Commit d18cd8b
test(e2e): 340 只数它自己那个目录里的 ninja (#557)
THE PREDICATE WAS RIGHT AND THE OBJECT WAS WRONG.
`count_ninja` counted every ninja on the machine. A baseline taken before the
build covers one that was ALREADY running; it does not cover one that appears
during the window, and a build server, a second checkout or a parallel test can
start one at any moment.
Measured in an ecosystem sandbox, which shares the host's PID namespace: an
unrelated session's ninja, in a different project entirely, made this assertion
fail on a build that has the fix.
Counting is now scoped by the process's working directory to the test's own
`$work` tree. Verified that scoping did not weaken it: the test still FAILS
against the released 2026.9.4.2, naming the orphan and its directory.
Co-authored-by: speak-agent <248744407+speak-agent@users.noreply.github.com>1 parent d3773e7 commit d18cd8b
1 file changed
Lines changed: 14 additions & 3 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
33 | 33 | | |
34 | 34 | | |
35 | 35 | | |
36 | | - | |
37 | | - | |
| 36 | + | |
| 37 | + | |
| 38 | + | |
| 39 | + | |
| 40 | + | |
| 41 | + | |
| 42 | + | |
| 43 | + | |
| 44 | + | |
| 45 | + | |
| 46 | + | |
| 47 | + | |
38 | 48 | | |
39 | 49 | | |
40 | 50 | | |
41 | | - | |
| 51 | + | |
| 52 | + | |
42 | 53 | | |
43 | 54 | | |
44 | 55 | | |
| |||
0 commit comments