UTF-8 edit ini file

[Grovkillen]

If I create a ini file with UTF-8 activated the edit ini file command aren't complying. See this test:

Code: Select all

WriteLn>%SCRIPT_DIR%\inifile1.ini,,######
EditIniFile>%SCRIPT_DIR%\inifile1.ini,section,entry,åäöÅÄÖ

Let>WLN_ENCODING=UTF8
WriteLn>%SCRIPT_DIR%\inifile2.ini,,######
EditIniFile>%SCRIPT_DIR%\inifile2.ini,section,entry,åäöÅÄÖ

ExecuteFile>%SCRIPT_DIR%\inifile1.ini
ExecuteFile>%SCRIPT_DIR%\inifile2.ini

[Marcus Tettmar]

Hi,

Looks like EditIniFile isn't UTF-8/Unicode compatible. Have added to list.

[Grovkillen]

Hi,

Looks like EditIniFile isn't UTF-8/Unicode compatible. Have added to list.

Thanks

[Grovkillen]

FYI: the read ini file command isn't parsing UTF-8 as well.

[Grovkillen]

To devs: is this on the horizon (as in a fix)? Or should I create a separate "clean up script" for this? I would rather it be the first options since I'm relying on the ini files to only be updated if there's a difference.. this could trigger false positives since I cannot know for sure if the value that differ is in fact a correct (cleaned) one.

[Marcus Tettmar]

Not on the immediate horizon. Have looked into it and the hard fact of the matter is that Windows doesn't natively support utf8 ini files either. We're wrapping the Windows Api functions which read/write to INI files. They don't do it, so therefore nor do we. We would therefore need to reinvent the wheel from scratch and write our own INI file handling functions. Microsoft may argue that INI files are old tech and that use of the registry supersedes them. I'll need to look into this more and weigh up the effort against the need.

[Grovkillen]

Oh, please don't be like Microsoft then I love my INI files. They are so portable and easy to backup compared to the registry.

[Grovkillen]

This is how I parse a ini file in JS. Perhaps something that could be done with MS?

Code: Select all

    'iniFileToObject': function(string, pipe2array = false) {
      let object = {};
      let sections = string.match(/^\[[^\]\r\n]+](?:[\r\n]([^[\r\n].*)?)*/gm);
      for (let i=0; i < sections.length; i++) {
          let sectionName = sections[i].split("\n")[0];
          sectionName = sectionName.replace(/[\[\]]/g, '').trim();
          object[sectionName] = {};
          let key = sections[i].match(/^((?!\/\*[\s\S]*?\*\/|([^\\:]|^)\/\/.*$).)+/gm);  //we remove comments "//" syntax, single row
          for (let k=1; k < key.length; k++) {
              let keyValue = key[k].split("=");
              if ((keyValue[1].charAt(0) === "0" && keyValue[1].length > 1) || isNaN(Number(keyValue[1]))) {   // leading zeros are interpreted as string values
                  if (pipe2array && keyValue[1].trim().includes("|")) {
                      object[sectionName][keyValue[0].trim()] = keyValue[1].trim().split("|");
                  } else if (keyValue[1].trim() === "null") {
                      object[sectionName][keyValue[0].trim()] = null;
                  } else {
                      object[sectionName][keyValue[0].trim()] = keyValue[1].trim();
                  }
              } else {
                  object[sectionName][keyValue[0].trim()] = Number(keyValue[1]);
              }
          }
      }
      return object;
    }

I have added a "pipe2array" thing to be able to have some keys automatically turned into arrays by separating them with a | "pipe".

This parser allow for comments on single rows or on key value rows using "//". Section rows must not contain any comments.

Writing to a ini file would be just to check if a section and key is already present, if so replace the string of that key with the new value, remember the comment if any and add it to the end of the string.

If no section and/or key is found, inject a new line with this information.

For easier parsing you may even allow for the ini file comments to be part of the ini object as simply add them as such:

SECTION1=section1s_name
SECTION1_ROW=row_of_section_in_ini_file
SECTION1_KEY1=keyname
SECTION1_KEY1_ROW=row_of_key_in_ini_file
SECTION1_KEY1_VALUE=value_of_key1
SECTION1_KEY1_COMMENT=comment_found_after_value
...
SECTIONn=section1s_name
SECTIONn_ROW=row_of_section_in_ini_file
SECTIONn_KEYn=keyname
SECTIONn_KEYn_ROW=row_of_key_in_ini_file
SECTIONn_KEYn_VALUE=value_of_keyN
SECTIONn_KEYn_COMMENT=comment_found_after_value
+
INI_FILE_COMMENT1=comment_string
INI_FILE_COMMENT1_ROW=row_of_comment_in_ini_file
...
INI_FILE_COMMENTn=comment_string
INI_FILE_COMMENTn_ROW=row_of_comment_in_ini_file

Just as an idea of how you could implement it, I'm not suggesting you should allow for the user to see this information.

You could then say that a ini file syntax must be used as such:

Sections [section] must reside on a single row with no comments, spaces before and after a section entry will be trimmed away.
Key-value entries must reside on a single row, spaces before and after a key will be trimmed away.
Key must be followed by a = or else the row will be treated as a comment.
Values must be single row and everything after a "//" is treated as a comment.

This way you wouldn't reinvent the wheel, simply make a better propulsion mechanism. And I would love MS even more

[Grovkillen]

Okay, I really hit rock bottom with this problem. I want to read the INI file and lookup a folder structure based on info in the INI. The illegal characters will not work when parsing the path to the file based on text with none-UTF-8 type.

[Grovkillen]

Okay, I managed to get my server side stuff to work by moving the encoding to the front-end. I read the ini files and parse them as binary data. So I'm all right with that solution. So no immediate hurry for my sake

I.E. I now use the default encoding (ANSI) and the non-ANSI characters are converted in the front-end instead of on server side script.

[hagchr]

Good you have a working solution. Not sure if relevant/helps, I played around and note that if I create and save an empty inifile2.ini in Notepad (as UTF-8), then when running your script in MS it looks ok (shows åäö).

[Grovkillen]

Good you have a working solution. Not sure if relevant/helps, I played around and note that if I create and save an empty inifile2.ini in Notepad (as UTF-8), then when running your script in MS it looks ok (shows åäö).

Thanks for testing. I didn't do that test but now I learned how to parse binary data in JavaScript. That's something positive