Bug 176962 – webkitpy: Provide accessor for device type

• Component: Tools / Tests
• Status: RESOLVED CONFIGURATION CHANGED

[Jonathan Bedard]

Provide a function to ask a device what type it is.  The device type will be of the form 'iPhone 5s' or 'iPad Air.'  For simulator, this device type will be derivative of the device name.  For example, 'iPhone 5s WebKit Tester 1' has a device type of 'iPhone 5s'.

[Radar WebKit Bug Importer]

<rdar://problem/34444987>

[Jonathan Bedard]

Created attachment 320836 [details]
Patch

[Daniel Bates]

Comment on attachment 320836 [details]
Patch

This patch assumes the convention that the name of the simulator will contain the name of the device. This is a convention. The name of a simulator device is arbitrary; => we should not parse the device name to determine the device type it corresponds to. Moreover, we do not have to resort to such manipulation. The command "simctl list" prints the list of device types under the "== Device Types ==" section and the parsing code in simulator.py knows how to parse this section. We need to teach the parsing code to pass the human readable name corresponding to the device type when instantiating a SimulatedDevice object.

[Jonathan Bedard]

(In reply to Daniel Bates from comment #3)
> Comment on attachment 320836 [details]
> Patch
> 
> This patch assumes the convention that the name of the simulator will
> contain the name of the device. This is a convention. The name of a
> simulator device is arbitrary; => we should not parse the device name to
> determine the device type it corresponds to. Moreover, we do not have to
> resort to such manipulation. The command "simctl list" prints the list of
> device types under the "== Device Types ==" section and the parsing code in
> simulator.py knows how to parse this section. We need to teach the parsing
> code to pass the human readable name corresponding to the device type when
> instantiating a SimulatedDevice object.

We need a map of device UDIDs to device types.  The 'device types' section only provides a mapping of device type to simulator runtime type.

I discovered this morning that this information is stored in the device.plist, which seems like the best place to retrieve this information from, I will update this patch.

[Jonathan Bedard]

Created attachment 321110 [details]
Patch

[Jonathan Bedard]

Created attachment 321111 [details]
Patch

[Jonathan Bedard]

(In reply to Jonathan Bedard from comment #6)
> Created attachment 321111 [details]
> Patch

Looking for deviceType, not runtime.

[Daniel Bates]

Would it be sufficient to be able to track the device type of devices we (webkitpy/webkitperl) created? Or do we need to be able to compute the device type for an arbitrary simulator device?

[Jonathan Bedard]

(In reply to Daniel Bates from comment #8)
> Would it be sufficient to be able to track the device type of devices we
> (webkitpy/webkitperl) created? Or do we need to be able to compute the
> device type for an arbitrary simulator device?

We support running on the currently running simulator (meaning webkitpy isn't managing the device), so we do need to be able to compute the device type for an arbitrary simulator.

I memoized the function so that we only do the computation once.

[Jonathan Bedard]

This was addressed when we shifted to the new simulator manager class <https://bugs.webkit.org/show_bug.cgi?id=180555>.