I think the file has a normal UCS-2 encoding. Your git client is
configured to transform between Unix and Windows line endings on
checkouts and commits. There seems to be a bug in your git client to not
handle UCS-2 correctly: It detects only the 0x0A as a Unix line ending
and does not notice that the 0x0D is two bytes before the 0x0A due to
the UCS-2 encoding. Therefore it converts the 0x0A to a 0x0D 0x0A (\R
\n). This destroys the UCS-2 encoding as your git client only inserts a
single byte and then the alignment of the 2-byte-characters is wrong in
every second line.
If you have a current version of your git client, you should probably
file a bug there. I think that the git client should notice that 0x00
0x0D 0x00 0x0A is in fact a Windows line ending in UCS-2 encoding and
not double-convert it to the Windows line ending format. I could not
reproduce the behaviour on my machine as I configured my git client to
not convert anything.
On 30.01.2014 05:47, Sid Shetye wrote:
Can someone confirm there is nothing malicious in
\bc-sharp\crypto\src\asn1\isismtt\x509\AdmissionSyntax.cs beyond incorrect
encoding as UCS-2 ? This message is invisible online @ GitHub; you need to
clone the repo locally and open it in Notepad++ or a hex editor.
I’ve yet to diff the GIT code off a golden reference, not sure if there are
other such files. Inline image below bounced off the mailing list, so the image
link is athttp://i.imgur.com/IfMzTew.png
