL, PCRE_INFO_SIZE, &pattern_size);
erts_pcre_fullinfo(result, NULL, PCRE_INFO_CAPTURECOUNT, &capture_count);
+ erts_pcre_fullinfo(result, NULL, PCRE_INFO_OPTIONS, &options);
+ options &= PCRE_NEWLINE_CR|PCRE_NEWLINE_LF | PCRE_NEWLINE_CRLF |
+ PCRE_NEWLINE_ANY | PCRE_NEWLINE_ANYCRLF;
+ use_crlf = (options == PCRE_NEWLINE_ANY ||
+ options == PCRE_NEWLINE_CRLF ||
+ options == PCRE_NEWLINE_ANYCRLF);
/* XXX: Optimize - keep in offheap binary to allow this to
be kept across traps w/o need of copying */
ret = new_binary(p, (byte *) result, pattern_size);
erts_pcre_free(result);
- hp = HAlloc(p, (with_ok) ? (3+5) : 5);
- ret = TUPLE4(hp,am_re_pattern, make_small(capture_count), make_small(unicode),ret);
+ hp = HAlloc(p, (with_ok) ? (3+6) : 6);
+ ret = TUPLE5(hp,am_re_pattern, make_small(capture_count), make_small(unicode),make_small(use_crlf),ret);
if (with_ok) {
- hp += 5;
+ hp += 6;
ret = TUPLE2(hp,am_ok,ret);
}
}
@@ -424,9 +489,12 @@ re_compile(Process* p, Eterm arg1, Eterm arg2)
int options = 0;
int pflags = 0;
int unicode = 0;
+#ifdef DEBUG
+ int buffres;
+#endif
- if (parse_options(arg2,&options,NULL,&pflags,NULL,NULL)
+ if (parse_options(arg2,&options,NULL,&pflags,NULL,NULL,NULL,NULL)
< 0) {
BIF_ERROR(p,BADARG);
}
@@ -445,16 +513,19 @@ re_compile(Process* p, Eterm arg1, Eterm arg2)
BIF_ERROR(p,BADARG);
}
expr = erts_alloc(ERTS_ALC_T_RE_TMP_BUF, slen + 1);
- if (erts_iolist_to_buf(arg1, expr, slen) != 0) {
- erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
- BIF_ERROR(p,BADARG);
- }
+#ifdef DEBUG
+ buffres =
+#endif
+ erts_iolist_to_buf(arg1, expr, slen);
+
+ ASSERT(buffres >= 0);
+
expr[slen]='\0';
result = erts_pcre_compile2(expr, options, &errcode,
&errstr, &errofset, default_table);
ret = build_compile_result(p, am_error, result, errcode,
- errstr, errofset, unicode, 1);
+ errstr, errofset, unicode, 1, NIL);
erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
BIF_RET(ret);
}
@@ -492,7 +563,7 @@ typedef struct _return_info {
} ReturnInfo;
typedef struct _restart_context {
- pcre_extra extra;
+ erts_pcre_extra extra;
void *restart_data;
Uint32 flags;
char *subject; /* to be able to free it when done */
@@ -502,6 +573,7 @@ typedef struct _restart_context {
} RestartContext;
#define RESTART_FLAG_SUBJECT_IN_BINARY 0x1
+#define RESTART_FLAG_REPORT_MATCH_LIMIT 0x2
static void cleanup_restart_context(RestartContext *rc)
{
@@ -542,13 +614,29 @@ static Eterm build_exec_return(Process *p, int rc, RestartContext *restartp, Ete
Eterm res;
Eterm *hp;
if (rc <= 0) {
- res = am_nomatch;
+ if (restartp->flags & RESTART_FLAG_REPORT_MATCH_LIMIT) {
+ if (rc == PCRE_ERROR_MATCHLIMIT) {
+ hp = HAlloc(p,3);
+ res = TUPLE2(hp,am_error,am_match_limit);
+ } else if (rc == PCRE_ERROR_RECURSIONLIMIT) {
+ hp = HAlloc(p,3);
+ res = TUPLE2(hp,am_error,am_match_limit_recursion);
+ } else {
+ res = am_nomatch;
+ }
+ } else {
+ res = am_nomatch;
+ }
} else {
- ReturnInfo *ri = restartp->ret_info;
+ ReturnInfo *ri;
ReturnInfo defri = {RetIndex,0,{0}};
- if (ri == NULL) {
+
+ if (restartp->ret_info == NULL) {
