Content-Transfer-Encoding: 7bit
Content-Type: multipart/mixed; boundary="_----------=_1587212471197570"
MIME-Version: 1.0
Date: Sat, 18 Apr 2020 12:21:11 +0000
From: "Roland Illig" <rillig@netbsd.org>
Subject: CVS commit: pkgsrc
To: pkgsrc-changes@NetBSD.org
Reply-To: rillig@netbsd.org
Message-Id: <20200418122111.22944FB27@cvs.NetBSD.org>
Sender: pkgsrc-changes-owner@NetBSD.org
Precedence: bulk

This is a multi-part message in MIME format.

--_----------=_1587212471197570
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset="US-ASCII"

Module Name:	pkgsrc
Committed By:	rillig
Date:		Sat Apr 18 12:21:11 UTC 2020

Modified Files:
	pkgsrc/mk: subst.mk
	pkgsrc/regress/infra-unittests: subst.sh

Log Message:
mk/subst.mk: avoid undefined behavior in regular expressions

https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap09.html
says in section 9.3.2 BRE Ordinary Characters that only very few
characters may be preceded with a backslash.

As a side effect, this change allows parentheses in the variable names
listed in SUBST_VARS (even if that will never happen in practice).

The reason that the regression test had not replaced VAR.[] before was
simply that this variable had not been listed in SUBST_VARS.


To generate a diff of this commit:
cvs rdiff -u -r1.76 -r1.77 pkgsrc/mk/subst.mk
cvs rdiff -u -r1.15 -r1.16 pkgsrc/regress/infra-unittests/subst.sh

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.


--_----------=_1587212471197570
Content-Disposition: inline
Content-Length: 2413
Content-Transfer-Encoding: binary
Content-Type: text/x-diff; charset=us-ascii

Modified files:

Index: pkgsrc/mk/subst.mk
diff -u pkgsrc/mk/subst.mk:1.76 pkgsrc/mk/subst.mk:1.77
--- pkgsrc/mk/subst.mk:1.76	Sat Apr 18 11:42:34 2020
+++ pkgsrc/mk/subst.mk	Sat Apr 18 12:21:10 2020
@@ -1,4 +1,4 @@
-# $NetBSD: subst.mk,v 1.76 2020/04/18 11:42:34 rillig Exp $
+# $NetBSD: subst.mk,v 1.77 2020/04/18 12:21:10 rillig Exp $
 #
 # The subst framework replaces text in one or more files in the WRKSRC
 # directory. Packages can define several ``classes'' of replacements.
@@ -128,7 +128,7 @@ SUBST_FILTER_CMD.${class}?=	LC_ALL=C ${S
 SUBST_VARS.${class}?=		# none
 SUBST_MESSAGE.${class}?=	Substituting "${class}" in ${SUBST_FILES.${class}}
 .  for v in ${SUBST_VARS.${class}}
-SUBST_FILTER_CMD.${class}+=	-e s,@${v:C|[^A-Za-z0-9_]|\\\\&|gW:Q}@,${${v}:S|\\|\\\\|gW:S|,|\\,|gW:S|&|\\\&|gW:S|${.newline}|\\${.newline}|gW:Q},g
+SUBST_FILTER_CMD.${class}+=	-e s,@${v:C|[.[\\*^${$}]|\\\\&|gW:Q}@,${${v}:S|\\|\\\\|gW:S|,|\\,|gW:S|&|\\\&|gW:S|${.newline}|\\${.newline}|gW:Q},g
 .  endfor
 .  if ${SUBST_SHOW_DIFF.${class}:U${SUBST_SHOW_DIFF}:tl} == yes
 _SUBST_KEEP.${class}?=		LC_ALL=C ${DIFF} -u "$$file" "$$tmpfile" || true

Index: pkgsrc/regress/infra-unittests/subst.sh
diff -u pkgsrc/regress/infra-unittests/subst.sh:1.15 pkgsrc/regress/infra-unittests/subst.sh:1.16
--- pkgsrc/regress/infra-unittests/subst.sh:1.15	Sat Mar 28 12:08:28 2020
+++ pkgsrc/regress/infra-unittests/subst.sh	Sat Apr 18 12:21:10 2020
@@ -801,7 +801,7 @@ if test_case_begin "SUBST_VARS for varia
 		'SUBST_CLASSES+=	vars' \
 		'SUBST_STAGE.vars=	pre-configure' \
 		'SUBST_FILES.vars=	vars.txt' \
-		'SUBST_VARS.vars=	VAR...... VAR.abcde VAR.() VAR.<>' \
+		'SUBST_VARS.vars=	VAR...... VAR.abcde VAR.() VAR.<> VAR.[]' \
 		'' \
 		'VAR......=	dots' \
 		'VAR.abcde=	letters' \
@@ -817,20 +817,19 @@ if test_case_begin "SUBST_VARS for varia
 		"@VAR.()@" \
 		"@VAR.<>@" \
 		"@VAR.[]@"
+cp "$tmpdir/testcase.mk" /tmp
 
 	run_bmake "testcase.mk" "pre-configure" \
 		1> "$tmpdir/stdout" \
 		2> "$tmpdir/stderr" \
 	&& exitcode=0 || exitcode=$?
 
-	# TODO: Why are the angle brackets replaced, but not the parentheses
-	# and square brackets?
 	assert_that "vars.txt" --file-is-lines \
 		"dots" \
 		"letters" \
-		"@VAR.()@" \
+		"parentheses" \
 		"angle brackets" \
-		"@VAR.[]@"
+		"square brackets"
 	assert_that "stdout" --file-is-lines \
 		"=> Substituting \"vars\" in vars.txt"
 	assert_that "stderr" --file-is-empty


--_----------=_1587212471197570--