"File/Revision Compare" window can't handle UTF-16 SQL set to text/plain (7.5.3)

I commited a file called "Testdaten.sql" that is encoded in UTF-16 (with proper BOM) and set the mime-type to "text/plain".    I changed the file and wanted to view the differences ("Query->Show Changes").    In the "File/Revision Compare" window the right side ("Working Copy") is shown correctly. The left side ("Pristine Copy") is apparently trying to display the file using some ANSI encoding. The BOM is displayed as two question marks and every second char is displayed as a "invalid char block".

Last updated 10 years ago

Yes, this window is really somewhat buggy. Also, if you have things like a copyright symbol (hex 0xA9) in a file, in the comparison view it becomes some weird character. I guess the encoding detection may need some work.  Luckily so far nothing go lost for me... :-)

M4rtin;129996Yes, this window is really somewhat buggy. Also, if you have things like a copyright symbol (hex 0xA9) in a file, in the comparison view it becomes some weird character. I guess the encoding detection may need some work.  Luckily so far nothing go lost for me... :-)

  That would be really really bad, but since it's only the comparer window I wouldn't expect that to happen. :)

Is there any solution to this problem ?  This bug block my team to update to this version (we are still use SmartSVN 6).    Thank you.

We fixed a similar issue in our latest intermediate build. Can you guys try this and see if the issue is resolved? You can get the latest build here: http://www.smartsvn.com/intermediate-builds

Thinks for your answer, I try it but it does not fix my bug.

Hervé Labboz;130443Thinks for your answer, I try it but it does not fix my bug.

 Well for me it fixes the bug with the copyright and special umlauts, for example. So at least its getting better... ;-)

Thank, "smartsvn-build-7_5_3-r13800" work fine. Intermediate build no this problem for file encoding "windows-1251"

hello, first i created a brand new topic with this, but it didn't get posted and i see that more people are having the same problem, so i'm posting the message in this thread instead:  when i browse through my repository and the hit a file with latin characters in it, the built-in compare utility shows malformed text and treats them as different files. however, the files are the same and the file browser shows the state as "Unchanged" too. even after i update the file via smartsvn, the file compare utility shows them as different  the svn storage is a unix server and i am running on a windows machine.  tldr: the files are exactly the same, and the client also tells me so, but the file compare utility treats them as different files 

Version 7.5.4 fixed the bug for me.