[ggo-discussion] Re: a new bug... (0.3.9pre2)

  • From: Peter Strempel <zotan@xxxxxx>
  • To: ggo-discussion@xxxxxxxxxxxxx
  • Date: Wed, 14 May 2003 17:55:16 +0200

Found, fixed.


The old version used an array of chars. This is so easy and dumb even
Windows did not manage to mess this up.

Now, when sending a string there are wonderful ways to mess with the
encoding. When sending through the pipe on Linux, a plain US-ASCII encoding
is used, no problem. However, when sending through the pipe on Windows, some
smartass at Redmond probably decided the Windows sockets need to alter the
string encoding from US-ASCII to Cp1252, the
Microsoft-selfmade-only-exists-on-Windoze Latin1 encoding. The recieving
side of the socket, expecting an US-ASCII encoded string, now was presented
with a Cp1252 encoded string, magically transformed by the Oh So Great And
Exiting Operating System.


I force it to UTF-8 now to solve the issue. Drawback is, that won't work anymore
on Windoze 95 or 98, but I dont care.


Other related posts: