linux watcher bug fixes and general slight improvements

[whyvn]

resolves #3

if you do not wish to merge this pr the bug fixes can still be merged

[whyvn]

1433508
and
f077375

are ideas for "invalid" watch descriptors although i couldn't get them to work (more information is provided in the commit paragraphs). why does inotify ignore updating future wds even if you remove a watch target :(

all commits up until here have just been bug fixes (and changing Event but i think that's pretty useful)

also someone must verify that the macos build is fine

[freref]

Hi Julia, thanks for your contributions, I will take a look at this after my finals as I would also have to adjust the macos build

[whyvn]

oh dont worry about it yet. i still need to implement the actual feature of this pr and i will make the draft ready once i get home and do that!

[whyvn]

Ok, ive come to the conclusion that, although i can see a use for this idea, its too niche to warrant extending the api in such a way that makes it extremely clunky (e.g. a hypothetical Watcher.poll() would have to reset Watcher.modified to prevent a race condition from the main event loop however what if the user wants to call poll multiple times? or what if they want to check modified without poll?)

also, the way it would work is not too far off with the user implementing it themselves like in #3. you still have to call poll in the main event loop and call the function from there unless we do fancy meta programming so yeah. ill just keep this pr as bug fixes i guess.

also, ive changed Event to include an item field which i think would be useful.

[freref]

Closing in favor of #5 which continues this work