ri = &defri;
+ } else {
+ ri = restartp->ret_info;
}
+
if (ri->type == RetNone) {
res = am_match;
} else if (ri->type == RetIndex){
@@ -577,6 +665,17 @@ static Eterm build_exec_return(Process *p, int rc, RestartContext *restartp, Ete
ri->num_spec * 2 * sizeof(Eterm));
for (i = 0; i < ri->num_spec; ++i) {
x = ri->v[i];
+ if (x < -1) {
+ int n = i-x+1;
+ int j;
+ for (j = i+1; j < ri->num_spec && j < n; ++j) {
+ if (restartp->ovector[(ri->v[j])*2] >= 0) {
+ x = ri->v[j];
+ break;
+ }
+ }
+ i = n-1;
+ }
if (x < rc && x >= 0) {
tmp_vect[n*2] = make_signed_integer(restartp->ovector[x*2],p);
tmp_vect[n*2+1] = make_signed_integer(restartp->ovector[x*2+1]-restartp->ovector[x*2],p);
@@ -658,6 +757,17 @@ static Eterm build_exec_return(Process *p, int rc, RestartContext *restartp, Ete
ri->num_spec * sizeof(Eterm));
for (i = 0; i < ri->num_spec; ++i) {
x = ri->v[i];
+ if (x < -1) {
+ int n = i-x+1;
+ int j;
+ for (j = i+1; j < ri->num_spec && j < n; ++j) {
+ if (restartp->ovector[(ri->v[j])*2] >= 0) {
+ x = ri->v[j];
+ break;
+ }
+ }
+ i = n-1;
+ }
if (x < rc && x >= 0) {
char *cp;
int len;
@@ -722,6 +832,49 @@ static Eterm build_exec_return(Process *p, int rc, RestartContext *restartp, Ete
*/
#define RINFO_SIZ(Num) (sizeof(ReturnInfo) + (sizeof(int) * (Num - 1)))
+#define PICK_INDEX(NameEntry) \
+ ((int) ((((unsigned) ((unsigned char *) (NameEntry))[0]) << 8) + \
+ ((unsigned) ((unsigned char *) (NameEntry))[1])))
+
+
+static void build_one_capture(const pcre *code, ReturnInfo **ri, int *sallocated, int has_dupnames, char *name)
+{
+ ReturnInfo *r = (*ri);
+ if (has_dupnames) {
+ /* Build a sequence of positions, starting with -size if
+ more than one, otherwise just put the index there... */
+ char *first,*last;
+ int esize = erts_pcre_get_stringtable_entries(code,name,&first,&last);
+ if (esize == PCRE_ERROR_NOSUBSTRING) {
+ r->v[r->num_spec - 1] = -1;
+ } else if(last == first) {
+ r->v[r->num_spec - 1] = PICK_INDEX(first);
+ } else {
+ int num = ((last - first) / esize) + 1;
+ int i;
+ ASSERT(num > 1);
+ r->v[r->num_spec - 1] = -num; /* A value less than -1 means
+ multiple indexes for same name */
+ for (i = 0; i < num; ++i) {
+ ++(r->num_spec);
+ if(r->num_spec > (*sallocated)) {
+ (*sallocated) += 10;
+ r = erts_realloc(ERTS_ALC_T_RE_SUBJECT, r,
+ RINFO_SIZ((*sallocated)));
+ }
+ r->v[r->num_spec - 1] = PICK_INDEX(first);
+ first += esize;
+ }
+ }
+ } else {
+ /* Use the faster binary search if no duplicate names are present */
+ if ((r->v[r->num_spec - 1] = erts_pcre_get_stringnumber(code,name)) ==
+ PCRE_ERROR_NOSUBSTRING) {
+ r->v[r->num_spec - 1] = -1;
+ }
+ }
+ *ri = r;
+}
static ReturnInfo *
build_capture(Eterm capture_spec[CAPSPEC_SIZE], const pcre *code)
@@ -770,13 +923,58 @@ build_capture(Eterm capture_spec[CAPSPEC_SIZE], const pcre *code)
}
ri->v[ri->num_spec - 1] = 0;
break;
+ case am_all_names:
+ {
+ int rc,i,top;
+ int entrysize;
+ unsigned char *nametable, *last = NULL;
+ int has_dupnames;
+ unsigned long options;
+
+ if (erts_pcre_fullinfo(code, NULL, PCRE_INFO_OPTIONS, &options) != 0)
