1951827 - AutoConfig: 'No such host' errors are no longer handled correctly - statusText vs. status [regression]

Status: RESOLVED FIXED

(Thunderbird :: Account Manager, defect)

Description

[Ben Bucksch (:BenB)]

(Re-filing bug 1907443 with less confusion)

In account setup, Exchange AutoDiscover, we try server names (like autodiscover.example.com) that may not exist. So, "host not found", firewalls blackholing us and never responding (not even a "no such host"), and similar errors are normal and common, by the very nature.

As of bug 1397646, FetchHTTP can't reliably use statusText any more as a means of detecting network failures. So if the autodiscover server doesn't provide any status text (e.g. because it uses HTTP/2), FetchHTTP now misidentifies these as network failures.

The fix is to use the status property instead.

Comment 2

[Ben Bucksch (:BenB)]

Fix: Last hunk of https://phabricator.services.mozilla.com/D216824 . We will re-attach it here as minimal patch.

Comment 3

[Ben Bucksch (:BenB)]

Attachment: Bug 1951827 - AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=neil

In account setup, Exchange AutoDiscover, we try server names that may not exist. So, "host not found", firewalls blackholing us and never responding (not even a "no such host"), and similar errors are normal and common, by the very nature.

As of bug 1397646, FetchHTTP can't reliably use statusText any more as a means of detecting network failures. So if the autodiscover server doesn't provide any status text (e.g. because it uses HTTP/2), FetchHTTP now misidentifies these as network failures.

The fix is to use the status property instead.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/89cbec0bfe46
AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=Neil,aleca

Comment 5

[Magnus Melin [:mkmelin]]

Target milestone is for "first landed in".

Comment 6

[Ben Bucksch (:BenB)]

Requesting uplift to TB 140, because this fixes a regression due to a Mozilla API change in XHR. The old code is now failing, the new code works as originally designed and tested before the regression.

Comment 7

[Magnus Melin [:mkmelin]]

Comment on attachment 9469852 [details]
Bug 1951827 - AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=neil

Uplift Approval Request

• Please state case for uplift consideration and ensure bug severity is set: Regression
• User impact if declined: Wrong error message
• Is this code covered by automated tests?: No
• Has the fix been verified in Daily?: Yes
• Has the fix been verified in Beta?: No
• Needs manual test from QA?: No
• If yes, steps to reproduce:
• List of other uplifts needed: none
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): It's a contained fix
• String changes made/needed: none

Comment 8

[Wayne Mery (:wsmwk)]

Comment on attachment 9469852 [details]
Bug 1951827 - AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=neil

[Triage Comment]
Sorry, this missed the last beta build for 140, which was cut on Monday.

Comment 9

[Magnus Melin [:mkmelin]]

We seem to be missing a approval-comm-esr140 flag

Comment 10

[Corey Bryant]

Comment on attachment 9469852 [details]
Bug 1951827 - AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=neil

[Triage Comment]
Approved for esr140

Comment 11

[Corey Bryant]

Comment on attachment 9469852 [details]
Bug 1951827 - AutoConfig: Handle 'No such host' errors correctly - statusText vs. status - fixes regression. r=neil

[Triage Comment]
Already in 141 beta

Comment 12

[Daniel Darnell [:dandarnell]]

Thunderbird 140.0.1esr:
https://hg.mozilla.org/releases/comm-esr140/rev/113d124e425858309976cff586e1e25d05ce7292