implement the compile option locale

README.rst

@@ -50,7 +50,7 @@ regex
 
 ::
 
-	new OBJ = regex(STRING pattern, BOOL allow_empty_class=0, BOOL anchored=0, ENUM {ANYCRLF,UNICODE} bsr=0, BOOL alt_bsux=0, BOOL alt_circumflex=0, BOOL alt_verbnames=0, BOOL caseless=0, BOOL dotall=0, BOOL dollar_endonly=0, BOOL dupnames=0, BOOL extended=0, BOOL firstline=0, STRING locale=0, BOOL match_unset_backref=0, INT max_pattern_len=0, BOOL multiline=0, BOOL never_backslash_c=0, BOOL never_ucp=0, BOOL never_utf=0, ENUM {CR,LF,CRLF,ANYCRLF,ANY} newline=0, BOOL no_auto_capture=0, BOOL no_auto_possess=0, BOOL no_dotstar_anchor=0, BOOL no_start_optimize=0, BOOL no_utf_check=0, INT parens_nest_limit=0, BOOL ucp=0, BOOL ungreedy=0, BOOL use_offset_limit=0, BOOL utf=0)
+	new OBJ = regex(STRING pattern, BOOL allow_empty_class=0, BOOL anchored=0, ENUM {ANYCRLF,UNICODE} bsr=0, BOOL alt_bsux=0, BOOL alt_circumflex=0, BOOL alt_verbnames=0, BOOL caseless=0, BOOL dollar_endonly=0, BOOL dotall=0, BOOL dupnames=0, BOOL extended=0, BOOL firstline=0, STRING locale=0, BOOL match_unset_backref=0, INT max_pattern_len=0, BOOL multiline=0, BOOL never_backslash_c=0, BOOL never_ucp=0, BOOL never_utf=0, ENUM {CR,LF,CRLF,ANYCRLF,ANY} newline=0, BOOL no_auto_capture=0, BOOL no_auto_possess=0, BOOL no_dotstar_anchor=0, BOOL no_start_optimize=0, BOOL no_utf_check=0, INT parens_nest_limit=0, BOOL ucp=0, BOOL ungreedy=0, BOOL use_offset_limit=0, BOOL utf=0)
 
 # XXX options for dfa_match, jit fast path, start_offset
 # XXX option to make saving the match ctx with PRIV_CALL optional

configure.ac

@@ -93,6 +93,12 @@ PKG_CHECK_MODULES([PCRE2], [libpcre2-8])
 AC_SUBST([PCRE2_CFLAGS])
 AC_SUBST([PCRE2_LIBS])
 
+if test ! -z `locale -a | grep '^fr_FR$'`; then
+   AC_SUBST([LOCALE_FR_VTC], [locale_fr.vtc.yes])
+else
+   AC_SUBST([LOCALE_FR_VTC], [locale_fr.vtc.no])
+fi
+
 # --enable-stack-protector
 AC_ARG_ENABLE(stack-protector,
        AS_HELP_STRING([--enable-stack-protector],[enable stack protector (default is YES)]),

src/Makefile.am

@@ -21,14 +21,18 @@ vcc_if.c: vcc_if.h
 vcc_if.h vmod_pcre2.man.rst: @VMODTOOL@ $(top_srcdir)/src/vmod_pcre2.vcc
 	@VMODTOOL@ $(top_srcdir)/src/vmod_pcre2.vcc
 
-VMOD_TESTS = $(top_srcdir)/src/tests/*.vtc
+VMOD_TESTS = $(top_srcdir)/src/tests/*.vtc $(top_srcdir)/src/tests/locale_fr.vtc
 .PHONY: $(VMOD_TESTS)
 
-$(top_srcdir)/src/tests/*.vtc: libvmod_pcre2.la
+locale_fr:
+	cp $(top_srcdir)/src/tests/@LOCALE_FR_VTC@ $(top_srcdir)/src/tests/locale_fr.vtc
+
+$(top_srcdir)/src/tests/*.vtc $(top_srcdir)/src/tests/locale_fr.vtc: locale_fr libvmod_pcre2.la
 	PATH=@LIBVARNISHAPI_SBINDIR@:$$PATH \
 	@VARNISHTEST@ -Dvmod_topbuild=$(abs_top_builddir) $@
 