+ goto error;
+ if ((rc = erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top)) != 0)
+ goto error;
+ if (top <= 0) {
+ ri->num_spec = 0;
+ ri->type = RetNone;
+ break;
+ }
+ if (erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize) != 0)
+ goto error;
+ if (erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable) != 0)
+ goto error;
+
+ has_dupnames = ((options & PCRE_DUPNAMES) != 0);
+
+ for(i=0;i<top;++i) {
+ if (last == NULL || !has_dupnames || strcmp((char *) last+2,(char *) nametable+2)) {
+ ASSERT(ri->num_spec >= 0);
+ ++(ri->num_spec);
+ if(ri->num_spec > sallocated) {
+ sallocated += 10;
+ ri = erts_realloc(ERTS_ALC_T_RE_SUBJECT, ri, RINFO_SIZ(sallocated));
+ }
+ if (has_dupnames) {
+ /* This could be more effective, we actually have
+ the names and could fill in the vector
+ immediately. Now we lookup the name again. */
+ build_one_capture(code,&ri,&sallocated,has_dupnames,(char *) nametable+2);
+ } else {
+ ri->v[ri->num_spec - 1] = PICK_INDEX(nametable);
+ }
+ }
+ last = nametable;
+ nametable += entrysize;
+ }
+ break;
+ }
default:
if (is_list(capture_spec[CAPSPEC_VALUES])) {
for(l=capture_spec[CAPSPEC_VALUES];is_list(l);l = CDR(list_val(l))) {
int x;
Eterm val = CAR(list_val(l));
- if (ri->num_spec < 0)
- ri->num_spec = 0;
+ ASSERT(ri->num_spec >= 0);
++(ri->num_spec);
if(ri->num_spec > sallocated) {
sallocated += 10;
@@ -785,6 +983,11 @@ build_capture(Eterm capture_spec[CAPSPEC_SIZE], const pcre *code)
if (term_to_int(val,&x)) {
ri->v[ri->num_spec - 1] = x;
} else if (is_atom(val) || is_binary(val) || is_list(val)) {
+ int has_dupnames;
+ unsigned long options;
+ if (erts_pcre_fullinfo(code, NULL, PCRE_INFO_OPTIONS, &options) != 0)
+ goto error;
+ has_dupnames = ((options & PCRE_DUPNAMES) != 0);
if (is_atom(val)) {
Atom *ap = atom_tab(atom_val(val));
if ((ap->len + 1) > tmpbsiz) {
@@ -799,6 +1002,10 @@ build_capture(Eterm capture_spec[CAPSPEC_SIZE], const pcre *code)
tmpb[ap->len] = '\0';
} else {
ErlDrvSizeT slen;
+#ifdef DEBUG
+ int buffres;
+#endif
+
if (erts_iolist_size(val, &slen)) {
goto error;
}
@@ -810,15 +1017,15 @@ build_capture(Eterm capture_spec[CAPSPEC_SIZE], const pcre *code)
(tmpbsiz = slen + 1));
}
}
- if (erts_iolist_to_buf(val, tmpb, slen) != 0) {
- goto error;
- }
+
+#ifdef DEBUG
+ buffres =
+#endif
+ erts_iolist_to_buf(val, tmpb, slen);
+ ASSERT(buffres >= 0);
tmpb[slen] = '\0';
}
- if ((ri->v[ri->num_spec - 1] = erts_pcre_get_stringnumber(code,tmpb)) ==
- PCRE_ERROR_NOSUBSTRING) {
- ri->v[ri->num_spec - 1] = -1;
- }
+ build_one_capture(code,&ri,&sallocated,has_dupnames,tmpb);
} else {
goto error;
}
@@ -867,15 +1074,18 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
unsigned long loop_count;
Eterm capture[CAPSPEC_SIZE] = CAPSPEC_INIT;
int is_list_cap;
+ int match_limit = 0;
+ int match_limit_recursion = 0;
- if (parse_options(arg3,&comp_options,&options,&pflags,&startoffset,capture)
+ if (parse_options(arg3,&comp_options,&options,&pflags,&startoffset,capture,
+ &match_limit,&match_limit_recursion)
< 0) {
BIF_ERROR(p,BADARG);
}
