4 ms·
Awesome, love to see more Java devs pushing back on the spring madness. In Db.java, with the write/update/delete locks, that mutate both in memory and on disk,
by jabradoodle 3y ago
Awesome, love to see more Java devs pushing back on the spring madness.
In Db.java, with the write/update/delete locks, that mutate both in memory and on disk, is there not a bunch of race conditions there?
- 7ep 3y agoThere are locks around those - for example, writeLock
- jabradoodle 3y agoI mean race conditions between the operations, e.g. an update followed by a delete: 1. update x in memory 2. delete x in memory 3. update fn checks for file on disk 4. delete x from disk 5. update x on disk Is this not possible and you end up with different state in memory and on disk.
- 7ep 3y agohmmm. it does seem possible. Let me look into that, thanks!
- 7ep 3y agogreat find. i’ve got some adjustments in mind, mainly carefully placed locks. i’d like to give you credit for the call out in the docs/acknowledgment file. is there some better way to reference you than jabradoodle? the code i’m considering is in the branch under_consideration.
- jabradoodle 3y agoThanks that is kind of you. For the time being I would rather not tie my real life identity to my Internet identity.