From 1e8026f143f96467ca15b45101adab2e212d35a2 Mon Sep 17 00:00:00 2001
From: Maxime Coste <mawww@kakoune.org>
Date: Sun, 29 Apr 2018 16:42:46 +1000
Subject: [PATCH] Regex: Use only 128 characters in start desc and encode
 others as 0

Using 257 was using lots of memory for no good reason, as > 127
codepoint are not common enough to be treated specially.
---
 src/regex_impl.cc | 12 +++++++-----
 src/regex_impl.hh | 14 ++++++++------
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/src/regex_impl.cc b/src/regex_impl.cc
index 38a2dfc6..e074f68b 100644
--- a/src/regex_impl.cc
+++ b/src/regex_impl.cc
@@ -15,6 +15,7 @@ namespace Kakoune
 {
 
 constexpr Codepoint CompiledRegex::StartDesc::other;
+constexpr Codepoint CompiledRegex::StartDesc::count;
 
 struct ParsedRegex
 {
@@ -906,15 +907,16 @@ private:
                 {
                     for (auto& range : character_class.ranges)
                     {
-                        auto min = std::min(CompiledRegex::StartDesc::other, range.min);
-                        auto max = std::min(CompiledRegex::StartDesc::other, range.max);
-                        for (Codepoint cp = min; cp <= max; ++cp)
+                        constexpr auto clamp = [](Codepoint cp) { return std::min(CompiledRegex::StartDesc::count, cp); };
+                        for (auto cp = clamp(range.min), end = clamp(range.max + 1); cp < end; ++cp)
                             start_desc.map[cp] = true;
+                        if (range.max >= CompiledRegex::StartDesc::count)
+                            start_desc.map[CompiledRegex::StartDesc::other] = true;
                     }
                 }
                 else
                 {
-                    for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::other; ++cp)
+                    for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::count; ++cp)
                     {
                         if (start_desc.map[cp] or is_character_class(character_class, cp))
                             start_desc.map[cp] = true;
@@ -926,7 +928,7 @@ private:
             case ParsedRegex::CharacterType:
             {
                 const CharacterType ctype = (CharacterType)node.value;
-                for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::other; ++cp)
+                for (Codepoint cp = 0; cp < CompiledRegex::StartDesc::count; ++cp)
                 {
                     if (is_ctype(ctype, cp))
                         start_desc.map[cp] = true;
diff --git a/src/regex_impl.hh b/src/regex_impl.hh
index 90e276d0..379e5539 100644
--- a/src/regex_impl.hh
+++ b/src/regex_impl.hh
@@ -103,9 +103,9 @@ struct CompiledRegex : RefCountable, UseMemoryDomain<MemoryDomain::Regex>
 
     struct StartDesc : UseMemoryDomain<MemoryDomain::Regex>
     {
-        static constexpr size_t count = 256;
-        static constexpr Codepoint other = 256;
-        bool map[count+1];
+        static constexpr Codepoint count = 128;
+        static constexpr Codepoint other = 0;
+        bool map[count];
     };
 
     std::unique_ptr<StartDesc> forward_start_desc;
@@ -213,7 +213,8 @@ public:
                     return false;
             }
             else if (start != config.end and
-                     not start_desc->map[std::min(*start, CompiledRegex::StartDesc::other)])
+                     not start_desc->map[*start < CompiledRegex::StartDesc::count ?
+                                         *start : CompiledRegex::StartDesc::other])
                 return false;
         }
 
@@ -509,8 +510,9 @@ private:
     void to_next_start(EffectiveIt& start, const EffectiveIt& end,
                        const CompiledRegex::StartDesc& start_desc)
     {
-        while (start != end and *start >= 0 and
-               not start_desc.map[std::min(*start, CompiledRegex::StartDesc::other)])
+        Codepoint cp;
+        while (start != end and (cp = *start) >= 0 and
+               not start_desc.map[cp < CompiledRegex::StartDesc::count ? cp : CompiledRegex::StartDesc::other])
             ++start;
     }