Page MenuHomeFreeBSD

fs/tmpfs/t_times.sh: remove local modification
AbandonedPublic

Authored by ngie on Sun, Feb 15, 7:03 PM.
Tags
None
Referenced Files
F145774625: D55290.diff
Tue, Feb 24, 8:01 AM
Unknown Object (File)
Mon, Feb 23, 9:58 PM
Unknown Object (File)
Sun, Feb 22, 8:28 PM
Unknown Object (File)
Sat, Feb 21, 9:55 PM
Unknown Object (File)
Sat, Feb 21, 9:40 AM
Unknown Object (File)
Thu, Feb 19, 11:06 PM
Unknown Object (File)
Tue, Feb 17, 6:39 AM
Unknown Object (File)
Tue, Feb 17, 3:45 AM
Subscribers

Details

Reviewers
des
Summary

This particular change was not accepted upstream and does not seem to be
needed. Don't carry it forward.

MFC after: 1 week

Test Plan

The test continues to pass after this line is removed.

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
Tests Skipped
Build Status
Buildable 70679
Build 67562: arc lint + arc unit

Event Timeline

ngie requested review of this revision.Sun, Feb 15, 7:03 PM
ngie added a reviewer: des.
des requested changes to this revision.Sun, Feb 15, 8:08 PM

This is a textbook example of Chesterton's Fence. I suggest you add a comment referencing bug 274615 instead.

This revision now requires changes to proceed.Sun, Feb 15, 8:08 PM
In D55290#1264539, @des wrote:

This is a textbook example of Chesterton's Fence. I suggest you add a comment referencing bug 274615 instead.

I suspected that might be the case, but I wanted to confirm via this differential.

I'll submit this upstream.

contrib/netbsd-tests/fs/tmpfs/t_times.sh
55

I'm going to retract my prior statement: this actually shouldn't be required here (anymore) since the assertions after line 56 (the echo foo >a) are actually checking to make sure the reported values are greater than the previous values (it mattered for the prior steps, e.g., line 50-52, since it was testing equality). All that being said, the ost values should really be reset between each operation to ensure sanity comparing the checkpoints in time between the various assertions.

It would have been helpful if the messages had helped note which test case was failing, but it seems pretty obvious based on the code that ost_atime prior to this would always be greater than it was before, unless the clock on the system went backwards somehow.

This revision now requires changes to proceed.Mon, Feb 23, 6:39 AM
contrib/netbsd-tests/fs/tmpfs/t_times.sh
55

It absolutely _is_ needed to verify that cat a above updated the atime.

contrib/netbsd-tests/fs/tmpfs/t_times.sh
55

correction: to verify that echo >a below updates the atime

ngie marked 2 inline comments as done.
ngie added inline comments.
contrib/netbsd-tests/fs/tmpfs/t_times.sh
55

correction: to verify that echo >a below updates the atime

You're right. Even though technically the assertion above should still pass, you want to ensure that the value changes. Adding this update ensures that the variable changes when echo foo>a is run.

(it's been fun keeping stuff straight this week 🤪..)

I'll file another NetBSD PR to update this too, I suppose.