From: Francois Fleuret <francois@fleuret.org>
Date: Mon, 16 Mar 2009 20:33:06 +0000 (+0100)
Subject: Merge branch 'master' of ssh://fleuret@login.idiap.ch/homes/fleuret/public/git/selector
X-Git-Url: https://www.fleuret.org/cgi-bin/gitweb/gitweb.cgi?a=commitdiff_plain;h=db57909a08ce6bf9ce426f37a383f7ab0678f171;hp=5ec53edf02493ba52127bedc508ab53bb5c2b9f4;p=selector.git

Merge branch 'master' of ssh://fleuret@login.idiap.ch/homes/fleuret/public/git/selector

Conflicts:
	selector.cc
---

diff --git a/selector.1 b/selector.1
index c1762c1..81c4aa4 100644
--- a/selector.1
+++ b/selector.1
@@ -7,8 +7,12 @@ selector - A simple command line for dynamic pattern selection
 .SH "DESCRIPTION"
 .PP
 \fBselector\fP is a command line dynamic string selection. As you type
-a list of strings separated by ';', the display is updated in real
-time to show only the lines containing all the said strings.
+a list of strings separated by ';' or a regexp, the display is updated
+in real time to show only the lines containing all the said strings,
+or matching the regexp.
+
+The ^R key switches between the standard multi-substring mode and the
+regexp mode.
 
 The main usage of selector is as an efficient search in the shell
 command history. With the correct option, it will inject the selected
@@ -40,8 +44,10 @@ inverse the order of the lines
 remove the numeric prefix from bash history
 .IP "\fB-z\fP" 10
 remove the time prefix from zsh history
-.IP "\fB-r\fP" 10
+.IP "\fB-d\fP" 10
 remove duplicated lines
+.IP "\fB-e\fP" 10
+start with the regexp mode activated
 .IP "\fB-c <fg modeline> <bg modeline> <fg highlight> <bg highlight>\fP" 10
 select the modline and highlight color numbers
 .IP "\fB-o <output filename>\fP" 10
@@ -57,7 +63,7 @@ specify the file to search into
 
 To use selector to search into your bash history, you can use
 
-.B selector -r -i -b -v -f <(history)
+.B selector -d -i -b -v -f <(history)
 
 .SH "BUGS"
 
diff --git a/selector.cc b/selector.cc
index af8ef84..e21025f 100644
--- a/selector.cc
+++ b/selector.cc
@@ -39,6 +39,7 @@
 #include <fcntl.h>
 #include <sys/ioctl.h>
 #include <termios.h>
+#include <regex.h>
 
 using namespace std;
 
@@ -55,6 +56,7 @@ int with_colors = 1;
 int zsh_history = 0, bash_history = 0;
 int inverse_order = 0;
 int remove_duplicates = 0;
+int use_regexp = 0;
 
 //////////////////////////////////////////////////////////////////////
 
@@ -65,22 +67,13 @@ void inject_into_tty_buffer(char *line) {
   tcgetattr(STDIN_FILENO,&oldtio);
   memset(&newtio, 0, sizeof(newtio));
   // Set input mode (non-canonical, *no echo*,...)
-  tcsetattr(STDIN_FILENO,TCSANOW, &newtio);
+  tcsetattr(STDIN_FILENO, TCSANOW, &newtio);
   // Put the selected line in the tty input buffer
   for(char *k = line; *k; k++) {
     ioctl(STDIN_FILENO, TIOCSTI, k);
   }
   // Restore the old settings
-  tcsetattr(STDIN_FILENO,TCSANOW, &oldtio);
-}
-
-//////////////////////////////////////////////////////////////////////
-
-int match(char *string, int nb_patterns, char **patterns) {
-  for(int n = 0; n < nb_patterns; n++) {
-    if(strstr(string, patterns[n]) == 0) return 0;
-  }
-  return 1;
+  tcsetattr(STDIN_FILENO, TCSANOW, &oldtio);
 }
 
