Closed leons727 closed 1 year ago
Very nice! Is there a way to get a test for this? I understand if it's too tricky to write, and I doubt anybody will break this feature in the future...
Also, you might want to rebase main, so you'll get rid of the clippy/chrono warning-errors.
There is probably a way to test this by generating bulk inserts with various sizes, for a table with a single BIT
column each row takes 3 bytes, so it'd be possible to construct an insert that trips buffer after appending TokenDone
. But that would be a bit fragile setup, and I am not sure if it's worth it...
Regarding rebasing main -- not sure what you mean -- my fork is based on main... I did miss running format, updated PR with format ran.
There is probably a way to test this by generating bulk inserts with various sizes, for a table with a single
BIT
column each row takes 3 bytes, so it'd be possible to construct an insert that trips buffer after appendingTokenDone
. But that would be a bit fragile setup, and I am not sure if it's worth it...Regarding rebasing main -- not sure what you mean -- my fork is based on main... I did miss running format, updated PR with format ran.
I mean there are some CI errors from the latest rust/clippy and chrono deprecations. Fetch our main branch, then on yours do git rebase main
so the tests on this PR are green and we can merge.
Ah, I thought I've done that, should be up to date now.
The new version is out. Thank you.
This fixes the issue when appending
TokenDone
to the buffer infinalize()
method causes buffer length exceed packet size.