Feature #158

- Added feature #158, regular expression support for filters and substitutions.
- Added unit tests to att.t and filter.t.
- Updated docs.

ChangeLog

@@ -5,6 +5,7 @@
  + Added burndown charts - burndown.daily, burndown.weekly, burndown.monthly,
    that use color.burndown.pending, color.burndown.started and
    color.burndown.done colors.
+ + Added feature #158, regular expression support for filters and substitutions.
  + Added feature #247, providing infinite width reports when redirecting output
    to a file, by setting defaultwidth to 0.
  + Added feature #546, which is a 'count' command that counts tasks, and is

NEWS

@@ -3,7 +3,8 @@ New Features in taskwarrior 1.9.4
 
   - New burndown charts.
   - New 'count' helper command.
-  - Inifinite width reports, when redirecting output, by using rc.defaultwidth=0.
+  - Inifinite width reports, when redirecting output.
+  - Regular expression support in filters and substitutions.
 
   Please refer to the ChangeLog file for full details.  There are too many to
   list here.
@@ -21,6 +22,11 @@ New configuration options in taskwarrior 1.9.4
     control the color of the burndown charts.
   - burndown.bias, which is a tweakable control for the completion estimation
     for the burndown charts, and is documented in taskrc(5).
+  - defaultwidth=0 is causes word-wrapping to be turned off, and effectively
+    set reports to potentially have infinite width .
+  - regex=on enables regular expression searches for filters (task list a...e
+    matches 'above') and substitutions (task <id> /a...e/over/ changes 'above'
+    to 'over').  Default is off, as this is an advanced feature.
 
 Newly deprecated features in taskwarrior 1.9.4
 

doc/man/task-faq.5

@@ -315,6 +315,26 @@ non-exact match:
 
 This will remove the second annotation - the first non-exact match.
 
+.TP
+.B Q: Does task support searching using regular expressions?
+Yes, taskwarrior supports IEEE Std 1003.2 (POSIX.2) regular expressions, but not
+by default.  You must enable this feature with the following command:
+
+    $ task config regex on
+
+Once enabled, all searches are considered regular expressions, for example:
+
+    $ task list ^the
+
+will list all tasks whose description or annotations start with "the".
+Substitutions also support regular expressions:
+
+    $ task 1 /^the/The/
+
+Note that regular expressions work in conjunction with the
+.B search.case.sensitive
+configuration setting.
+
 .TP
 .B Q: How can I help?
 There are lots of ways.  Here are some:

doc/man/taskrc.5

@@ -255,6 +255,18 @@ names you have used, or just the ones used in active tasks.  The default value i
 May be yes or no, and determines whether keyword lookup and substitutions on the
 description and annotations are done in a case sensitive way.  Defaults to yes.
 
+.TP
+.B regex=on
+Enables regular expression searches in filters (task list ^Fix), and
+substitutions (task <id> /^the/The/).
+
+Note that this feature works in conjunction with the
+.B search.case.sensitive
+setting.
+
+The default value is off, because this advanced feature could cause confusion
+among users that are not comfortable with regular expressions.
+
 .TP
 .B _forcecolor=no
 Taskwarrior shuts off color automatically when the output is not sent directly

src/Att.cpp

@@ -29,13 +29,14 @@
 #include <sstream>
 #include <stdlib.h>
 #include <string.h>
-#include "text.h"
+#include <text.h>
-#include "Color.h"
+#include <rx.h>
-#include "util.h"
+#include <Color.h>
-#include "Date.h"
+#include <util.h>
-#include "Duration.h"
+#include <Date.h>
-#include "Context.h"
+#include <Duration.h>
-#include "Att.h"
+#include <Context.h>
+#include <Att.h>
 
 extern Context context;
 
@@ -533,11 +534,22 @@ void Att::parse (Nibbler& n)
 ////////////////////////////////////////////////////////////////////////////////
 // "this" is the attribute that has modifiers.  "other" is the attribute from a
 // Record that does not have modifiers, but may have a value.
+//
+// In other words, the filter:
+//   task list description.contains:foo
+//
+// Is represented with:
+//   this  = filter (description.contains:foo)
+//   other = actual task data to be matched
+//
 bool Att::match (const Att& other) const
 {
   // All matches are assumed to pass, any short-circuit on non-match.
   bool case_sensitive = context.config.getBoolean ("search.case.sensitive");
 
