From acbb68d13fa5597020b969297764f077896530ed Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?S=C3=B6ren=20Tempel?= <soeren+git@soeren-tempel.net>
Date: Sat, 28 Sep 2024 12:17:40 +0200
Subject: [PATCH] awk: Mark test for handling of "start of word" pattern as
 known-broken
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

If REG_STARTEND is not supported by the regex implementation, then
this code does not work correctly. This is even documented in a
source code comment within awk.c. Hence, without REG_STARTEND (e.g. on
musl libc), this test fails.

See:

* https://git.busybox.net/busybox/commit/?id=2ca39ffd447ca874fcea933194829717d5573247
* https://git.busybox.net/busybox/tree/editors/awk.c?id=2ca39ffd447ca874fcea933194829717d5573247#n2537

Signed-off-by: Sören Tempel <soeren+git@soeren-tempel.net>
---
 testsuite/awk.tests | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/testsuite/awk.tests b/testsuite/awk.tests
index be25f6696..31c430b98 100755
--- a/testsuite/awk.tests
+++ b/testsuite/awk.tests
@@ -608,6 +608,7 @@ testing 'awk references to empty fields' \
 b
 ' '' 'a\nb\n'
 
+test x"$SKIP_KNOWN_BUGS" = x"" && {
 # The "b" in "abc" should not match <b* pattern.
 # Currently we use REG_STARTEND ("This flag is a BSD extension, not present in POSIX")
 # to implement the code to handle this correctly, but if your libc has no REG_STARTEND,
@@ -616,5 +617,6 @@ testing 'awk gsub erroneous word start match' \
 	"awk 'BEGIN { a=\"abc\"; gsub(/\<b*/,\"\",a); print a }'" \
 	'abc\n' \
 	'' ''
+}
 
 exit $FAILCOUNT