Missing utf_mode handling for PCRE2

[stephane-chazelas]

PCRE regexp support had:

        pcre *comp = pcre_compile(pattern,
                        ((utf_mode) ? PCRE_UTF8 | PCRE_NO_UTF8_CHECK : 0) |

Meaning regexps would work properly at matching UTF-8 encoded text in UTF-8 locales, but with PCRE2 (with less built with --with-regex=pcre2)

echo Stéphane | LESSCHARSET=utf-8 LC_ALL=en_GB.UTF-8 ./less +/St.ph

Reports Pattern not found, and one needs: (*UTF)St.p instead of just St.p as the regexp when matching UTF-8 encoded text.

According to NEWS, PCRE2 support was added at the same time as those PCRE_UTF8 flags were added to PCRE, which seems odd to me. Why not add it for both PCRE and PCRE2 at the time?

It seems to me the fix should just be a matter of:

diff --git a/pattern.c b/pattern.c
index 0c651bd..6534156 100644
--- a/pattern.c
+++ b/pattern.c
@@ -81,6 +81,7 @@ static int compile_pattern2(constant char *pattern, int search_type, PATTERN_TYP
 	PCRE2_SIZE erroffset;
 	PARG parg;
 	pcre2_code *comp = pcre2_compile((PCRE2_SPTR)pattern, strlen(pattern),
+			((utf_mode) ? PCRE2_UTF | PCRE2_NO_UTF_CHECK : 0) |
 			(is_caseless ? PCRE2_CASELESS : 0),
 			&errcode, &erroffset, NULL);
 	if (comp == NULL)

To do the same like with PCRE.

[gwsw]

Fixed in 5f60178.