Sotos wrote:ok ... I see what is all this about now. It is your "save designs" thing. What I don't understand is how those unwanted characters end up in your files to begin with. Are they entered by the user? If that's the case then you should be doing the validation and any conversion before you write the file, not when you read it. You should also make sure you use the same encoding everywhere. In the head of your HTML (<meta charset="utf-8">), in the fs read/write functions (sync or async shouldn't make a difference), maybe there is a setting in Chromium/node-webkit for setting character encoding? (never used those myself). If everything is set correctly then maybe there is bug in that stack you are using... a quick search turned this: http://stackoverflow.com/questions/2329 ... ode-webkit So maybe try utf16le for the read/write functions.
The offending file-opening function I posted belongs to a new generator I’ve created. This one reads in a text file and converts it into a hypertext HTML5. (Image below)
In theory it sounds relatively easy to do but it turned out to be a very challenging generator to create.
All the research and exhaustive tests I’ve done since the last time we spoke points to a node-webkit bug like you found on SO, although I'm using v0.10.4… man I’m gonna kill that little chink who wrote it!