diff options
author | Roman / Linnea Gräf <roman.graef@gmail.com> | 2023-02-18 15:24:52 +0100 |
---|---|---|
committer | GitHub <noreply@github.com> | 2023-02-18 15:24:52 +0100 |
commit | 0d281d483909d71272783033b2aba8f33dcbce36 (patch) | |
tree | e59f05fe4b22f989b5a600e66d7b8c9aaecda37c /buildSrc/moulsign.sh | |
parent | b12da67a08b6808363c0ad130fe8c3bc6f00f7b2 (diff) | |
download | NotEnoughUpdates-0d281d483909d71272783033b2aba8f33dcbce36.tar.gz NotEnoughUpdates-0d281d483909d71272783033b2aba8f33dcbce36.tar.bz2 NotEnoughUpdates-0d281d483909d71272783033b2aba8f33dcbce36.zip |
ApiCache: Fix apparent NullPointerException (#619)
`CacheResult` should in theory have either a `file != null` a `future !=
null` or be `disposed`. Apparently this invariant of `CacheResult` is
either being violated somewhere, or the `synchronized` blocks arent as
synchronized as id hoped they were. In fact, `dispose()` does not even
delete the file, so i can really only see this happening because the
first `synchronized` block that writes the file and the second
`synchronized` block that reads from the file hold the same lock.
I have no idea how this would happen, but hopefully this fixes it (since
the dispose didn't have a threading issue reported so far, i feel more
confident leaving the .deleteOnExit in there, but I'm also wrapping any
potential IOExceptions during access, because I am just so confused how
the internal state was broken.
Diffstat (limited to 'buildSrc/moulsign.sh')
0 files changed, 0 insertions, 0 deletions