 //////////////////////////////////////////////////////////////////////
@@ -95,17 +88,111 @@ void check_opt(int argc, char **argv, int n_opt, int n, const char *help) {
   }
 }
 
+//////////////////////////////////////////////////////////////////////
+// A quick and dirty hash table
+
+int *new_hash_table(int hash_table_size) {
+  int *result;
+  result = new int[hash_table_size];
+  for(int k = 0; k < hash_table_size; k++) {
+    result[k] = -1;
+  }
+  return result;
+}
+
+int test_and_add(char *new_string, int new_index,
+                 char **strings, int *hash_table, int hash_table_size) {
+  unsigned int code = 0;
+
+  for(int k = 0; new_string[k]; k++) {
+    code += int(new_string[k]) << (8 * k%4);
+  }
+
+  code = code % hash_table_size;
+
+  while(hash_table[code] >= 0) {
+    if(strcmp(new_string, strings[hash_table[code]]) == 0) return 1;
+    code = (code + 1) % hash_table_size;
+  }
+
+  hash_table[code] = new_index;
+
+  return 0;
+}
+
+//////////////////////////////////////////////////////////////////////
+// A matcher matches either with a collection of substrings, or with a
+// regexp
+
+struct matcher_t {
+  regex_t preg;
+  int regexp_error;
+  int nb_patterns;
+  char *splitted_patterns, **patterns;
+};
+
+int match(char *string, matcher_t *matcher) {
+  if(matcher->nb_patterns >= 0) {
+    for(int n = 0; n < matcher->nb_patterns; n++) {
+      if(strstr(string, matcher->patterns[n]) == 0) return 0;
+    }
+    return 1;
+  } else {
+    return regexec(&matcher->preg, string, 0, 0, 0) == 0;
+  }
+}
+
+void free_matcher(matcher_t *matcher) {
+  if(matcher->nb_patterns >= 0) {
+    delete[] matcher->splitted_patterns;
+    delete[] matcher->patterns;
+  } else {
+    if(!matcher->regexp_error) regfree(&matcher->preg);
+  }
+}
+
+void initialize_matcher(int use_regexp, matcher_t *matcher, const char *pattern) {
+  if(use_regexp) {
+    matcher->nb_patterns = -1;
+    matcher->regexp_error = regcomp(&matcher->preg, pattern, REG_ICASE);
+  } else {
+    matcher->regexp_error = 0;
+    matcher->nb_patterns = 1;
+
+    for(const char *s = pattern; *s; s++) {
+      if(*s == pattern_separator) {
+        matcher->nb_patterns++;
+      }
+    }
+
+    matcher->splitted_patterns = new char[strlen(pattern) + 1];
+    matcher->patterns = new char*[matcher->nb_patterns];
+
+    strcpy(matcher->splitted_patterns, pattern);
+
+    int n = 0;
+    char *last_pattern_start = matcher->splitted_patterns;
+    for(char *s = matcher->splitted_patterns; n < matcher->nb_patterns; s++) {
+      if(*s == pattern_separator || *s == '\0') {
+        *s = '\0';
+        matcher->patterns[n++] = last_pattern_start;
+        last_pattern_start = s + 1;
+      }
+    }
+  }
+}
+
 //////////////////////////////////////////////////////////////////////
 
-int previous_visible(int current_line, int nb_lines, char **lines, int nb_patterns, char **patterns) {
+int previous_visible(int current_line, int nb_lines, char **lines, matcher_t *matcher) {
   int line = current_line - 1;
-  while(line >= 0 && !match(lines[line], nb_patterns, patterns)) line--;
+  while(line >= 0 && !match(lines[line], matcher)) line--;
   return line;
 }
 
-int next_visible(int current_line, int nb_lines, char **lines, int nb_patterns, char **patterns) {
+int next_visible(int current_line, int nb_lines, char **lines, matcher_t *matcher) {
   int line = current_line + 1;
-  while(line < nb_lines && !match(lines[line], nb_patterns, patterns)) line++;
+  while(line < nb_lines && !match(lines[line], matcher)) line++;
 
   if(line < nb_lines)
     return line;
@@ -113,36 +200,16 @@ int next_visible(int current_line, int nb_lines, char **lines, int nb_patterns,
     return -1;
 }
 
