API Review: Frame process info #3675

JosephJin0815 · 2023-07-31T17:20:52Z

This is a review for the new FrameProcessInfo API.

Add Spec for FrameProcessInfo.md

oldnewthing · 2023-08-10T19:51:00Z

specs/FrameProcesssInfo.md

+    if (environment14)
+    {
+        //! [GetProcessInfosWithDetails]
+        CHECK_FAILURE(environment14->GetProcessInfosWithDetails(


Why do we need a new method at all? Is the problem that getting the AssociatedFrameInfos is expensive, so we don't want to do the work unless explicitly requested?

(Also, we're taking the name "WithDetails", which means that if we extend ProcessInfo more, we're stuck with a clunky name like "WithMoreDetails" or "NowEvenDetailier".)

Yeah, it could be expensive. Considered we want to use the new method to acquire more process information detail.
Current design is to populate data in CoreWebView2Environment.GetProcessInfosWithDetails and extend CoreWebView2ProcessInfo.

This was my first question too - why is GetProcessInfosWithDetails even necessary, given you need to QI for ICoreWebView2ProcessInfo2? Either the QI or the call to get_AssociatedFrameInfos provide ample opportunity for lazy evaluation of the AssociatedFrameInfos.

oldnewthing · 2023-08-10T19:52:08Z

specs/FrameProcesssInfo.md

+                            CHECK_FAILURE(
+                                processInfo->QueryInterface(IID_PPV_ARGS(&processInfo2)));
+                            wil::com_ptr<ICoreWebView2FrameInfoCollection> frameInfoCollection;
+                            CHECK_FAILURE(processInfo2->get_AssociatedFrameInfos(


Does this return the associated frame infos as of the point of call? Or does it return the frame infos for the frames that were associated at the time GetProcessInfosWithDetails was called?

In other words, is this a snapshot or a live object?

It's when GetProcessInfosWithDetails gets called. It's a snapshot.

New names discussed: GetProcessExtendedInfosAsync returns ProcessExtendedInfo.

oldnewthing · 2023-08-10T19:56:26Z

specs/FrameProcesssInfo.md

+{
+    wil::com_ptr<ICoreWebView2FrameInfo> mainFrameInfo;
+    wil::com_ptr<ICoreWebView2FrameInfo2> frameInfo2;
+    while (frameInfo)


Just checking: A frame that is an immediate child of the top-level document reports no parent frame. In other words, the top-level document does not count as a "frame" for this purpose.

We inherit internal definition and treat top-level document as the MainFrame. For an immediate child of the top-level document, we reports the corresponding MainFrame as its parent frame.

oldnewthing · 2023-08-10T19:57:39Z

specs/FrameProcesssInfo.md

+    COREWEBVIEW2_FRAME_KIND frameKind = COREWEBVIEW2_FRAME_KIND_OTHER;
+
+    CHECK_FAILURE(frameInfo->get_Name(&nameRaw));
+    std::wstring name = ((std::wstring)(nameRaw.get())).empty() ? L"none" : nameRaw.get();


Suggested change

std::wstring name = ((std::wstring)(nameRaw.get())).empty() ? L"none" : nameRaw.get();

std::wstring name = nameRaw.get()[0] ? nameRaw.get() : L"none";

No need to create a temporary std::wstring for the sole purpose of checking if the string is empty.

oldnewthing · 2023-08-10T19:59:55Z

specs/FrameProcesssInfo.md

+std::wstring ProcessComponent::FrameKindToString(const COREWEBVIEW2_FRAME_KIND kind)
+{
+    switch (kind)
+    {                                                                


I think a macro definition is missing

#define KIND_ENTRY(kindValue) case kindValue: return L#kindValue;

oldnewthing · 2023-08-10T20:11:12Z

specs/FrameProcesssInfo.md

+    CHECK_FAILURE(frameInfo2->get_FrameId(&mainFrameId));
+
+    wil::com_ptr<ICoreWebView2FrameInfo> firstLevelFrameInfo =
+        GetAncestorFirstLevelFrameInfo(frameInfo);


C++ sample forgot to include the code for this function. (But I can infer it from the C# sample.)

Terminology is a little confusing. "First level frame" sounds like the main frame (since the main frame comes first), but I think here "first level frame" is "first level below the main frame." So it's the first subframe.

