Remove duplicated chars from user input in []

Since `[aabbcc]` is the same as `[abc]`, txt2regex will remove the
repeated characters the user may have typed to put inside a list.

It's safe to do it because in `getCharList()` we only expect a list of
literal characters, not ranges or negated list identifier `^`.

CHANGELOG.md

@@ -65,6 +65,8 @@ The format is based on [Keep a Changelog].
 - lex regexes: now using GNU flex
 - PHP regexes: switch from old `ereg` to `preg` (PCRE)
 - Changed the default programs: +egrep +grep -perl -php -postgres
+- Remove repeated characters inside a list `[]` (if the user has typed
+  `abbbca`, make it `[abc]`)
 - Now `--showmeta` also shows the version for each program
 - Now the "!! not supported" legend only appears when there are
   unsupported metacharacters in the current regex

tests/features.md

@@ -24,6 +24,17 @@ $ txt2regex --prog egrep --history '215¤a5!6'
 $
 ```
 
+## User input: Remove duplicated chars from [] — getCharList()
+
+When informing literal characters to be put inside a `[]` list, txt2regex will deduplicate those characters, because the repetition in this case is not meaningful (`[aabbcc]` is the same as `[abc]`).
+
+```console
+$ txt2regex --prog egrep --history '24¤aabbbcab'
+ Regex egrep: [abc]
+
+$
+```
+
 ## User input: Rearrange [] special elements — getCharList()
 
 When informing literal characters to be put inside a `[]` list, some special cases have to be handled:

txt2regex.sh

@@ -621,6 +621,19 @@ charInText() {
     return 1
 }
 
+# Remove all duplicated chars from the $1 text
+uniqChars() {
+    local text="$1"
+    local text_uniq=''
+    local i
+
+    for ((i = 0; i < ${#text}; i++)); do
+        charInText "${text:$i:1}" "$text_uniq" ||
+            text_uniq="$text_uniq${text:$i:1}"
+    done
+    printf '%s\n' "$text_uniq"
+}
+
 # Escape each $1 in $2 using $3
 escapeChars() {
     local special_chars="$1"
@@ -987,7 +1000,6 @@ getChar() {
     F_ESCCHAR=1
 }
 
-#TODO 1st of all, take out repeated chars
 getCharList() {
     gotoxy $x_prompt2 $y_prompt
 
@@ -999,6 +1011,10 @@ getCharList() {
         doNextHistArg
         uin=$histarg
     fi
+
+    # dedup is safe because $uin contains only literal chars (no ranges)
+    uin="$(uniqChars "$uin")"
+
     uins="${uins}¤$uin"
 
     # putting not special chars in not special places: [][^-]
@@ -1009,6 +1025,7 @@ getCharList() {
     # if any $1, negated list
     [ -n "$1" ] && uin="^$uin"
 
+    # make it a list
     uin="[$uin]"
     F_ESCCHARLIST=1
 }