+//////////////////////////////////////////////////////////////////////
+
 void update_screen(int *current_line, int *temporary_line, int motion,
                    int nb_lines, char **lines,
                    char *pattern_list) {
 
   char buffer[buffer_size];
+  matcher_t matcher;
 
-  // We split the pattern list into individual patterns
-
-  int nb_patterns = 1;
-
-  for(char *s = pattern_list; *s; s++) {
-    if(*s == pattern_separator) {
-      nb_patterns++;
-    }
-  }
-
-  char splitted_patterns[strlen(pattern_list) + 1];
-  char *patterns[nb_patterns];
-
-  strcpy(splitted_patterns, pattern_list);
-
-  int n = 0;
-  char *last_pattern_start = splitted_patterns;
-  for(char *s = splitted_patterns; n < nb_patterns; s++) {
-    if(*s == pattern_separator || *s == '\0') {
-      *s = '\0';
-      patterns[n++] = last_pattern_start;
-      last_pattern_start = s + 1;
-    }
-  }
+  initialize_matcher(use_regexp, &matcher, pattern_list);
 
   // We now take care of printing the lines per se
 
@@ -152,140 +219,144 @@ void update_screen(int *current_line, int *temporary_line, int motion,
   // First, we find a visible line. In priority: The current, or the
   // first visible after it, or the first visible before it.
 
-  int new_line;
-  if(match(lines[*current_line], nb_patterns, patterns)) {
-    new_line = *current_line;
+  int nb_printed_lines = 0;
+
+  clear();
+  use_default_colors();
+  addstr("\n");
+
+  if(matcher.regexp_error) {
+    addstr("[regexp error]");
   } else {
-    new_line = next_visible(*current_line, nb_lines, lines, nb_patterns, patterns);
-    if(new_line < 0) {
-      new_line = previous_visible(*current_line, nb_lines, lines, nb_patterns, patterns);
+
+    int new_line;
+    if(match(lines[*current_line], &matcher)) {
+      new_line = *current_line;
+    } else {
+      new_line = next_visible(*current_line, nb_lines, lines, &matcher);
+      if(new_line < 0) {
+        new_line = previous_visible(*current_line, nb_lines, lines, &matcher);
+      }
     }
-  }
 
-  // If we found a visible line and we should move, let's move
+    // If we found a visible line and we should move, let's move
 
-  if(new_line >= 0 && motion != 0) {
-    int l = new_line;
-    if(motion > 0) {
-      // We want to go down, let's find the first visible line below
-      for(int m = 0; l >= 0 && m < motion; m++) {
-        l = next_visible(l, nb_lines, lines, nb_patterns, patterns);
-        if(l >= 0) {
-          new_line = l;
+    if(new_line >= 0 && motion != 0) {
+      int l = new_line;
+      if(motion > 0) {
+        // We want to go down, let's find the first visible line below
+        for(int m = 0; l >= 0 && m < motion; m++) {
+          l = next_visible(l, nb_lines, lines, &matcher);
+          if(l >= 0) {
+            new_line = l;
+          }
         }
-      }
-    } else {
-      // We want to go up, let's find the first visible line above
-      for(int m = 0; l >= 0 && m < -motion; m++) {
-        l = previous_visible(l, nb_lines, lines, nb_patterns, patterns);
-        if(l >= 0) {
-          new_line = l;
+      } else {
+        // We want to go up, let's find the first visible line above
+        for(int m = 0; l >= 0 && m < -motion; m++) {
+          l = previous_visible(l, nb_lines, lines, &matcher);
+          if(l >= 0) {
+            new_line = l;
+          }
         }
       }
     }
-  }
-
-  clear();
-
-  use_default_colors();
-
-  addstr("\n");
 
