Closed Bug 1165816 Opened 9 years ago Closed 9 years ago

Cancel remote application reputation request after a certain timeout

Categories

(Toolkit :: Downloads API, defect)

38 Branch
defect
Not set
normal

Tracking

()

VERIFIED FIXED
mozilla41
Tracking Status
firefox41 --- verified

People

(Reporter: hectorz, Assigned: francois)

References

(Blocks 1 open bug)

Details

Attachments

(2 files, 1 obsolete file)

As a result of the unreliable connections to any Google service, many users in China are experiencing a long delay (in minutes) before a downloaded executable file can be marked as safe/blocked.

Beijing office would like to request introducing a timeout for application reputation check, similar to what happened in bug 1024555 for safe browsing gethash request, thanks!
I'll add a similarly-configurable timeout of 5 seconds.
Assignee: nobody → francois
Status: NEW → ASSIGNED
Here's a test case to easily reproduce this bug.
/r/9239 - Bug 1165816 - Cancel remote application reputation requests after a certain timeout

Pull down this commit:

hg pull -r fcd4d3c590da5749760c48921afc47a71531e78f https://reviewboard-hg.mozilla.org/gecko/
Component: General → Download Manager
Product: Firefox → Toolkit
Comment on attachment 8609152 [details]
MozReview Request: bz://1165816/francois

/r/9239 - Bug 1165816 - Cancel remote application reputation requests after a certain timeout

Pull down this commit:

hg pull -r fcd4d3c590da5749760c48921afc47a71531e78f https://reviewboard-hg.mozilla.org/gecko/
Attachment #8609152 - Flags: review?(gpascutto)
Comment on attachment 8609152 [details]
MozReview Request: bz://1165816/francois

/r/9239 - Bug 1165816 - Cancel remote application reputation requests after a certain timeout

Pull down this commit:

hg pull -r d0a637cdb6698deebc0fe02b64139d9d8f36371c https://reviewboard-hg.mozilla.org/gecko/
QA Contact: mwobensmith
Attachment #8609152 - Flags: review?(gpascutto) → review+
As per Google's suggestion, I have set the default timeout to 10 seconds.
Bug 1165816 - Cancel remote application reputation requests after a certain timeout. r=gcp
Attachment #8616639 - Flags: review?(gpascutto)
This new revision of the patch fixes the assertion error in Windows debug builds:

https://treeherder.mozilla.org/#/jobs?repo=try&revision=4ab17fcff34b
Comment on attachment 8616639 [details]
MozReview Request: Bug 1165816 - Cancel remote application reputation requests after a certain timeout. r=gcp

https://reviewboard.mozilla.org/r/9239/#review9181

::: toolkit/components/downloads/ApplicationReputation.cpp:755
(Diff revisions 2 - 3)
> +    mTimeoutTimer->Cancel();

Maybe null that pointer here?
Attachment #8616639 - Flags: review?(gpascutto)
Comment on attachment 8616639 [details]
MozReview Request: Bug 1165816 - Cancel remote application reputation requests after a certain timeout. r=gcp

https://reviewboard.mozilla.org/r/9239/#review9183

Ship It!
Attachment #8616639 - Flags: review+
Blocks: downloadprotection
No longer blocks: 1090754
Depends on: 1172689
No longer depends on: 1172689
See Also: → 1172689
https://hg.mozilla.org/mozilla-central/rev/5389f1e4087c
Status: ASSIGNED → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla41
Attachment #8609152 - Attachment is obsolete: true
The links to malware on http://testsafebrowsing.appspot.com are no longer blocked, sadly, and I have no idea why... nor do I know how else to verify that this has been fixed.
(In reply to Matt Wobensmith from comment #19)
> The links to malware on http://testsafebrowsing.appspot.com are no longer
> blocked, sadly, and I have no idea why... nor do I know how else to verify
> that this has been fixed.

That's a problem. I can't reproduce that though, if I let my Nightly running overnight so it can update the DB, I get the blocks.

Are you on a new profile?
Yes, this is with a new profile that has been given plenty of time to download the list.

I believe all of us working on this feature have experienced intermittent issues w/r/t blocking these links, but no one has an explanation. 

I will dig into this deeper in the next week.
Today I was able to verify that this was fixed, using the links on that page. I verified using yesterday's Nightly Fx41, against an affected Nightly from 2015-05-17. 

Again, I don't know why this intermittently fails, but for the purposes of this bug, I would consider it fixed.

I still plan on investigating the intermittent failure.
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: