Add tests for sub() when Unicode is enabled.

src/Makefile.am

@@ -21,22 +21,29 @@ 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 = $(builddir)/tests/*.vtc $(builddir)/tests/locale_fr.vtc $(builddir)/tests/compile_utf8.vtc
+VMOD_TESTS = $(builddir)/tests/*.vtc \
+	$(builddir)/tests/locale_fr.vtc \
+	$(builddir)/tests/compile_utf8.vtc \
+	$(builddir)/tests/sub_utf8.vtc
 .PHONY: $(VMOD_TESTS)
 
 locale_fr:
-	cp $(top_srcdir)/src/tests/@LOCALE_FR_VTC@ $(top_srcdir)/src/tests/locale_fr.vtc
+	cp $(builddir)/tests/@LOCALE_FR_VTC@ $(builddir)/tests/locale_fr.vtc
 
 utf8:
-	cp $(builddir)/tests/compile_utf8.vtc.@PCRE2_UTF8@ $(builddir)/tests/compile_utf8.vtc
+	cp $(builddir)/tests/compile_utf8.vtc.@PCRE2_UTF8@ \
+	   $(builddir)/tests/compile_utf8.vtc
+	cp $(builddir)/tests/sub_utf8.vtc.@PCRE2_UTF8@ \
+	   $(builddir)/tests/sub_utf8.vtc
 
-$(builddir)/tests/*.vtc $(builddir)/tests/locale_fr.vtc $(builddir)/tests/compile_utf8.vtc: locale_fr utf8 libvmod_pcre2.la
+$(VMOD_TESTS): locale_fr utf8 libvmod_pcre2.la
 	PATH=@LIBVARNISHAPI_SBINDIR@:$$PATH \
 	@VARNISHTEST@ -Dvmod_topbuild=$(abs_top_builddir) $@
 
 check: $(VMOD_TESTS)
-	rm $(top_srcdir)/src/tests/locale_fr.vtc
+	rm $(builddir)/tests/locale_fr.vtc
 	rm $(builddir)/tests/compile_utf8.vtc
+	rm $(builddir)/tests/sub_utf8.vtc
 
 EXTRA_DIST = \
 	vmod_pcre2.vcc \

src/tests/sub_utf8.vtc.no

+varnishtest "not testing substitutions with UTF-8"
+
+# If Unicode is disabled for libpcre2, then this test is run as
+# sub_utf8.vtc, and does nothing.

src/tests/sub_utf8.vtc.yes

+# -*-mode: vcl; coding: raw-text;-*-
+
+varnishtest "substitutions with UTF-8"
+
+# This test is only run if Unicode is enabled for libpcre2.
+
+# Tests from PCRE2 testoutput5 and testoutput10
+
+# Test r5 uses the extended sub syntax, including strings of the form
+# ${...}. varnishtest interprets them as macros in a vtc script and
+# rejects the test, so we include the line from another file.
+
+varnish v1 -vcl {
+	import pcre2 from "${vmod_topbuild}/src/.libs/libvmod_pcre2.so";
+	backend b { .host = "${bad_ip}"; }
+
+	sub vcl_init {
+		new r1 = pcre2.regex("ábc", utf=true);
+		new r2 = pcre2.regex("(?<=abc)(|def)", utf=true);
+		new r3 = pcre2.regex("abcd", utf=true);
+		new r4
+		    = pcre2.regex("a(\x{e0}\x{101})(\x{c0}\x{102})", utf=true);
+		new r5
+		    = pcre2.regex("((?<digit>\d)|(?<letter>\p{L}))", utf=true);
+		new r6 = pcre2.regex("abc", utf=true);
+	}
+
+	sub vcl_recv {
+		return(synth(200));
+	}
+
+	sub vcl_synth {
+		set resp.http.r1 = r1.sub("123ábc123", "XሴZ");
+		set resp.http.r2
+		    = r2.sub("123abcáyzabcdef789abcሴqr",
+		             "<$0>", suball=true);
+		set resp.http.r3
+		    = r3.sub("abcd", "x\x{824}y\o{3333}z(\Q12\$34$$\x34\E5$$)",
+		    	     sub_extended=true);
+		set resp.http.r4 = r4.sub("aàāÀĂ",
+		"a\u$1\U$1\E$1\l$2\L$2\Eab\U\x{e0}\x{101}\L\x{d0}\x{160}\EDone",
+		    	                  sub_extended=true);
+
+		include "${vmod_topbuild}/src/tests/sub_utf8_macro.vcl";
+
+		set resp.http.r6 = r6.sub("abc", "");
+
+		return(deliver);
+	}
+} -start
+
+client c1 {
+	txreq
+	rxresp
+	expect resp.status == "200"
+	expect resp.http.r1 == "123XሴZ123"
+	expect resp.http.r2 == "123abc<>áyzabc<><def>789abc<>ሴqr"
+	expect resp.http.r3 == {xࠤyۛz(12\$34$$\x345$)}
+	expect resp.http.r4 == {aÀāÀĀàāàĂàăabÀĀðšDone}
+	expect resp.http.r5 == "<not digit; letter><not digit; letter><digit; not a letter><digit; not a letter><not digit; letter><not digit; letter><not digit; letter>"
+	expect resp.http.r6 == ""
+} -run

src/tests/sub_utf8_macro.vcl

+# This line is included from sub_utf8.vtc.yes, run as sub_utf8.vtc
+# when Unicode is enabled for libpcre2. The extended syntax for pcre2
+# substitutions includes the form ${...}, which varnishtest always
+# interprets as a macro in a vtc script. So this snippet is included
+# as a workaround.
+set resp.http.r5 = r5.sub("ab12cde", suball=true, replacement=
+		"<${digit:+digit; :not digit; }${letter:+letter:not a letter}>",
+		    	  sub_extended=true);