Conversation
We're still checking against the 'é' symbol. The resource filename has been renamed, as some systems can't deal with the symbol and therefore produce NPE.
|
I actually had a non standard file name, to ensure the library was correctly encoding the filename strings. Which platform is this broken on? And can we instead fix it to work? |
|
@bramp Testing for special characters is reasonable. However, this PR is not concerned with the code itself. The revert in commit c4c5e1b makes sure the test still handles special symbols. |
|
As a reference, on a system with posix locale, the produced filename is: This PR is only concerned with not testing OS behaviour we can't control |
|
So this sounds like the file system doesn't support the character. Which I find very surprising in this day and age of UTF-8. I of course don't want to test the OS, but I do want to test we interact with the OS correctly. But ok, I see the changes to the test, are really just fixing the test, not breaking the behaviour of reading non-ascii named media. |
Some systems don't allow the non asci letter é in file-names.
This results in the testProbeVide2 test failing, as the related file can't be found.
This commit renames the file, in order to allow the test to pass.
The filename wthin the ffmpeg output is not altered, to keep testing the parsing of such names