-  int nb_printed_lines = 0;
-
-  // Here new_line is either a line number matching the patterns, or -1
+    // Here new_line is either a line number matching the patterns, or -1
 
-  if(new_line >= 0) {
+    if(new_line >= 0) {
 
-    int first_line = new_line, last_line = new_line, nb_match = 1;
+      int first_line = new_line, last_line = new_line, nb_match = 1;
 
-    // We find the first and last line to show, so that the total of
-    // visible lines between them (them include) is console_height - 1
+      // We find the first and last line to show, so that the total of
+      // visible lines between them (them include) is console_height - 1
 
-    while(nb_match < console_height-1 && (first_line > 0 || last_line < nb_lines - 1)) {
+      while(nb_match < console_height-1 && (first_line > 0 || last_line < nb_lines - 1)) {
 
-      if(first_line > 0) {
-        first_line--;
-        while(first_line > 0 && !match(lines[first_line], nb_patterns, patterns)) {
+        if(first_line > 0) {
           first_line--;
+          while(first_line > 0 && !match(lines[first_line], &matcher)) {
+            first_line--;
+          }
+          if(match(lines[first_line], &matcher)) {
+            nb_match++;
+          }
         }
-        if(match(lines[first_line], nb_patterns, patterns)) {
-          nb_match++;
-        }
-      }
 
-      if(nb_match < console_height - 1 && last_line < nb_lines - 1) {
-        last_line++;
-        while(last_line < nb_lines - 1 && !match(lines[last_line], nb_patterns, patterns)) {
+        if(nb_match < console_height - 1 && last_line < nb_lines - 1) {
           last_line++;
-        }
+          while(last_line < nb_lines - 1 && !match(lines[last_line], &matcher)) {
+            last_line++;
+          }
 
-        if(match(lines[last_line], nb_patterns, patterns)) {
-          nb_match++;
+          if(match(lines[last_line], &matcher)) {
+            nb_match++;
+          }
         }
       }
-    }
 
-    // Now we display them
+      // Now we display them
 
-    for(int l = first_line; l <= last_line; l++) {
-      if(match(lines[l], nb_patterns, patterns)) {
-        int k = 0;
+      for(int l = first_line; l <= last_line; l++) {
+        if(match(lines[l], &matcher)) {
+          int k = 0;
 
-        while(lines[l][k] && k < buffer_size - 2 && k < console_width - 2) {
-          buffer[k] = lines[l][k];
-          k++;
-        }
-
-        // We fill the rest of the line with blanks if either we did
-        // not clear() or if this is the highlighted line
-
-        if(l == new_line) {
-          while(k < console_width) {
-            buffer[k++] = ' ';
+          while(lines[l][k] && k < buffer_size - 2 && k < console_width - 2) {
+            buffer[k] = lines[l][k];
+            k++;
           }
-        }
 
-        buffer[k++] = '\n';
-        buffer[k++] = '\0';
+          // We fill the rest of the line with blanks if either we did
+          // not clear() or if this is the highlighted line
 
-        // Highlight the highlighted line ...
+          if(l == new_line) {
+            while(k < console_width) {
+              buffer[k++] = ' ';
+            }
+          }
 
-        if(l == new_line) {
-          if(with_colors) {
-            attron(COLOR_PAIR(2));
-            addnstr(buffer, console_width);
-            attroff(COLOR_PAIR(2));
+          buffer[k++] = '\n';
+          buffer[k++] = '\0';
+
+          // Highlight the highlighted line ...
+
+          if(l == new_line) {
+            if(with_colors) {
+              attron(COLOR_PAIR(2));
+              addnstr(buffer, console_width);
+              attroff(COLOR_PAIR(2));
+            } else {
+              attron(A_STANDOUT);
+              addnstr(buffer, console_width);
+              attroff(A_STANDOUT);
+            }
           } else {
-            attron(A_STANDOUT);
             addnstr(buffer, console_width);
-            attroff(A_STANDOUT);
           }
-        } else {
-          addnstr(buffer, console_width);
+
+          nb_printed_lines++;
         }
+      }
 