+  // Are regular expressions being used in place of string comparison?
+  bool regex = context.config.getBoolean ("regex");
+
   // If there are no mods, just perform a straight compare on value.
   if (mMod == "")
   {
@@ -557,7 +569,13 @@ bool Att::match (const Att& other) const
     }
     else
     {
-      if (!compare (mValue, other.mValue, (bool) case_sensitive))
+      if (regex)
+      {
+        std::string pattern = "^" + mValue + "$";
+        if (!regexMatch (other.mValue, pattern, case_sensitive))
+          return false;
+      }
+      else if (!compare (mValue, other.mValue, (bool) case_sensitive))
         return false;
     }
   }
@@ -565,21 +583,38 @@ bool Att::match (const Att& other) const
   // has = contains as a substring.
   else if (mMod == "has" || mMod == "contains") // TODO i18n
   {
-    if (find (other.mValue, mValue, (bool) case_sensitive) == std::string::npos)
+    if (regex)
+    {
+      if (!regexMatch (other.mValue, mValue, case_sensitive))
+        return false;
+    }
+    else if (find (other.mValue, mValue, (bool) case_sensitive) == std::string::npos)
       return false;
   }
 
   // is = equal.  Nop.
   else if (mMod == "is" || mMod == "equals") // TODO i18n
   {
-    if (!compare (mValue, other.mValue, (bool) case_sensitive))
+    if (regex)
+    {
+      std::string pattern = "^" + mValue + "$";
+      if (!regexMatch (other.mValue, pattern, case_sensitive))
+        return false;
+    }
+    else if (!compare (mValue, other.mValue, (bool) case_sensitive))
       return false;
   }
 
   // isnt = not equal.
   else if (mMod == "isnt" || mMod == "not") // TODO i18n
   {
-    if (compare (mValue, other.mValue, (bool) case_sensitive))
+    if (regex)
+    {
+      std::string pattern = "^" + mValue + "$";
+      if (!regexMatch (other.mValue, pattern, case_sensitive))
+        return false;
+    }
+    else if (compare (mValue, other.mValue, (bool) case_sensitive))
       return false;
   }
 
@@ -599,6 +634,14 @@ bool Att::match (const Att& other) const
 
   // startswith = first characters must match.
   else if (mMod == "startswith" || mMod == "left") // TODO i18n
+  {
+    if (regex)
+    {
+      std::string pattern = "^" + mValue;
+      if (!regexMatch (other.mValue, pattern, case_sensitive))
+        return false;
+    }
+    else
     {
       if (other.mValue.length () < mValue.length ())
         return false;
@@ -606,9 +649,18 @@ bool Att::match (const Att& other) const
       if (!compare (mValue, other.mValue.substr (0, mValue.length ()), (bool) case_sensitive))
         return false;
     }
+  }
 
   // endswith = last characters must match.
   else if (mMod == "endswith" || mMod == "right") // TODO i18n
+  {
+    if (regex)
+    {
+      std::string pattern = mValue + "$";
+      if (!regexMatch (other.mValue, pattern, case_sensitive))
+        return false;
+    }
+    else
     {
       if (other.mValue.length () < mValue.length ())
         return false;
@@ -618,11 +670,17 @@ bool Att::match (const Att& other) const
                       std::string::npos), (bool) case_sensitive))
         return false;
     }
+  }
 
   // hasnt = does not contain as a substring.
   else if (mMod == "hasnt") // TODO i18n
   {
-    if (find (other.mValue, mValue, (bool) case_sensitive) != std::string::npos)
+    if (regex)
+    {
+      if (regexMatch (other.mValue, mValue, case_sensitive))
+        return false;
+    }
+    else if (find (other.mValue, mValue, (bool) case_sensitive) != std::string::npos)
       return false;
   }
 
@@ -704,6 +762,21 @@ bool Att::match (const Att& other) const
 
   // word = contains as a substring, with word boundaries.
   else if (mMod == "word") // TODO i18n
+  {
+    if (regex)
+    {
+      std::vector <int> start;
+      std::vector <int> end;
+      if (!regexMatch (start, end, other.mValue, mValue, case_sensitive))
+        return false;
+
+      if (!isWordStart (other.mValue, start[0]))
+        return false;
+
+      if (!isWordEnd (other.mValue, end[0]))
+        return false;
+    }
+    else
     {
       // Fail if the substring is not found.
       std::string::size_type sub = find (other.mValue, mValue, (bool) case_sensitive);
@@ -717,9 +790,21 @@ bool Att::match (const Att& other) const
       if (!isWordEnd (other.mValue, sub + mValue.length () - 1))
         return false;
     }
+  }
 
