Merge: Added contributing guidelines and link from readme
[nit.git] / lib / core / text / flat.nit
index 0fc5663..6273609 100644 (file)
@@ -225,6 +225,22 @@ redef class FlatText
                                req_esc += 1
                        else if c == 0x5Cu8 then
                                req_esc += 1
+                       else if c == 0x3Fu8 then
+                               var j = pos + 1
+                               if j < length then
+                                       var next = its[j]
+                                       # We ignore `??'` because it will be escaped as `??\'`.
+                                       if
+                                               next == 0x21u8 or
+                                               next == 0x28u8 or
+                                               next == 0x29u8 or
+                                               next == 0x2Du8 or
+                                               next == 0x2Fu8 or
+                                               next == 0x3Cu8 or
+                                               next == 0x3Du8 or
+                                               next == 0x3Eu8
+                                       then req_esc += 1
+                               end
                        else if c < 32u8 then
                                req_esc += 3
                        end
@@ -280,6 +296,27 @@ redef class FlatText
                                nns[opos] = 0x5Cu8
                                nns[opos + 1] = 0x5Cu8
                                opos += 2
+                       else if c == 0x3Fu8 then
+                               var j = pos + 1
+                               if j < length then
+                                       var next = its[j]
+                                       # We ignore `??'` because it will be escaped as `??\'`.
+                                       if
+                                               next == 0x21u8 or
+                                               next == 0x28u8 or
+                                               next == 0x29u8 or
+                                               next == 0x2Du8 or
+                                               next == 0x2Fu8 or
+                                               next == 0x3Cu8 or
+                                               next == 0x3Du8 or
+                                               next == 0x3Eu8
+                                       then
+                                               nns[opos] = 0x5Cu8
+                                               opos += 1
+                                       end
+                               end
+                               nns[opos] = 0x3Fu8
+                               opos += 1
                        else if c < 32u8 then
                                nns[opos] = 0x5Cu8
                                nns[opos + 1] = 0x30u8
@@ -369,6 +406,10 @@ redef class FlatText
                end
                return res
        end
+
+       redef fun copy_to_native(dst, n, src_off, dst_off) do
+               _items.copy_to(dst, n, first_byte + src_off, dst_off)
+       end
 end
 
 # Immutable strings of characters.
@@ -411,7 +452,7 @@ abstract class FlatString
 
                if from < 0 then
                        count += from
-                       if count < 0 then return ""
+                       if count <= 0 then return ""
                        from = 0
                end
 
@@ -666,15 +707,15 @@ private class ASCIIFlatString
        end
 
        redef fun substring(from, count) do
+               var ln = _length
+               if count <= 0 then return ""
+               if (count + from) > ln then count = ln - from
                if count <= 0 then return ""
-
                if from < 0 then
                        count += from
-                       if count < 0 then return ""
+                       if count <= 0 then return ""
                        from = 0
                end
-               var ln = _length
-               if (count + from) > ln then count = ln - from
                return new ASCIIFlatString.full_data(_items, count, from + _first_byte, count)
        end
 
@@ -922,9 +963,12 @@ class FlatBuffer
 
        redef fun clear do
                is_dirty = true
-               if written then reset
                _bytelen = 0
                _length = 0
+               if written then
+                       _capacity = 16
+                       reset
+               end
        end
 
        redef fun empty do return new Buffer
@@ -933,12 +977,13 @@ class FlatBuffer
        do
                var c = capacity
                if cap <= c then return
-               while c <= cap do c = c * 2 + 2
+               if c <= 16 then c = 16
+               while c <= cap do c = c * 2
                # The COW flag can be set at false here, since
                # it does a copy of the current `Buffer`
                written = false
                var bln = _bytelen
-               var a = new NativeString(c+1)
+               var a = new NativeString(c)
                if bln > 0 then
                        var it = _items
                        if bln > 0 then it.copy_to(a, bln, 0, 0)
@@ -990,22 +1035,17 @@ class FlatBuffer
        init from(s: Text)
        do
                _items = new NativeString(s.bytelen)
-               if s isa FlatText then
-                       _items = s._items
-               else
-                       for i in substrings do i.as(FlatString)._items.copy_to(_items, i._bytelen, 0, 0)
-               end
+               for i in s.substrings do i._items.copy_to(_items, i._bytelen, first_byte, 0)
                _bytelen = s.bytelen
                _length = s.length
                _capacity = _bytelen
-               written = true
        end
 
