From c5389dfa4237a24085c5db3d3f5866862e2293e5 Mon Sep 17 00:00:00 2001
From: Jon Michael Aanes <jonjmaa@gmail.com>
Date: Mon, 23 Oct 2017 14:40:30 +0200
Subject: [PATCH] Implemented shortening of decimal escaped charaters, when at
 end of cut-strings.

---
 common.lua            |  1 -
 pstring.lua           | 27 ++++++++++++++++++++-------
 test/test_pstring.lua | 29 +++++++++++++++++++++++++++--
 3 files changed, 47 insertions(+), 10 deletions(-)

diff --git a/common.lua b/common.lua
index 3f96154..830b0b2 100644
--- a/common.lua
+++ b/common.lua
@@ -27,7 +27,6 @@ local UNICODE_ZERO_WIDTH_CHARACTERS = {}
 for i = 128, 191 do  UNICODE_ZERO_WIDTH_CHARACTERS['\204'..string.char(i)] = true  end
 for i = 128, 175 do  UNICODE_ZERO_WIDTH_CHARACTERS['\205'..string.char(i)] = true  end
 
-
 local function iterate_utf8_chars (str)
 	-- TODO: Detect invalid codepoints.
 	return str:gmatch(UNICODE_CHAR_PATTERN)
diff --git a/pstring.lua b/pstring.lua
index db6d47c..18f99ac 100644
--- a/pstring.lua
+++ b/pstring.lua
@@ -109,14 +109,27 @@ local function safe_cut (str, si, ei)
 	-- Calculate
 	local cut_str = str:sub(si, ei)
 
-	-- Search for the number of backslashes just before the send of the string.
-	-- If that number is even, it's a sequence of backslashes, if not it's a
-	-- broken escape string.
+	-- Search for the number of backslashes and digits at the end of the string.
+	-- If the number of backslashes is even, it's a sequence of backslashes, if
+	-- not it's a broken escape string.
 	local start_of_backslashes, start_of_digits = cut_str:match '()\\*()%d?%d?$'
 	local nr_backslashes_before_end  =  start_of_digits - start_of_backslashes
-	if nr_backslashes_before_end % 2 == 1 then  cut_str = cut_str:sub(1, start_of_backslashes - 1)  end
+	if nr_backslashes_before_end % 2 == 1 then
+		-- Lets see if we can't shorten the escape code, to fit within the
+		-- cut limit.
+		local space_left = #cut_str - (start_of_digits - 2)
+		cut_str = cut_str:sub(1, start_of_digits - 2)
+		ei = ei - space_left
 
-	return cut_str
+		local digits, after_digits = str:match('^\\(%d?%d?%d?)()', si - 1 + start_of_digits - 1)
+
+		if space_left >= 1 + 3 - #digits:match '0*' then
+			ei = after_digits - 1
+			cut_str = cut_str .. ('\\%0'..(space_left-1)..'i'):format(digits)
+		end
+	end
+
+	return cut_str, ei
 end
 
 
@@ -149,8 +162,8 @@ local function format_concatted_string (str, _, l)
 	-- Cut strings
 	local sub_strings, str_i = {}, 1
 	repeat
-		local sub_str  =  safe_cut(str, str_i, str_i + width_without_overhead - 1)
-		str_i = str_i + #sub_str
+		local sub_str, ei  =  safe_cut(str, str_i, str_i + width_without_overhead - 1)
+		str_i = ei + 1
 		sub_strings[#sub_strings+1] = sub_str
 	until str_i >= #str
 
diff --git a/test/test_pstring.lua b/test/test_pstring.lua
index e7c6bb2..addcaff 100644
--- a/test/test_pstring.lua
+++ b/test/test_pstring.lua
@@ -130,6 +130,13 @@ format_test {
     expect = '{\n    \'Lorem ipsum dolor sit amet, consec\\t\'...\n}',
 }
 
+format_test {
+    name   = 'Cut strings are not cut in the middle of backslash sequence',
+    not_idempotent = true,
+    input  = {'Lorem ipsum dolor sit amet, consec\\\\\\\\\\\\\\tetur adipiscing elit. Nunc vestibulum tempus ligula. Sed ac lobortis mi.'},
+    expect = '{\n    \'Lorem ipsum dolor sit amet, consec\\\\\'...\n}',
+}
+
 format_test {
     name   = 'Cut strings are not cut in the middle of escaping \\',
     not_idempotent = true,
@@ -145,13 +152,19 @@ format_test {
 }
 
 format_test {
-    -- NOTE: Not priority functionallity.
-    name   = 'Cut strings can shorten decimal escape codes, if nessesary and possible',
+    name   = 'Cut strings can shorten decimal escape codes, if necessary and possible',
     not_idempotent = true,
     input  = {'Lorem ipsum dolor sit amet, consec\014tetur adipiscing elit. Nunc vestibulum tempus ligula. Sed ac lobortis mi.'},
     expect = '{\n    \'Lorem ipsum dolor sit amet, consec\\14\'...\n}',
 }
 
+format_test {
+    name   = 'Cut strings can shorten decimal escape codes, if necessary and possible, but will keep them as long as possible',
+    not_idempotent = true,
+    input  = {'Lorem ipsum dolor sit amet, consec\004tetur adipiscing elit. Nunc vestibulum tempus ligula. Sed ac lobortis mi.'},
+    expect = '{\n    \'Lorem ipsum dolor sit amet, consec\\04\'...\n}',
+}
+
 --------------------------------------------------------------------------------
 -- Concatted Strings
 
@@ -167,6 +180,18 @@ format_test {
     expect = [['Lorem ipsum dolor sit amet, consectetur adipiscing elit.\004\002\000Nunc ve' ..]]..'\n'..[['stibulum tempus ligula. Sed ac lobortis mi.']],
 }
 
+format_test {
+    name   = 'Concatted string with decimal escape at border',
+    input  = 'Lorem ipsum dolor sit amet, consectetur adipiscing elit.Nunc vestibu\004\255\000lum tempus ligula. Sed ac lobortis mi.',
+    expect = [['Lorem ipsum dolor sit amet, consectetur adipiscing elit.Nunc vestibu\004' ..]]..'\n'..[['\255\000lum tempus ligula. Sed ac lobortis mi.']],
+}
+
+format_test {
+    name   = 'Concatted string with decimal escape at border 2',
+    input  = 'Lorem ipsum dolor sit amet, consectetur adipiscing elit.Nunc vestibu\004\002\000lum tempus ligula. Sed ac lobortis mi.',
+    expect = [['Lorem ipsum dolor sit amet, consectetur adipiscing elit.Nunc vestibu\004\02' ..]]..'\n'..[['\000lum tempus ligula. Sed ac lobortis mi.']],
+}
+
 
 --------------------------------------------------------------------------------
 -- Longform Strings