1585698 - Exception: Unknown key in Tree Status response: tags in mozautomation/treestatus.py

Status: RESOLVED FIXED

(Developer Services :: Mercurial: mozext, defect, P3)

Description

[Steve Fink [:sfink] [:s:]]

Recently, the treestatus response format changed, and mozautomation's treestatus.py throws an Exception on any unrecognized key.

Comment 1

[Steve Fink [:sfink] [:s:]]

Attachment: Handle new keys in tree status response

Note that I won't be able to make sure this gets in (and don't have commit privs anyway), as I'll be on PTO for the rest of the week.

Comment 2

[Steve Fink [:sfink] [:s:]]

Attachment: Handle new keys in tree status response,

(fixed commit message)

Comment 3

[Connor Sheehan [:sheehan]]

Steve, can you send this on Phabricator, and tag me for review? Thanks.

Comment 4

[Rok Garbas [:garbas]]

Comment on attachment 9098268 [details] [diff] [review]
Handle new keys in tree status response,

>+            elif k == 'tags':
>+                self.tags = d[k]
>+            elif k in ('log_id',):
>+                pass
>             else:
>                 raise Exception('Unknown key in Tree Status response: %s' % k)

The api change was a minor version bump, since api got only extended (new fields) and existing fields change.

I'm not aware of a reason why checking full api response is actually needed. I would ignore fields that you are not using in the script. That would make it more future proof if there will be another field that will be added, for whatever reason.

Comment 5

[Steve Fink [:sfink] [:s:]]

Attachment: Bug 1585698 - Handle new keys in tree status response, r=sheehan

Comment 6

[Steve Fink [:sfink] [:s:]]

(In reply to Rok Garbas [:garbas] from comment #4)

I'm not aware of a reason why checking full api response is actually needed.
I would ignore fields that you are not using in the script. That would make
it more future proof if there will be another field that will be added, for
whatever reason.

Sorry, I shouldn't have tagged you for review (but I was leaving and in a rush). Connor is a better one to make decisions about what makes sense for this code. I also don't have a reason to go one way or the other, I was just mostly following existing style. I've redirected the review to Connor.

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/hgcustom/version-control-tools/rev/fdb9aada493e
Handle new keys in tree status response, r=sheehan