Oh no. I had been wondering for a while why this code was ever written this way. It seems to so clearly have the exact flaw that we are trying to fix. Well, this test explains why. Some of these write errors may be transient failures that will be resolve when a lock is released. In _that_ case, it does actually make sense to try again, though I don't know that trying again infinitely was a good idea. I think that the way that this ought to work is to, if possible, detect these transient failures and use a different error code if they are detected. Then, in that case, we can try again up to some maximum number of times. This project just keeps getting more and more complex than I originally intended for it to be. Let's chat with Amir about how to proceed here.
Bug 1853821 Comment 7 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Oh no. I had been wondering for a while why this code was ever written this way. It seems to so clearly have the exact flaw that we are trying to fix. Well, this test explains why. Some of these write errors may be transient failures that will be resolved when a lock is released. In _that_ case, it does actually make sense to try again, though I don't know that trying again infinitely was a good idea. I think that the way that this ought to work is to, if possible, detect these transient failures and use a different error code if they are detected. Then, in that case, we can try again up to some maximum number of times. This project just keeps getting more and more complex than I originally intended for it to be. Let's chat with Amir about how to proceed here.