QA/Desktop Firefox/WebRTC: Difference between revisions
< QA | Desktop Firefox
Jump to navigation
Jump to search
| Line 13: | Line 13: | ||
[[File:topoplogy.png]] | [[File:topoplogy.png]] | ||
* | The following test series should be run, with one machine on each network with Nightly. | ||
Full test series on AppRTC | |||
* A call on [http://apprtc.webrtc.org apprtc] between each pair of desktop endpoints | |||
** (9 calls in all) | |||
* A call on [http://apprtc.webrtc.org apprtc] between Firefox and Chrome (Stable + Canary) on any platform | |||
** (2 calls in all). | |||
* Each call should last no less than 5 minutes with no major failures (failure to connect, sustained freezes, etc.) | |||
Connect series on WebRTCMe | |||
* A call on [http://webrtcme.herokuapp.com/r/ webrtcme] between two Firefox endpoints (any platform) | |||
* A call on [http://webrtcme.herokuapp.com/r/ webrtcme] between Firefox and Chrome Stable (and platform) | |||
* Each call should connect and work for at least one minute | |||
aPlatforms to include all Windows, Mac OSX, Linux, Android, FirefoxOS | |||
* Daily regression testing with Firefox Nightly across all platforms to include: | * Daily regression testing with Firefox Nightly across all platforms to include: | ||
** Calling with [http://apprtc.webrtc.org apprtc] using various combinations of Firefox and Chrome | ** Calling with [http://apprtc.webrtc.org apprtc] using various combinations of Firefox and Chrome | ||
| Line 20: | Line 40: | ||
** [http://mozilla.github.com/webrtc-landing WebRTC-Landing] tests pass | ** [http://mozilla.github.com/webrtc-landing WebRTC-Landing] tests pass | ||
** ''Daily might be too ambitious, can it be done weekly?'' | ** ''Daily might be too ambitious, can it be done weekly?'' | ||
== QA Tasks == | == QA Tasks == | ||
Revision as of 21:46, 23 July 2013
See: https://wiki.mozilla.org/Media/WebRTC/Testing for background on automated testing.
Routine Manual Testing Tasks
The following tasks are essentially mechanical, but at least for now the need a person.
The basic topology should include two separate machines on separate networks, each with their own NAT. We have some flexibility about which NAT, but for concreteness, I suggest a pair of Apple Airport Express NATs. I.e., the following topology:
The following test series should be run, with one machine on each network with Nightly.
Full test series on AppRTC
- A call on apprtc between each pair of desktop endpoints
** (9 calls in all)
- A call on apprtc between Firefox and Chrome (Stable + Canary) on any platform
** (2 calls in all).
- Each call should last no less than 5 minutes with no major failures (failure to connect, sustained freezes, etc.)
Connect series on WebRTCMe
- A call on webrtcme between two Firefox endpoints (any platform)
- A call on webrtcme between Firefox and Chrome Stable (and platform)
- Each call should connect and work for at least one minute
aPlatforms to include all Windows, Mac OSX, Linux, Android, FirefoxOS
- Daily regression testing with Firefox Nightly across all platforms to include:
- Calling with apprtc using various combinations of Firefox and Chrome
- Calling with webrtcme using various combinations of Firefox and Chrome
- Evaluation of audio/video delay
- WebRTC-Landing tests pass
- Daily might be too ambitious, can it be done weekly?
QA Tasks
- Daily verification of whatever fixes landed in that day's Nightly
- Weekly triage and attempted confirmation of unconfirmed/regression bugs
- Monthly re-evaluation of regression rate to inform adjustments to the above approach
Testing
- Manual
- What must be done manually?
- Automated
- What can be automated?