   // noword = does not contain as a substring, with word boundaries.
   else if (mMod == "noword") // TODO i18n
+  {
+    if (regex)
+    {
+      std::vector <int> start;
+      std::vector <int> end;
+      if (regexMatch (start, end, other.mValue, mValue, case_sensitive) &&
+          isWordStart (other.mValue, start[0])                     &&
+          isWordEnd (other.mValue, end[0]))
+        return false;
+    }
+    else
     {
       // Fail if the substring is not found.
       std::string::size_type sub = find (other.mValue, mValue);
@@ -730,6 +815,7 @@ bool Att::match (const Att& other) const
         return false;
       }
     }
+  }
 
   return true;
 }

src/Config.cpp

@@ -81,6 +81,7 @@ std::string Config::defaults =
   "recurrence.limit=1                             # Number of future recurring pending tasks\n"
   "undo.style=side                                # Undo style - can be 'side', or 'diff'\n"
   "burndown.bias=0.666                            # Weighted mean bias toward recent data\n"
+  "regex=no                                       # Assume all search/filter strings are regexes\n"
   "\n"
   "# Dates\n"
   "dateformat=m/d/Y                               # Preferred input and display date format\n"

src/Subst.cpp

@@ -30,6 +30,7 @@
 #include <Directory.h>
 #include <Context.h>
 #include <text.h>
+#include <rx.h>
 #include <i18n.h>
 
 extern Context context;
@@ -130,6 +131,59 @@ void Subst::apply (
   bool sensitive = context.config.getBoolean ("search.case.sensitive");
 
   if (mFrom != "")
+  {
+    if (context.config.getBoolean ("regex"))
+    {
+      // Insert capturing parentheses, if necessary.
+      std::string pattern;
+      if (mFrom.find ('(') != std::string::npos)
+        pattern = mFrom;
+      else
+        pattern = "(" + mFrom + ")";
+
+      std::vector <int> start;
+      std::vector <int> end;
+
+      // Perform all subs on description.
+      int counter = 0;
+      if (regexMatch (start, end, description, pattern, sensitive))
+      {
+        for (unsigned int i = 0; i < start.size (); ++i)
+        {
+          description.replace (start[i], end[i] - start[i], mTo);
+          if (!mGlobal)
+            break;
+
+          if (++counter > 1000)
+            throw ("Terminated substitution because more than a thousand changes were made - infinite loop protection.");
+        }
+      }
+
+      // Perform all subs on annotations.
+      counter = 0;
+      std::vector <Att>::iterator i;
+      for (i = annotations.begin (); i != annotations.end (); ++i)
+      {
+        std::string annotation = i->value ();
+        start.clear ();
+        end.clear ();
+
+        if (regexMatch (start, end, annotation, pattern, sensitive))
+        {
+          for (unsigned int match = 0; match < start.size (); ++match)
+          {
+            annotation.replace (start[match], end[match] - start[match], mTo);
+            i->value (annotation);
+            if (!mGlobal)
+              break;
+
+            if (++counter > 1000)
+              throw ("Terminated substitution because more than a thousand changes were made - infinite loop protection.");
+          }
+        }
+      }
+    }
+    else
     {
       if (mGlobal)
       {
@@ -188,6 +242,7 @@ void Subst::apply (
         }
       }
     }
+  }
 }
 
 ////////////////////////////////////////////////////////////////////////////////

src/command.cpp

@@ -897,7 +897,7 @@ int handleShow (std::string& outs)
       "color.burndown.pending color.burndown.started color.overdue color.pri.H "
       "color.pri.L color.pri.M color.pri.none color.recurring color.tagged "
       "color.footnote color.header color.debug color.alternate color.calendar.today "
-      "color.calendar.due color.calendar.due.today color.calendar.overdue "
+      "color.calendar.due color.calendar.due.today color.calendar.overdue regex "
       "color.calendar.weekend color.calendar.holiday color.calendar.weeknumber "
       "color.summary.background color.summary.bar color.history.add "
       "color.history.done color.history.delete color.undo.before "

src/tests/att.t.cpp

@@ -34,7 +34,7 @@ Context context;
 ////////////////////////////////////////////////////////////////////////////////
 int main (int argc, char** argv)
 {
-  UnitTest t (99);
+  UnitTest t (117);
 
   Att a;
   t.notok (a.valid ("name"),            "Att::valid name         -> fail");
@@ -86,7 +86,7 @@ int main (int argc, char** argv)
   t.is (a6.value_int (), 7, "Att::value_int set/get");
   t.is (a6.value (), "7", "Att::value 7");
 
-  // Att::mod
+  // Att::mod - straight comparisons.
   bool good = true;
   try {a6.mod ("is");} catch (...) {good = false;}
   t.ok (good, "Att::mod (is)");
@@ -159,6 +159,81 @@ int main (int argc, char** argv)
   try {a6.mod ("unrecognized");} catch (...) {good = false;}
   t.notok (good, "Att::mod (unrecognized)");
 
