implement range restrictions on match options match_limit, offset_limit

and recursion_limit, and test compile option use_offset_limit

README.rst

@@ -61,7 +61,7 @@ regex.match
 
 ::
 
-	BOOL regex.match(PRIV_CALL, PRIV_TASK, STRING, INT len=0, BOOL anchored=0, INT match_limit=0, INT offset_limit=0, BOOL notbol=0, BOOL noteol=0, BOOL notempty=0, BOOL notempty_atstart=0, BOOL no_jit=0, BOOL no_utf_check=0, ENUM {HARD,SOFT} partial=0, INT recursion_limit=0)
+	BOOL regex.match(PRIV_CALL, PRIV_TASK, STRING subject, INT len=0, BOOL anchored=0, INT match_limit=0, INT offset_limit=0, BOOL notbol=0, BOOL noteol=0, BOOL notempty=0, BOOL notempty_atstart=0, BOOL no_jit=0, BOOL no_utf_check=0, ENUM {HARD,SOFT} partial=0, INT recursion_limit=0)
 
 .. _func_config_bool:
 

src/tests/compile_opts.vtc

@@ -12,16 +12,18 @@ varnishtest "compile options"
 # extended option is tested in compile.vtc and match.vtc
 
 # XXX
-# - tests for alt_circumflex will require backrefs
+# - tests for alt_circumflex, no_auto_capture and ungreedy require
+#   backrefs
 # - tests for alt_verbnames will require retrieving the mark
 # - auto_callout not implemented yet
 # - test for dupnames will require capturing named refs
 # - how to test never_backslash_c? pcre2 testdata only mentions it in
 #   one test, and vtc can't parse the pattern, since it has more {'s
 #   than }'s.
-# - test for no_auto_capture requires backrefs
 # - test for no_auto_possess, no_dotstar_anchor and no_start_optimize
 #   require callouts and/or pattern_info
+# - some of the tests for use_offset_limit should use the startoffset
+#   option for match() (not yet implemented)
 
 # allow_empty_class
 varnish v1 -vcl {
@@ -777,3 +779,64 @@ varnish v1 -errvcl {vmod pcre2 error: parens_nest_limit (4294967296) out of rang
 		new r = pcre2.regex("a", parens_nest_limit=4294967296);
 	}
 }
+
+# use_offset_limit
+varnish v1 -vcl {
+	import pcre2 from "${vmod_topbuild}/src/.libs/libvmod_pcre2.so";
+	backend b { .host = "${bad_ip}"; }
+
+	sub vcl_init {
+		new r1 = pcre2.regex("abc", use_offset_limit=true);
+		new r2 = pcre2.regex("(?<=abc)", use_offset_limit=true);
+		new r3 = pcre2.regex("abc");
+		new r4 = pcre2.regex("^\w", use_offset_limit=true,
+		                     multiline=true);
+	}
+
+	sub vcl_recv {
+		return(synth(200));
+	}
+
+	sub vcl_synth {
+		set resp.http.r1-1 = r1.match("1234abcde", offset_limit=100);
+		set resp.http.r1-2 = r1.match("1234abcde", offset_limit=9);
+		set resp.http.r1-3 = r1.match("1234abcde", offset_limit=4);
+#		set resp.http.r1-4 = r1.match("1234abcde", offset_limit=4,
+#		    startoffset=4);
+#		set resp.http.r1-5 = r1.match("1234abcde", offset_limit=4,
+#		    startoffset=5);
+		set resp.http.r1-6 = r1.match("1234abcde", offset_limit=3);
+		set resp.http.r2-1 = r2.match("1234abc", offset_limit=7);
+		set resp.http.r2-2 = r2.match("1234abc", offset_limit=6);
+		set resp.http.r3 = r3.match("1234abcde", offset_limit=4);
+		set resp.http.r4-1 = r4.match(offset_limit=3, subject={"
+..
+aa"});
+		set resp.http.r4-2 = r4.match(offset_limit=4, subject={"
+..
+aa"});
+	}
+}
+
+logexpect l1 -v v1 -d 0 -g vxid -q "VCL_Error" {
+        expect 0 * Begin req
+        expect * = VCL_Error "^vmod pcre2 error: in r3.match..: offset limit set without PCRE2_USE_OFFSET_LIMIT"
+        expect * = End
+} -start
+
+client c1 {
+	txreq
+	rxresp
+	expect resp.status == "200"
+	expect resp.http.r1-1 == "true"
+	expect resp.http.r1-2 == "true"
+	expect resp.http.r1-3 == "true"
+	expect resp.http.r1-6 == "false"
+	expect resp.http.r2-1 == "true"
+	expect resp.http.r2-2 == "false"
+	expect resp.http.r3 == "false"
+	expect resp.http.r4-1 == "false"
+	expect resp.http.r4-2 == "true"
+} -run
+
+logexpect l1 -wait

src/vmod_pcre2.c

@@ -445,6 +445,29 @@ vmod_regex_match(VRT_CTX, struct vmod_pcre2_regex *regex,
 	if (mctx_call->priv == NULL) {
 		pcre2_match_context *mctx;
 
+		/* XXX DRY */
+		if (match_limit < 0 || match_limit > UINT32_MAX) {
+			VERR(ctx, "match_limit (%lld) out of range in "
+			     "%s.match() (must be >= 0 and <= %" PRIu32 ")",
+			     (long long)match_limit, regex->vcl_name,
+			     UINT32_MAX);
+			return 0;
+		}
+		if (offset_limit < 0 || offset_limit > UINT32_MAX) {
+			VERR(ctx, "offset_limit (%lld) out of range in "
+			     "%s.match() (must be >= 0 and <= %" PRIu32 ")",
+			     (long long)offset_limit, regex->vcl_name,
+			     UINT32_MAX);
+			return 0;
+		}
+		if (recursion_limit < 0 || recursion_limit > UINT32_MAX) {
+			VERR(ctx, "recursion_limit (%lld) out of range in "
+			     "%s.match() (must be >= 0 and <= %" PRIu32 ")",
+			     (long long)recursion_limit, regex->vcl_name,
+			     UINT32_MAX);
+			return 0;
+		}
+
 		if ((mctx = pcre2_match_context_create(NULL)) == NULL) {
 			VERRNOMEM(ctx, "creating match context in %s.match()",
 				  regex->vcl_name);

src/vmod_pcre2.vcc

@@ -32,11 +32,12 @@ $Object regex(STRING pattern, BOOL allow_empty_class=0, BOOL anchored=0,
 
 # XXX options for dfa_match, jit fast path, start_offset
 # XXX option to make saving the match ctx with PRIV_CALL optional
-$Method BOOL .match(PRIV_CALL, PRIV_TASK, STRING, INT len=0, BOOL anchored=0,
-                    INT match_limit=0, INT offset_limit=0, BOOL notbol=0,
-		    BOOL noteol=0, BOOL notempty=0, BOOL notempty_atstart=0,
-		    BOOL no_jit=0, BOOL no_utf_check = 0,
-		    ENUM {HARD, SOFT} partial=0, INT recursion_limit=0)
+$Method BOOL .match(PRIV_CALL, PRIV_TASK, STRING subject, INT len=0,
+                    BOOL anchored=0, INT match_limit=0, INT offset_limit=0,
+		    BOOL notbol=0, BOOL noteol=0, BOOL notempty=0,
+		    BOOL notempty_atstart=0, BOOL no_jit=0,
+		    BOOL no_utf_check = 0, ENUM {HARD, SOFT} partial=0,
+		    INT recursion_limit=0)
 
 $Function BOOL config_bool(ENUM {JIT, STACKRECURSE, UNICODE})