is_list_cap = ((pflags & PARSE_FLAG_CAPTURE_OPT) &&
(capture[CAPSPEC_TYPE] == am_list));
- if (is_not_tuple(arg2) || (arityval(*tuple_val(arg2)) != 4)) {
+ if (is_not_tuple(arg2) || (arityval(*tuple_val(arg2)) != 5)) {
if (is_binary(arg2) || is_list(arg2) || is_nil(arg2)) {
/* Compile from textual RE */
ErlDrvSizeT slen;
@@ -885,6 +1095,9 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
const char *errstr = "";
int errofset = 0;
int capture_count;
+#ifdef DEBUG
+ int buffres;
+#endif
if (pflags & PARSE_FLAG_UNICODE &&
(!is_binary(arg2) || !is_binary(arg1) ||
@@ -897,18 +1110,32 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
}
expr = erts_alloc(ERTS_ALC_T_RE_TMP_BUF, slen + 1);
- if (erts_iolist_to_buf(arg2, expr, slen) != 0) {
- erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
- BIF_ERROR(p,BADARG);
- }
+
+#ifdef DEBUG
+ buffres =
+#endif
+ erts_iolist_to_buf(arg2, expr, slen);
+
+ ASSERT(buffres >= 0);
+
expr[slen]='\0';
result = erts_pcre_compile2(expr, comp_options, &errcode,
&errstr, &errofset, default_table);
if (!result) {
- erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
/* Compilation error gives badarg except in the compile
- function */
- BIF_ERROR(p,BADARG);
+ function or if we have PARSE_FLAG_REPORT_ERRORS */
+ if (pflags & PARSE_FLAG_REPORT_ERRORS) {
+ res = build_compile_result(p, am_error, result, errcode,
+ errstr, errofset,
+ (pflags &
+ PARSE_FLAG_UNICODE) ? 1 : 0,
+ 1, am_compile);
+ erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
+ BIF_RET(res);
+ } else {
+ erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
+ BIF_ERROR(p,BADARG);
+ }
}
if (pflags & PARSE_FLAG_GLOBAL) {
Eterm precompiled =
@@ -917,7 +1144,7 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
errstr, errofset,
(pflags &
PARSE_FLAG_UNICODE) ? 1 : 0,
- 0);
+ 0, NIL);
Eterm *hp,r;
erts_free(ERTS_ALC_T_RE_TMP_BUF, expr);
hp = HAlloc(p,4);
@@ -947,7 +1174,8 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
tp = tuple_val(arg2);
if (tp[1] != am_re_pattern || is_not_small(tp[2]) ||
- is_not_small(tp[3]) || is_not_binary(tp[4])) {
+ is_not_small(tp[3]) || is_not_small(tp[4]) ||
+ is_not_binary(tp[5])) {
BIF_ERROR(p,BADARG);
}
@@ -967,9 +1195,9 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
}
ovsize = 3*(unsigned_val(tp[2])+1);
- code_size = binary_size(tp[4]);
+ code_size = binary_size(tp[5]);
if ((code_tmp = (const pcre *)
- erts_get_aligned_binary_bytes(tp[4], &temp_alloc)) == NULL) {
+ erts_get_aligned_binary_bytes(tp[5], &temp_alloc)) == NULL) {
erts_free_aligned_binary_bytes(temp_alloc);
BIF_ERROR(p, BADARG);
}
@@ -994,6 +1222,16 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
restart.extra.restart_flags = 0;
restart.extra.loop_counter_return = &loop_count;
restart.ret_info = NULL;
+
+ if (pflags & PARSE_FLAG_MATCH_LIMIT) {
+ restart.extra.flags |= PCRE_EXTRA_MATCH_LIMIT;
+ restart.extra.match_limit = match_limit;
+ }
+
+ if (pflags & PARSE_FLAG_MATCH_LIMIT_RECURSION) {
+ restart.extra.flags |= PCRE_EXTRA_MATCH_LIMIT_RECURSION;
+ restart.extra.match_limit_recursion = match_limit_recursion;
