mirrors/jj
1
0
Fork 1
mirror of https://github.com/martinvonz/jj.git synced 2025-01-15 00:44:33 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 13

tests: add upper bound to test_concurrent_read_write_commit() loop

Browse source 
Hopefully this will fix the unfinished Windows CI issue. A possible scenario
is that recent migration to gitoxide made this test flaky on Windows. For
example, gitoxide might have in-memory object cache that relies on file mtime,
and occasionally fails to detect new object on Windows.
This commit is contained in:
Yuya Nishihara 2023-11-24 12:33:32 +09:00
parent e79c8b6f62
commit b37293fa68
1 changed files with 15 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
lib/tests/test_git.rs
View file

@ -2909,7 +2909,12 @@ fn test_concurrent_read_write_commit() {
pending_commit_ids.rotate_left(i); // start lookup from different place
s.spawn(move || {
barrier.wait();
while !pending_commit_ids.is_empty() {
// This loop should finish within a couple of retries, but terminate in case
// it doesn't.
for _ in 0..100 {
if pending_commit_ids.is_empty() {
break;
}
repo = repo.reload_at_head(settings).unwrap();
let git_backend = get_git_backend(&repo);
let mut tx = repo.start_transaction(settings, &format!("reader {i}"));
@ -2936,6 +2941,15 @@ fn test_concurrent_read_write_commit() {
}
thread::yield_now();
}
if !pending_commit_ids.is_empty() {
// It's not an error if some of the readers couldn't observe the commits. It's
// unlikely, but possible if the git backend had strong negative object cache
// for example.
eprintln!(
"reader {i} couldn't observe the following commits: \
{pending_commit_ids:#?}"
);
}
});
}
});