Rename getPath() to getRealPath() and getAllPaths() to getAllRealPaths()

[lenton]

This is a much more descriptive name for the two methods rather than the generic 'getPath'. It makes it much clearer that these methods get the real file path from their virtual one.

[lenton]

It seems HHVM doesn't work with vfsStream and it's not going to be fixed any time soon: facebook/hhvm#1971

I guess we're going to have to remove the HHVM travis check for now, unless anyone knows a work around?

[lenton]

I've made it so HHVM is now not required to pass its build: 1b2e619

Although when restarting the build it doesn't take into account the change, so you'll just have to ignore the failure for this PR.

[acoulton]

@lenton how very dull. The only workaround I can think of would be to drop vfsStream and make a temporary directory structure in the sys_get_temp_dir but that's not great either. An idealistic solution would use vfsStream on other platforms and a real filesystem on HHVM I guess...

As is, I think we should just skip the known bad tests - otherwise in future something we care about could fail on HHVM but the commit could still get a green tick. See #9

To get Travis to consider the changed configuration (either yours in master or mine), you'd need to rebase this branch - just restarting the build will build the same commit with the same travis configuration.

[lenton]

@acoulton Thanks, that's a better work around.

To get Travis to consider the changed configuration (either yours in master or mine), you'd need to rebase this branch - just restarting the build will build the same commit with the same travis configuration.

😞 How silly of me, that never crossed my mind.