-        nb_printed_lines++;
+      if(motion != 0) {
+        *current_line = new_line;
       }
     }
 
-    if(motion != 0) {
-      *current_line = new_line;
-    }
-  }
+    *temporary_line = new_line;
 
-  *temporary_line = new_line;
-
-  if(nb_printed_lines == 0) {
-    addnstr("[no selection]\n", console_width);
+    if(nb_printed_lines == 0) {
+      addnstr("[no selection]\n", console_width);
+    }
   }
 
   // Draw the modeline
 
-  sprintf(buffer, "%d/%d pattern: %s",
+  sprintf(buffer, "%d/%d pattern: %s%s",
           nb_printed_lines,
           nb_lines,
-          pattern_list);
+          pattern_list,
+          use_regexp ? " [regexp]" : "");
 
   for(int k = strlen(buffer); k < console_width; k++) buffer[k] = ' ';
   buffer[console_width] = '\0';
@@ -304,6 +375,7 @@ void update_screen(int *current_line, int *temporary_line, int motion,
   // We are done
 
   refresh();
+  free_matcher(&matcher);
 }
 
 //////////////////////////////////////////////////////////////////////
@@ -371,11 +443,16 @@ int main(int argc, char **argv) {
       i++;
     }
 
-    else if(strcmp(argv[i], "-r") == 0) {
+    else if(strcmp(argv[i], "-d") == 0) {
       remove_duplicates = 1;
       i++;
     }
 
+    else if(strcmp(argv[i], "-e") == 0) {
+      use_regexp = 1;
+      i++;
+    }
+
     else if(strcmp(argv[i], "-l") == 0) {
       check_opt(argc, argv, i, 1, "<maximum number of lines>");
       nb_lines_max = atoi(argv[i+1]);
@@ -400,7 +477,8 @@ int main(int argc, char **argv) {
            << " [-h]"
            << " [-v]"
            << " [-m]"
-           << " [-r]"
+           << " [-d]"
+           << " [-e]"
            << " [-z]"
            << " [-i]"
            << " [-c <fg modeline> <bg modeline> <fg highlight> <bg highlight>]"
@@ -433,10 +511,18 @@ int main(int argc, char **argv) {
     return 1;
   }
 
+  int hash_table_size = nb_lines_max * 10;
+  int *hash_table = 0;
+
+  if(remove_duplicates) {
+    hash_table = new_hash_table(hash_table_size);
+  }
+
   while(nb_lines < nb_lines_max && !file.eof()) {
     file.getline(buffer, buffer_size);
     if(strcmp(buffer, "") != 0) {
       char *s = buffer;
+
       if(zsh_history && *s == ':') {
         while(*s && *s != ';') s++;
         if(*s == ';') s++;
@@ -446,7 +532,7 @@ int main(int argc, char **argv) {
         while(*s == ' ' || (*s >= '0' && *s <= '9')) s++;
       }
 
-      if(!remove_duplicates || nb_lines == 0 || strcmp(lines[nb_lines - 1], s)) {
+      if(!hash_table || !test_and_add(s, nb_lines, lines, hash_table, hash_table_size)) {
         lines[nb_lines] = new char[strlen(s) + 1];
         strcpy(lines[nb_lines], s);
         nb_lines++;
@@ -454,6 +540,8 @@ int main(int argc, char **argv) {
     }
   }
 
+  delete[] hash_table;
+
   if(inverse_order) {
     for(int i = 0; i < nb_lines/2; i++) {
       char *s = lines[nb_lines - 1 - i];
@@ -537,6 +625,10 @@ int main(int argc, char **argv) {
       motion = -1;
     }
 
+    else if(key == '') {
+      use_regexp = !use_regexp;
+    }
+
     else if(key == KEY_DOWN || key == '') {
       motion = 1;
     }
@@ -578,6 +670,7 @@ int main(int argc, char **argv) {
   for(int l = 0; l < nb_lines; l++) {
     delete[] lines[l];
   }
+
   delete[] lines;
 
   exit(0);