        # Create a new empty string with a given capacity.
        init with_capacity(cap: Int)
        do
                assert cap >= 0
-               _items = new NativeString(cap + 1)
+               _items = new NativeString(cap)
                capacity = cap
                _bytelen = 0
        end
@@ -1053,6 +1093,21 @@ class FlatBuffer
                return new FlatBuffer.with_infos(r_items, byte_length, byte_length, count)
        end
 
+       redef fun append_substring_impl(s, from, length) do
+               if length <= 0 then return
+               if not s isa FlatText then
+                       super
+                       return
+               end
+               var bytest = s.char_to_byte_index(from)
+               var bytend = s.char_to_byte_index(from + length - 1)
+               var btln = bytend - bytest + 1
+               enlarge(btln + _bytelen)
+               s._items.copy_to(_items, btln, bytest, _bytelen)
+               _bytelen += btln
+               _length += length
+       end
+
        redef fun reverse
        do
                written = false
@@ -1355,37 +1410,26 @@ redef class NativeString
        #
        # Very unsafe, make sure to have room for this char prior to calling this function.
        private fun set_char_at(pos: Int, c: Char) do
-               if c.code_point < 128 then
-                       self[pos] = c.code_point.to_b
+               var cp = c.code_point
+               if cp < 128 then
+                       self[pos] = cp.to_b
                        return
                end
                var ln = c.u8char_len
-               native_set_char(pos, c, ln)
-       end
-
-       private fun native_set_char(pos: Int, c: Char, ln: Int) `{
-               char* dst = self + pos;
-               switch(ln){
-                       case 1:
-                               dst[0] = c;
-                               break;
-                       case 2:
-                               dst[0] = 0xC0 | ((c & 0x7C0) >> 6);
-                               dst[1] = 0x80 | (c & 0x3F);
-                               break;
-                       case 3:
-                               dst[0] = 0xE0 | ((c & 0xF000) >> 12);
-                               dst[1] = 0x80 | ((c & 0xFC0) >> 6);
-                               dst[2] = 0x80 | (c & 0x3F);
-                               break;
-                       case 4:
-                               dst[0] = 0xF0 | ((c & 0x1C0000) >> 18);
-                               dst[1] = 0x80 | ((c & 0x3F000) >> 12);
-                               dst[2] = 0x80 | ((c & 0xFC0) >> 6);
-                               dst[3] = 0x80 | (c & 0x3F);
-                               break;
-               }
-       `}
+               if ln == 2 then
+                       self[pos] = (0xC0 | ((cp & 0x7C0) >> 6)).to_b
+                       self[pos + 1] = (0x80 | (cp & 0x3F)).to_b
+               else if ln == 3 then
+                       self[pos] = (0xE0 | ((cp & 0xF000) >> 12)).to_b
+                       self[pos + 1] = (0x80 | ((cp & 0xFC0) >> 6)).to_b
+                       self[pos + 2] = (0x80 | (cp & 0x3F)).to_b
+               else if ln == 4 then
+                       self[pos] = (0xF0 | ((cp & 0x1C0000) >> 18)).to_b
+                       self[pos + 1] = (0x80 | ((cp & 0x3F000) >> 12)).to_b
+                       self[pos + 2] = (0x80 | ((cp & 0xFC0) >> 6)).to_b
+                       self[pos + 3] = (0x80 | (cp & 0x3F)).to_b
+               end
+       end
 end
 
 redef class Int