1924927 - FolderCompactor doesn't commit the message summary file before making a backup

Status: RESOLVED FIXED

(MailNews Core :: General, defect)

Description

[Geoff Lankow (:darktrojan)]

We could be backing up stale data.

Comment 1

[Geoff Lankow (:darktrojan)]

Attachment: Bug 1924927 - Commit message summary file before backing it up for compaction. r=#thunderbird-reviewers

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/3f74315e90d3
Commit message summary file before backing it up for compaction. r=darktrojan

Comment 3

[Geoff Lankow (:darktrojan)]

Comment on attachment 9431173 [details]
Bug 1924927 - Commit message summary file before backing it up for compaction. r=#thunderbird-reviewers

[Approval Request Comment]
Regression caused by (bug #): bug 1890448 (probably)
User impact if declined: recently deleted messages could become undeleted if compact fails
Testing completed (on c-c, etc.): in 133.0b1
Risk to taking this patch (and alternatives if risky): very low

Comment 4

[Corey Bryant]

Comment on attachment 9431173 [details]
Bug 1924927 - Commit message summary file before backing it up for compaction. r=#thunderbird-reviewers

[Triage Comment]
Approved for esr128

Comment 5

[Daniel Darnell [:dandarnell]]

Thunderbird 128.4.2esr:
https://hg.mozilla.org/releases/comm-esr128/rev/43052c849f97

Comment 6

[Vlad Lucaci, Thunderbird QA (:vlucaci)]

Hello,

Are there any specific STR that QA could perform in order to reproduce the faulty behaviour(crash) on an affected build and properly verify and confirm that it is fixed?

Comment 7

[Corey Bryant]

Comment on attachment 9431173 [details]
Bug 1924927 - Commit message summary file before backing it up for compaction. r=#thunderbird-reviewers

[Triage Comment]
Approved for release

Comment 8

[Daniel Darnell [:dandarnell]]

Thunderbird 132.0.1:
https://hg.mozilla.org/releases/comm-release/rev/c19955be3f91

Comment 9

[Geoff Lankow (:darktrojan)]

(In reply to Vlad Lucaci, Thunderbird QA (:vlucaci) from comment #6)

Hello,

Are there any specific STR that QA could perform in order to reproduce the faulty behaviour(crash) on an affected build and properly verify and confirm that it is fixed?

No, this is a difficult thing to test without somehow causing compaction to fail in the middle and leave the backup file behind. I'm already convinced that it works.