+  // Att::mod - regex comparisons.
+  context.config.set ("regex", "on");
+
+  good = true;
+  try {a6.mod ("is");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (is)");
+
+  good = true;
+  try {a6.mod ("before");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (before)");
+
+  good = true;
+  try {a6.mod ("after");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (after)");
+
+  good = true;
+  try {a6.mod ("none");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (none)");
+
+  good = true;
+  try {a6.mod ("any");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (any)");
+
+  good = true;
+  try {a6.mod ("over");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (over)");
+
+  good = true;
+  try {a6.mod ("under");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (under)");
+
+  good = true;
+  try {a6.mod ("above");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (above)");
+
+  good = true;
+  try {a6.mod ("below");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (below)");
+
+  good = true;
+  try {a6.mod ("isnt");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (isnt)");
+
+  good = true;
+  try {a6.mod ("has");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (has)");
+
+  good = true;
+  try {a6.mod ("contains");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (contains)");
+
+  good = true;
+  try {a6.mod ("hasnt");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (hasnt)");
+
+  good = true;
+  try {a6.mod ("startswith");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (startswith)");
+
+  good = true;
+  try {a6.mod ("endswith");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (endswith)");
+
+  good = true;
+  try {a6.mod ("word");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (word)");
+
+  good = true;
+  try {a6.mod ("noword");} catch (...) {good = false;}
+  t.ok (good, "Att::mod (noword)");
+
+  good = true;
+  try {a6.mod ("unrecognized");} catch (...) {good = false;}
+  t.notok (good, "Att::mod (unrecognized)");
+
   // Att::parse
   Nibbler n ("");
   Att a7;

src/tests/filter.t

@@ -28,7 +28,7 @@
 
 use strict;
 use warnings;
-use Test::More tests => 131;
+use Test::More tests => 180;
 
 # Create the rc file.
 if (open my $fh, '>', 'filter.rc')
@@ -209,6 +209,70 @@ unlike ($output, qr/five/,  'r5');
 unlike ($output, qr/six/,   'r6');
 unlike ($output, qr/seven/, 'r7');
 
+# Regex filters.
+$output = qx{../task rc:filter.rc list rc.regex:on project:[A-Z]};
+like   ($output, qr/one/,   's1');
+like   ($output, qr/two/,   's2');
+like   ($output, qr/three/, 's3');
+unlike ($output, qr/four/,  's4');
+unlike ($output, qr/five/,  's5');
+unlike ($output, qr/six/,   's6');
+unlike ($output, qr/seven/, 's7');
+
+$output = qx{../task rc:filter.rc list rc.regex:on project:.};
+like   ($output, qr/one/,   't1');
+like   ($output, qr/two/,   't2');
+like   ($output, qr/three/, 't3');
+unlike ($output, qr/four/,  't4');
+unlike ($output, qr/five/,  't5');
+unlike ($output, qr/six/,   't6');
+unlike ($output, qr/seven/, 't7');
+
+$output = qx{../task rc:filter.rc rc.regex:on list fo\{2\}};
+like   ($output, qr/one/,   'u1');
+unlike ($output, qr/two/,   'u2');
+unlike ($output, qr/three/, 'u3');
+unlike ($output, qr/four/,  'u4');
+unlike ($output, qr/five/,  'u5');
+like   ($output, qr/six/,   'u6');
+like   ($output, qr/seven/, 'u7');
+
+$output = qx{../task rc:filter.rc rc.regex:on list f.. b..};
+unlike ($output, qr/one/,   'v1');
+unlike ($output, qr/two/,   'v2');
+unlike ($output, qr/three/, 'v3');
+unlike ($output, qr/four/,  'v4');
+unlike ($output, qr/five/,  'v5');
+unlike ($output, qr/six/,   'v6');
+like   ($output, qr/seven/, 'v7');
+
+$output = qx{../task rc:filter.rc rc.regex:on list ^s};
+unlike ($output, qr/one/,   'w1');
+unlike ($output, qr/two/,   'w2');
+unlike ($output, qr/three/, 'w3');
+unlike ($output, qr/four/,  'w4');
+unlike ($output, qr/five/,  'w5');
+like   ($output, qr/six/,   'w6');
+like   ($output, qr/seven/, 'w7');
+
+$output = qx{../task rc:filter.rc rc.regex:on list ^.i};
+unlike ($output, qr/one/,   'x1');
+unlike ($output, qr/two/,   'x2');
+unlike ($output, qr/three/, 'x3');
+unlike ($output, qr/four/,  'x4');
+like   ($output, qr/five/,  'x5');
+like   ($output, qr/six/,   'x6');
+unlike ($output, qr/seven/, 'x7');
+
+$output = qx{../task rc:filter.rc rc.regex:on list "two|five"};
+unlike ($output, qr/one/,   'y1');
+like   ($output, qr/two/,   'y2');
+unlike ($output, qr/three/, 'y3');
+unlike ($output, qr/four/,  'y4');
+like   ($output, qr/five/,  'y5');
+unlike ($output, qr/six/,   'y6');
+unlike ($output, qr/seven/, 'y7');
+
 # Cleanup.
 unlink 'pending.data';
 ok (!-r 'pending.data', 'Removed pending.data');