 check: $(VMOD_TESTS)
+	rm $(top_srcdir)/src/tests/locale_fr.vtc
 
 EXTRA_DIST = \
 	vmod_pcre2.vcc \
@@ -38,4 +42,5 @@ CLEANFILES = \
 	$(builddir)/vcc_if.c \
 	$(builddir)/vcc_if.h \
 	$(builddir)/vmod_pcre2.rst \
-	$(builddir)/vmod_pcre2.man.rst
+	$(builddir)/vmod_pcre2.man.rst \
+	$(top_srcdir)/src/tests/locale_fr.vtc

src/tests/compile_opts.vtc

@@ -462,3 +462,62 @@ client c1 {
 	expect resp.http.r3-1 == "true"
 	expect resp.http.r3-2 == "false"
 } -run
+
+# Test the locale option with the POSIX locale, which is the only
+# locale we can presume exists (at least on POSIX-conforming
+# systems). AFAICT the results are no different at all from the pcre2
+# standard character tables, but at least this tests against obvious
+# bugs.
+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("(?i)abc \d \D \w \W \s \S \babc\B",
+		    		     locale="POSIX");
+		new r2 = pcre2.regex("[[:alnum:]]+ [[:alpha:]]+ [[:ascii:]]+",
+		    		     locale="POSIX");
+		new r3 = pcre2.regex("[[:blank:]]+_[[:cntrl:]]+ [[:digit:]]+",
+		    		     locale="POSIX");
+		new r4 = pcre2.regex("[[:graph:]]+ [[:lower:]]+ [[:print:]]+",
+		    		     locale="POSIX");
+		new r5 = pcre2.regex("[[:space:]]+_[[:upper:]]+ [[:word:]]+",
+		    		     locale="POSIX");
+		new r6 = pcre2.regex("[[:xdigit:]]+", locale="POSIX");
+	}
+
+	sub vcl_recv {
+		return(synth(200));
+	}
+
+	sub vcl_synth {
+		set resp.http.r1 = r1.match("aBc 0 x x !   x AbCx");
+		set resp.http.r2 = r2.match("a1 bb ~~");
+		set resp.http.r3 = r3.match({" 	_
 123456789"});
+		set resp.http.r4 = r4.match({"ab12!"§$%&/ xyz !"§$%&/"});
+		set resp.http.r5 = r5.match({" 	_ABCXYZ _abcXYZ987"});
+		set resp.http.r6 = r6.match("0123456789abcdef");
+	}
+}
+
+client c1 {
+	txreq
+	rxresp
+	expect resp.status == "200"
+	expect resp.http.r1 == "true"
+	expect resp.http.r2 == "true"
+	expect resp.http.r3 == "true"
+	expect resp.http.r4 == "true"
+	expect resp.http.r5 == "true"
+	expect resp.http.r6 == "true"
+} -run
+
+# Pretty sure that this locale will not exist anywhere
+varnish v1 -errvcl {vmod pcre2 error: Cannot set locale lPb2PZDb4KcfbKJPSww1cm6omvoF in r constructor} {
+	import pcre2 from "${vmod_topbuild}/src/.libs/libvmod_pcre2.so";
+	backend b { .host = "${bad_ip}"; }
+
+	sub vcl_init {
+		new r = pcre2.regex("", locale="lPb2PZDb4KcfbKJPSww1cm6omvoF");
+	}
+}

src/tests/locale_fr.vtc.no

+varnishtest "not testing compile option locale=fr_FR"
+
+# If this test is run as locale_fr.vtc, then compile option
+# locale="fr_FR" is not being tested for this build, because autoconf
+# could not verify the existence of that locale on this host.
+#
+# Check the command 'locale -a' if you think this is an error -- you
+# may be able to generate the locale on your platform and try again.
+#
+# Simpler tests for the locale option are run in compile_opt.vtc.
+#
+# If the locale had been found, this test would have executed more
+# detailed tests for option locale, following the self-tests for
+# libpcre2.

