From 96b5b4420bbf27071447e00aa9887aa36941ce09 Mon Sep 17 00:00:00 2001
From: Myk <myk002@yahoo.com>
Date: Sun, 10 Apr 2022 21:18:01 -0700
Subject: [PATCH] Add string:escape_pattern() utility function (#2082)

* add string:escape_pattern() to dfhack.lua

stolen from devel/query.lua. will migrate scripts to use the common implementation later
---
 docs/Lua API.rst        | 45 +++++++++++++++++++++++++++++++++++++++++
 docs/changelog.txt      |  2 ++
 library/lua/dfhack.lua  |  6 ++++++
 test/library/string.lua | 15 ++++++++++++++
 4 files changed, 68 insertions(+)

diff --git a/docs/Lua API.rst b/docs/Lua API.rst
index 6b58574de..5c73dbbd0 100644
--- a/docs/Lua API.rst	
+++ b/docs/Lua API.rst	
@@ -2643,6 +2643,51 @@ environment by the mandatory init file dfhack.lua:
   Walks a sequence of dereferences, which may be represented by numbers or strings.
   Returns *nil* if any of obj or indices is *nil*, or a numeric index is out of array bounds.
 
+.. _lua-string:
+
+String class extentions
+-----------------------
+
+DFHack extends Lua's basic string class to include a number of convenience
+functions. These are invoked just like standard string functions, e.g.::
+
+    if imastring:startswith('imaprefix') then
+
+* ``string:startswith(prefix)``
+
+  Returns ``true`` if the first ``#prefix`` characters of the string are equal
+  to ``prefix``. Note that ``prefix`` is not interpreted as a pattern.
+
+* ``string:endswith(suffix)``
+
+  Returns ``true`` if the last ``#suffix`` characters of the string are equal
+  to ``suffix``. Note that ``suffix`` is not interpreted as a pattern.
+
+* ``string:split([delimiter[, plain]])``
+
+  Split a string by the given delimiter. If no delimiter is specified, space
+  (``' '``) is used. The delimter is treated as a pattern unless a ``plain`` is
+  specified and set to ``true``. To treat multiple successive delimiter
+  characters as a single delimiter, e.g. to avoid getting empty string elements,
+  pass a pattern like ``' +'``. Be aware that passing patterns that match empty
+  strings (like ``' *'``) will result in improper string splits.
+
+* ``string:trim()``
+
+  Removes spaces (i.e. everything that matches ``'%s'``) from the start and end
+  of a string. Spaces between non-space characters are left untouched.
+
+* ``string:wrap([width])``
+
+  Inserts newlines into a string so no individual line exceeds the given width.
+  Lines are split at space-separated word boundaries. Any existing newlines are
+  kept in place. If a single word is longer than width, it is split over
+  multiple lines. If width is not specified, 72 is used.
+
+* ``string:escape_pattern()``
+
+  Escapes regex special chars in a string. E.g. ``'a+b'`` -> ``'a%+b'``.
+
 utils
 =====
 
diff --git a/docs/changelog.txt b/docs/changelog.txt
index d7fcb9617..c4e1fed4e 100644
--- a/docs/changelog.txt
+++ b/docs/changelog.txt
@@ -64,6 +64,7 @@ changelog.txt uses a syntax similar to RST, with a few special sequences:
 - Updated information regarding obtaining a compatible Windows build environment
 - `confirm`: Correct the command name in the plugin help text
 - ``Quickfort Blueprint Editing Guide``: added screenshots to the Dreamfort case study and overall clarified text
+- Document DFHack `lua-string` (``startswith``, ``endswith``, ``split``, ``trim``, ``wrap``, and ``escape_pattern``).
 
 ## API
 - Added functions reverse-engineered from ambushing unit code: ``Units::isHidden``, ``Units::isFortControlled``, ``Units::getOuterContainerRef``, ``Items::getOuterContainerRef``
@@ -71,6 +72,7 @@ changelog.txt uses a syntax similar to RST, with a few special sequences:
 ## Lua
 - Lua wrappers for functions reverse-engineered from ambushing unit code: ``isHidden(unit)``, ``isFortControlled(unit)``, ``getOuterContainerRef(unit)``, ``getOuterContainerRef(item)``
 - ``dwarfmode.enterSidebarMode()``: passing ``df.ui_sidebar_mode.DesignateMine`` now always results in you entering ``DesignateMine`` mode and not ``DesignateChopTrees``, even when you looking at the surface where the default designation mode is ``DesignateChopTrees``
+- New string class function: ``string:escape_pattern()`` escapes regex special characters within a string
 
 # 0.47.05-r4
 
diff --git a/library/lua/dfhack.lua b/library/lua/dfhack.lua
index 11eebcbf7..5fe377b50 100644
--- a/library/lua/dfhack.lua
+++ b/library/lua/dfhack.lua
@@ -462,6 +462,12 @@ function string:wrap(width)
     return table.concat(wrapped_text, '\n')
 end
 
+-- Escapes regex special chars in a string. E.g. "a+b" -> "a%+b"
+local regex_chars_pattern = '(['..('%^$()[].*+-?'):gsub('(.)', '%%%1')..'])'
+function string:escape_pattern()
+    return self:gsub(regex_chars_pattern, '%%%1')
+end
+
 -- String conversions
 
 function dfhack.persistent:__tostring()
diff --git a/test/library/string.lua b/test/library/string.lua
index 350f07d5e..d45be4b3f 100644
--- a/test/library/string.lua
+++ b/test/library/string.lua
@@ -68,3 +68,18 @@ function test.wrap()
     expect.eq('hel\nloo\nwor\nldo', ('helloo  worldo'):wrap(3))
     expect.eq('', (''):wrap())
 end
+
+function test.escape_pattern()
+    -- no change expected
+    expect.eq('', (''):escape_pattern())
+    expect.eq(' ', (' '):escape_pattern())
+    expect.eq('abc', ('abc'):escape_pattern())
+    expect.eq('a,b', ('a,b'):escape_pattern())
+    expect.eq('"a,b"', ('"a,b"'):escape_pattern())
+
+    -- excape regex chars
+    expect.eq('iz for me%?', ('iz for me?'):escape_pattern())
+    expect.eq('%.%*', ('.*'):escape_pattern())
+    expect.eq('%( %) %. %% %+ %- %* %? %[ %] %^ %$',
+              ('( ) . % + - * ? [ ] ^ $'):escape_pattern())
+end