+ }
if (pflags & PARSE_FLAG_CAPTURE_OPT) {
if ((restart.ret_info = build_capture(capture,restart.code)) == NULL) {
@@ -1002,7 +1240,7 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
BIF_ERROR(p,BADARG);
}
}
-
+
/* Optimized - if already in binary off heap, keep that and avoid
copying, also binary returns can be sub binaries in that case */
@@ -1029,6 +1267,9 @@ re_run(Process *p, Eterm arg1, Eterm arg2, Eterm arg3)
restart.subject = (char *) (pb->bytes+offset);
restart.flags |= RESTART_FLAG_SUBJECT_IN_BINARY;
} else {
+#ifdef DEBUG
+ int buffres;
+#endif
handle_iolist:
if (erts_iolist_size(arg1, &slength)) {
erts_free(ERTS_ALC_T_RE_SUBJECT, restart.ovector);
@@ -1040,24 +1281,30 @@ handle_iolist:
}
restart.subject = erts_alloc(ERTS_ALC_T_RE_SUBJECT, slength);
- if (erts_iolist_to_buf(arg1, restart.subject, slength) != 0) {
- erts_free(ERTS_ALC_T_RE_SUBJECT, restart.ovector);
- erts_free(ERTS_ALC_T_RE_SUBJECT, restart.code);
- erts_free(ERTS_ALC_T_RE_SUBJECT, restart.subject);
- if (restart.ret_info != NULL) {
- erts_free(ERTS_ALC_T_RE_SUBJECT, restart.ret_info);
- }
- BIF_ERROR(p,BADARG);
- }
+#ifdef DEBUG
+ buffres =
+#endif
+ erts_iolist_to_buf(arg1, restart.subject, slength);
+ ASSERT(buffres >= 0);
}
+ if (pflags & PARSE_FLAG_REPORT_ERRORS) {
+ restart.flags |= RESTART_FLAG_REPORT_MATCH_LIMIT;
+ }
#ifdef DEBUG
loop_count = 0xFFFFFFFF;
#endif
+
+ rc = erts_pcre_exec(restart.code, &(restart.extra), restart.subject,
+ slength, startoffset,
+ options, restart.ovector, ovsize);
+
+ if (rc == PCRE_ERROR_BADENDIANNESS || rc == PCRE_ERROR_BADMAGIC) {
+ cleanup_restart_context(&restart);
+ BIF_ERROR(p,BADARG);
+ }
- rc = erts_pcre_exec(restart.code, &(restart.extra), restart.subject, slength, startoffset,
- options, restart.ovector, ovsize);
ASSERT(loop_count != 0xFFFFFFFF);
BUMP_REDS(p, loop_count / LOOP_FACTOR);
if (rc == PCRE_ERROR_LOOP_LIMIT) {
@@ -1077,7 +1324,7 @@ handle_iolist:
arg2 /* To avoid GC of precompiled code, XXX: not utilized yet */,
magic_bin);
}
-
+
res = build_exec_return(p, rc, &restart, arg1);
cleanup_restart_context(&restart);
@@ -1149,6 +1396,120 @@ static BIF_RETTYPE re_exec_trap(BIF_ALIST_3)
BIF_RET(res);
}
+BIF_RETTYPE
+re_inspect_2(BIF_ALIST_2)
+{
+ Eterm *tp,*tmp_vec,*hp;
+ int i,top,j;
+ int entrysize;
+ unsigned char *nametable, *last,*name;
+ int has_dupnames;
+ unsigned long options;
+ int num_names;
+ Eterm res;
+ const pcre *code;
+ byte *temp_alloc = NULL;
+#ifdef DEBUG
+ int infores;
+#endif
+
+
+ if (is_not_tuple(BIF_ARG_1) || (arityval(*tuple_val(BIF_ARG_1)) != 5)) {
+ goto error;
+ }
+ tp = tuple_val(BIF_ARG_1);
+ if (tp[1] != am_re_pattern || is_not_small(tp[2]) ||
+ is_not_small(tp[3]) || is_not_small(tp[4]) ||
+ is_not_binary(tp[5])) {
+ goto error;
+ }
+ if (BIF_ARG_2 != am_namelist) {
+ goto error;
+ }
+ if ((code = (const pcre *)
+ erts_get_aligned_binary_bytes(tp[5], &temp_alloc)) == NULL) {
+ goto error;
+ }
+
+ /* OK, so let's try to get some info */
+
+ if (erts_pcre_fullinfo(code, NULL, PCRE_INFO_OPTIONS, &options) != 0)