src/tests/locale_fr.vtc.yes

+# -*-mode: vcl; coding: iso-8859-1 -*-
+
+# The non-ASCII letters in this source correspond to ISO-8859-1 (aka
+# ISO-Latin-1). You might want to set up your editor accordingly.
+
+varnishtest "compile option locale using a non-standard locale"
+
+# tests from pcre2 testoutput3
+
+# XXX add the tests that require backrefs and pcre2_pattern_info
+
+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("^[\w]+", locale="fr_FR");
+		new r2 = pcre2.regex("^[\w]+");
+		new r3 = pcre2.regex("^[\W]+", locale="fr_FR");
+		new r4 = pcre2.regex("^[\W]+");
+		new r5 = pcre2.regex("[\b]", locale="fr_FR");
+		new r6 = pcre2.regex("[\b]");
+		new r7 = pcre2.regex("^\w+", locale="fr_FR");
+		new r8 = pcre2.regex("^\w+");
+		new r9 = pcre2.regex("(.+)\b(.+)", locale="fr_FR");
+		new r10 = pcre2.regex("(.+)\b(.+)");
+		new r11 = pcre2.regex("cole", locale="fr_FR", caseless=true);
+		new r12 = pcre2.regex("cole", caseless=true);
+		new r13 = pcre2.regex("^[\xc8-\xc9]", locale="fr_FR",
+					caseless=true);
+		new r14 = pcre2.regex("^[\xc8-\xc9]", locale="fr_FR");
+	}
+
+	sub vcl_recv {
+		return(synth(200));
+	}
+
+	sub vcl_synth {
+		set resp.http.r1 = r1.match("cole");
+		set resp.http.r2 = r2.match("cole");
+		set resp.http.r3 = r3.match("cole");
+		set resp.http.r4 = r4.match("cole");
+		set resp.http.r5-1 = r5.match("");
+		set resp.http.r6-1 = r6.match("");
+		set resp.http.r5-2 = r5.match("a");
+		set resp.http.r6-2 = r6.match("a");
+		set resp.http.r7 = r7.match("cole");
+		set resp.http.r8 = r8.match("cole");
+		set resp.http.r9 = r9.match("cole");
+		set resp.http.r10 = r10.match("cole");
+		set resp.http.r11-1 = r11.match("cole");
+		set resp.http.r12-1 = r12.match("cole");
+		set resp.http.r11-2 = r11.match("cole");
+		set resp.http.r12-2 = r12.match("cole");
+		set resp.http.r13-1 = r13.match("cole");
+		set resp.http.r13-2 = r13.match("cole");
+		set resp.http.r14-1 = r14.match("cole");
+		set resp.http.r14-2 = r14.match("cole");
+	}
+} -start
+
+client c1 {
+	txreq
+	rxresp
+	expect resp.status == "200"
+	expect resp.http.r1 == "true"
+	expect resp.http.r2 == "false"
+	expect resp.http.r3 == "false"
+	expect resp.http.r4 == "true"
+	expect resp.http.r5-1 == "true"
+	expect resp.http.r5-2 == "false"
+	expect resp.http.r6-1 == "true"
+	expect resp.http.r6-2 == "false"
+	expect resp.http.r7 == "true"
+	expect resp.http.r8 == "false"
+	expect resp.http.r9 == "false"
+	expect resp.http.r10 == "true"
+	expect resp.http.r11-1 == "true"
+	expect resp.http.r11-2 == "true"
+	expect resp.http.r12-1 == "true"
+	expect resp.http.r12-2 == "false"
+	expect resp.http.r13-1 == "true"
+	expect resp.http.r13-2 == "true"
+	expect resp.http.r14-1 == "true"
+	expect resp.http.r14-2 == "false"
+} -run

src/vmod_pcre2.c

@@ -29,6 +29,7 @@
 
 #include <stdint.h>
 #include <string.h>
+#include <locale.h>
 
 #include "vcl.h"
 #include "cache/cache.h"
@@ -248,9 +249,25 @@ vmod_regex__init(VRT_CTX, struct vmod_pcre2_regex **regexp,
 		AZ(pcre2_set_bsr(ccontext, val));
 	}
 	if (locale != NULL) {
-		VERR(ctx, "locales not implemented yet in %s constructor",
-		     vcl_name);
-		return;
+		const unsigned char *tables;
+		char *saved_lc = setlocale(LC_CTYPE, NULL);
+
+		AN(saved_lc);
+		if (setlocale(LC_CTYPE, locale) == NULL) {
+			AN(setlocale(LC_CTYPE, saved_lc));
+			VERR(ctx, "Cannot set locale %s in %s constructor",
+			     locale, vcl_name);
+			return;
+		}
+		tables = pcre2_maketables(NULL);
+		if (tables == NULL) {
+			AN(setlocale(LC_CTYPE, saved_lc));
+			VERRNOMEM(ctx, "Creating tables for locale %s in "
+				  "%s constructor", locale, vcl_name);
+			return;
+		}
+		AZ(pcre2_set_character_tables(ccontext, tables));
+		AN(setlocale(LC_CTYPE, saved_lc));
 	}
 	if (max_pattern_len != 0)
 		AZ(pcre2_set_max_pattern_length(ccontext,