From d2daa921e9a45801ff418b0719d694e171d517db Mon Sep 17 00:00:00 2001
From: ridiculousfish <rf@fishshell.com>
Date: Sat, 29 Oct 2022 12:51:13 -0700
Subject: [PATCH] Introduce re::make_anchored

This allows adjusting a pattern string so that it matches an entire
string, by wrapping the regex in a group like ^(?:...)$

This is a workaround for the fact that PCRE2_ENDANCHORED is unavailable
on PCRE2 prior to 2017, so we have to adjust the pattern instead.

Also introduce an overload of match() which creates its own
match_data_t.
---
 src/fish_tests.cpp | 15 +++++++++++++++
 src/re.cpp         | 15 +++++++++++++++
 src/re.h           |  8 ++++++++
 3 files changed, 38 insertions(+)

diff --git a/src/fish_tests.cpp b/src/fish_tests.cpp
index 28c38bd37..75dabd33a 100644
--- a/src/fish_tests.cpp
+++ b/src/fish_tests.cpp
@@ -6867,6 +6867,21 @@ static void test_re_basic() {
     }
     do_test(join_strings(matches, L',') == L"AA,CC,11");
     do_test(join_strings(captures, L',') == L"A,C,1");
+
+    // Test make_anchored
+    re = regex_t::try_compile(make_anchored(L"ab(.+?)"));
+    do_test(re.has_value());
+    do_test(!re->match(L""));
+    do_test(!re->match(L"ab"));
+    do_test((re->match(L"abcd") == match_range_t{0, 4}));
+    do_test((re->match(L"abcdefghij") == match_range_t{0, 10}));
+
+    re = regex_t::try_compile(make_anchored(L"(a+)|(b+)"));
+    do_test(re.has_value());
+    do_test(!re->match(L""));
+    do_test(!re->match(L"aabb"));
+    do_test((re->match(L"aaaa") == match_range_t{0, 4}));
+    do_test((re->match(L"bbbb") == match_range_t{0, 4}));
 }
 
 static void test_re_reset() {
diff --git a/src/re.cpp b/src/re.cpp
index 5b1424350..9dba0e18d 100644
--- a/src/re.cpp
+++ b/src/re.cpp
@@ -130,6 +130,11 @@ maybe_t<match_range_t> regex_t::match(match_data_t &md, const wcstring &subject)
     return match_range_t{ovector[0], ovector[1]};
 }
 
+maybe_t<match_range_t> regex_t::match(const wcstring &subject) const {
+    match_data_t md = this->prepare();
+    return this->match(md, subject);
+}
+
 maybe_t<match_range_t> regex_t::group(const match_data_t &md, size_t group_idx) const {
     if (group_idx >= md.max_capture || group_idx >= pcre2_get_ovector_count(get_md(md.data))) {
         return none();
@@ -288,3 +293,13 @@ regex_t::regex_t(adapters::bytecode_ptr_t &&code) : code_(std::move(code)) {
 }
 
 wcstring re_error_t::message() const { return message_for_code(this->code); }
+
+wcstring re::make_anchored(wcstring pattern) {
+    // PATTERN -> ^(:?PATTERN)$.
+    const wchar_t *prefix = L"^(?:";
+    const wchar_t *suffix = L")$";
+    pattern.reserve(pattern.size() + wcslen(prefix) + wcslen(suffix));
+    pattern.insert(0, prefix);
+    pattern.append(suffix);
+    return pattern;
+}
diff --git a/src/re.h b/src/re.h
index 719cc0ff5..134b01c5e 100644
--- a/src/re.h
+++ b/src/re.h
@@ -111,6 +111,9 @@ class regex_t : noncopyable_t {
     /// \return a range on a successful match, none on no match.
     maybe_t<match_range_t> match(match_data_t &md, const wcstring &subject) const;
 
+    /// A convenience function which calls prepare() for you.
+    maybe_t<match_range_t> match(const wcstring &subject) const;
+
     /// \return the matched range for an indexed or named capture group. 0 means the entire match.
     maybe_t<match_range_t> group(const match_data_t &md, size_t group_idx) const;
     maybe_t<match_range_t> group(const match_data_t &md, const wcstring &name) const;
@@ -145,5 +148,10 @@ class regex_t : noncopyable_t {
     adapters::bytecode_ptr_t code_;
 };
 
+/// Adjust a pattern so that it is anchored at both beginning and end.
+/// This is a workaround for the fact that PCRE2_ENDANCHORED is unavailable on pre-2017 PCRE2
+/// (e.g. 10.21, on Xenial).
+wcstring make_anchored(wcstring pattern);
+
 }  // namespace re
 #endif