Partial? Fix for panic when opening binary files

[wilsonk]

Not completely sure if this the right way to do this, but it seems to work for the few file types I have tested and directories (that is part of the problem...we may want to be able to traverse directories?). Anyways, nothing happens if you try to choose a binary/non-text file to open.

Perhaps we should also change the 'Search:' line to a different color, or show some indication that the open failed?

Fixes(kindof) #191 (LIL-59)

[tauraamui]

I'm about to go to bed, but I'm sure this is fine for now (I'll double check tomorrow). I'd like to refer you to this open branch tho, https://github.com/tauraamui/lilly/tree/spike/investigate-file-encodings/encoding-test this is a collection files of which 99% of will fatal crash Lilly on open, due to rendering problem. However, I think the default behaviour of Lilly should be to reject or refuse to open any file whose encoding it is unable to infer from a BOM or otherwise, eg., anything outside of all major encoding present in that examples directory.

Our inability to render these documents is a separate issue, in the meantime we should eventually add some kind of file inspection stage to loading a workspace where we ahead of time identify each files encoding type, and based on this we either:

• just exclude files from the list (I don't like that idea)
• make them disabled entries in the list, ie., you can't open them
• or some kind of error output response on attempt to open them

[wilsonk]

I definitely like option 2 of those options (disabled entries). I will check out that branch also.

[tauraamui]