2012-06-06, 19:12
OK that initial run resulted in it compiling GIT:20120604-6b2bb78. Got an immediate crash as soon as I tried to play a movie. Moving on to the next one.
(2012-06-06, 20:24)bobo1on1 Wrote: Possibly fixed in https://github.com/xbmc/xbmc/commit/b6e1...b1c16d2650
(2012-06-06, 20:41)amet Wrote: Why is it a messy code? it's done in he same coding style as everything else...
Also, it compiles and runs just fine on osx, no crashes in about a week. PR has been up on github for 2 or so weeks and no one could be bothered to compile on Linux , it got signed off and vdrfan fixed it in an hour as it was not compiling on Linux.
I can't test it on every platform, that's why the PR was posted....
(2012-06-06, 20:58)john.doe Wrote: Well here are just a few of the problems:
* std:trlen instead of strlen
* mismatch in several prototypes and their actual definitions, such as InitVFSTypes(bool bInitTypes) vs InitVFSTypes(void)
* still crashing even after prototypes are fixed, meaning there are more bugs
(2012-06-06, 19:29)john.doe Wrote: May I also say that DAMN that "extend xbmcvfs" is badly coded. Such messy code.
(2012-06-06, 21:36)davilla Wrote:(2012-06-06, 19:29)john.doe Wrote: May I also say that DAMN that "extend xbmcvfs" is badly coded. Such messy code.
Unless you intend to start submitting patches or contributing to pull/reqs, how about backing off on the ponking please. Nightlies can be volatile, we try our best but sometimes opps will slip in.
(2012-06-06, 22:16)bobo1on1 Wrote:(2012-06-06, 20:58)john.doe Wrote: * still crashing even after prototypes are fixed, meaning there are more bugs
The fact that reverting the commit makes the crashes go away is not proof that the vfs code is responsible for the crashes.