On Fri, Nov 25, 2016 at 01:38:02PM +0100, Hugo González wrote: > Apart from that, I tried to fix the diff file myself, just to check if the > error was only during the generation of the diff file or also to apply an > Unicode diff file. I edited the changes.diff file using Notepad++ and > removed the \0 codes, so I had an ANSI diff file called > changes_fixed.diff that also fails in the same way to apply the patch. I > also converted the ANSI diff file into an Unicode UCS2-LittleEndian file > with Notepad++, and now Subversion cannot apply the patch So, it seems > there could be three bugs there: one in the creation of diff files, another > one applying ANSI diff files to Unicode files, and another just working with > Unicode diff files. I attach all the examples again, with the manually fixed > diff files. Best regards, > > > > Hugo González
Hi Hugo, Subversion only supports Unicode it if is encoded as UTF-8. Unicode represented in encodings other than UTF-8 is usually treated like a binary file. Fixing this would be a major effort. See https://issues.apache.org/jira/browse/SVN-2194 If you convert files you want to use with 'svn patch' to UTF-8, then things should "just work". Note that conversion between UTF-8 and UTF-16 is generally lossless, so in the worst case this would add extra (but nonetheless inconvenient) conversion steps to your workflow. Regards, Stefan