+ goto error;
+
+#ifdef DEBUG
+ infores =
+#endif
+ erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMECOUNT, &top);
+
+ ASSERT(infores == 0);
+
+ if (top <= 0) {
+ hp = HAlloc(BIF_P, 3);
+ res = TUPLE2(hp,am_namelist,NIL);
+ erts_free_aligned_binary_bytes(temp_alloc);
+ BIF_RET(res);
+ }
+#ifdef DEBUG
+ infores =
+#endif
+ erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMEENTRYSIZE, &entrysize);
+
+ ASSERT(infores == 0);
+
+#ifdef DEBUG
+ infores =
+#endif
+ erts_pcre_fullinfo(code, NULL, PCRE_INFO_NAMETABLE, &nametable);
+
+ ASSERT(infores == 0);
+
+ has_dupnames = ((options & PCRE_DUPNAMES) != 0);
+ /* First, count the names */
+ num_names = 0;
+ last = NULL;
+ name = nametable;
+ for(i=0;i<top;++i) {
+ if (last == NULL || !has_dupnames || strcmp((char *) last+2,
+ (char *) name+2)) {
+ ++num_names;
+ }
+ last = name;
+ name += entrysize;
+ }
+ tmp_vec = erts_alloc(ERTS_ALC_T_RE_TMP_BUF,
+ num_names * sizeof(Eterm));
+ /* Re-iterate and fill tmp_vec */
+ last = NULL;
+ name = nametable;
+ j = 0;
+ for(i=0;i<top;++i) {
+ if (last == NULL || !has_dupnames || strcmp((char *) last+2,
+ (char *) name+2)) {
+ tmp_vec[j++] = new_binary(BIF_P, (byte *) name+2, strlen((char *) name+2));
+ }
+ last = name;
+ name += entrysize;
+ }
+ ASSERT(j == num_names);
+ hp = HAlloc(BIF_P, 3+2*j);
+ res = NIL;
+ for(i = j-1 ;i >= 0; --i) {
+ res = CONS(hp,tmp_vec[i],res);
+ hp += 2;
+ }
+ res = TUPLE2(hp,am_namelist,res);
+ erts_free_aligned_binary_bytes(temp_alloc);
+ erts_free(ERTS_ALC_T_RE_TMP_BUF, tmp_vec);
+ BIF_RET(res);
+
+ error:
+ /* tmp_vec never allocated when we reach here */
+ erts_free_aligned_binary_bytes(temp_alloc);
+ BIF_ERROR(BIF_P,BADARG);
+}
+
+
+
diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c
index bd2be7afca..c71f8f0712 100644
--- a/erts/emulator/beam/utils.c
+++ b/erts/emulator/beam/utils.c
@@ -3021,6 +3021,14 @@ buf_to_intlist(Eterm** hpp, const char *buf, size_t len, Eterm tail)
** Return remaining bytes in buffer on success
** ERTS_IOLIST_TO_BUF_OVERFLOW on overflow
** ERTS_IOLIST_TO_BUF_TYPE_ERROR on type error (including that result would not be a whole number of bytes)
+**
+** Note!
+** Do not detect indata errors in this fiunction that are not detected by erts_iolist_size!
+**
+** A caller should be able to rely on a successful return from erts_iolist_to_buf
+** if erts_iolist_size is previously successfully called and erts_iolist_to_buf
+** is called with a buffer at least as large as the value given by erts_iolist_size.
+**
*/
ErlDrvSizeT erts_iolist_to_buf(Eterm obj, char* buf, ErlDrvSizeT alloced_len)
@@ -3127,6 +3135,11 @@ ErlDrvSizeT erts_iolist_to_buf(Eterm obj, char* buf, ErlDrvSizeT alloced_len)
/*
* Return 0 if successful, and non-zero if unsuccessful.
+ *
+ * It is vital that if erts_iolist_to_buf would return an error for
+ * any type of term data, this function should do so as well.
+ * Any input term error detected in erts_iolist_to_buf should also
+ * be detected in this function!
*/
int erts_iolist_size(Eterm obj, ErlDrvSizeT* sizep)
{