Agreed. Please update. Main frame's child or don't treat this specially and don't give it a term.

oldnewthing · 2023-08-10T20:37:27Z

specs/FrameProcesssInfo.md

+        CHECK_FAILURE(frameInfo2->get_FrameId(&firstLevelFrameId));
+    }
+
+    result << L"{frame name:" << name << L" | frame Id:" << std::to_wstring(frameId)


Suggested change

result << L"{frame name:" << name << L" | frame Id:" << std::to_wstring(frameId)

result << L"{frame name:" << name << L" | frame Id:" << frameId

Don't need to std::to_wstring the uint32. << can accept uint32's already.

oldnewthing · 2023-08-10T20:46:05Z

specs/FrameProcesssInfo.md

+                            CHECK_FAILURE(
+                                processInfo->QueryInterface(IID_PPV_ARGS(&processInfo2)));
+                            wil::com_ptr<ICoreWebView2FrameInfoCollection> frameInfoCollection;
+                            CHECK_FAILURE(processInfo2->get_AssociatedFrameInfos(


What happens if you call the old GetProcessInfos (no WithDetails) and then ask for the AssociatedFrameInfos? Do you get an empty list? Do you get nullptr? Do you get an exception?

It will return an empty list,

CoreWebView2ProcessInfo objects obtained via CoreWebView2Environment. GetProcessInfos or for non-renderer processes will always have an empty AssociatedFrameInfos.

This is a little confusing as there would be no way to guess that just from reading the api.

Would it instead be possible to do something like:

runtimeclass CoreWebView2 { Windows.Foundation.IAsyncOperation<IVectorView<WebView2ProcessAndFrameInfo>> GetProcessAndFrameInfosAsync() } runtimeclass WebView2ProcessAndFrameInfo { CoreWebView2ProcessInfo ProcessInfo {get; } IVectorView<CoreWebView2FrameInfo> FrameInfos { get; }; }

This way you would never end up with an object that included 'dummy' properties.

Or CoreWebView2ProcessInfo.GetAssociatedFrameInfos()?

This is another reason not to introduce GetProcessInfosWithDetails - it fragments the behavior of GetProcessInfos. Just add ICoreWebView2ProcessInfo2 with lazy evaluation of get_AssociatedFrameInfos.

oldnewthing · 2023-08-10T20:47:28Z

specs/FrameProcesssInfo.md

+                            CHECK_FAILURE(
+                                processInfo->QueryInterface(IID_PPV_ARGS(&processInfo2)));
+                            wil::com_ptr<ICoreWebView2FrameInfoCollection> frameInfoCollection;
+                            CHECK_FAILURE(processInfo2->get_AssociatedFrameInfos(


What about top-level documents? This lets me find the process that is hosting a specific frame, but how do I find the process that is hosting my top-level document?

The documentation for ParentFrameInfo suggests that you can get a FrameInfo for the top-level document, since it talks about the value of the ParentFrameInfo property in that case:

This property is also null for the top most document in the WebView2 which has no parent frame.
But how do you get the FrameInfo for the top-level document?

The top-level documents also returning as part of GetProcessInfosWithDetails. CoreWebView2FrameKind.MainFrame is the type for top-level documents.

oldnewthing · 2023-08-10T21:10:31Z

specs/FrameProcesssInfo.md

+[v1_enum]
+typedef enum COREWEBVIEW2_FRAME_KIND {
+  /// Indicates that the frame is type of frame we don't differentiate yet.
+  COREWEBVIEW2_FRAME_KIND_OTHER,


Do we ever report this for any other kind of frames beyond terminated frames? Could just call this TERMINATED and just use new values for future types of frames?

Calling it OTHER makes it hard to add new types of frames later. Suppose we add "3DFrames" later. Do they show up as COREWEBVIEW2_FRAME_KIND_3DFRAME and confuse old clients? Or do we report them as COREWEBVIEW2_FRAME_KIND_OTHER and confuse new clients?

FrankC-msft · 2023-08-10T20:26:58Z

specs/FrameProcesssInfo.md

+
+We provide the `GetProcessInfos` API for host applications to understand which
+processes are part of their WebView2s. That API provides enough information for
+a host application to understand the overall performance impact(memory, CPU 


Suggested change

a host application to understand the overall performance impact(memory, CPU

a host application to understand the overall performance impact (memory, CPU

nit, missing space. This occurs all three times parentheticals are used in Background + Description.

FrankC-msft · 2023-08-10T20:47:27Z

specs/FrameProcesssInfo.md

+[v1_enum]
+typedef enum COREWEBVIEW2_FRAME_KIND {
+  /// Indicates that the frame is type of frame we don't differentiate yet.
+  COREWEBVIEW2_FRAME_KIND_OTHER,


Do we need a reserved Other enum for future expansion here? (vs adding new concrete frame kinds when they becomes defined)
Are there real other types we know about but aren't including?
I'm wondering if including Other increases risk of non-extensible usage. Will developers' apps break if some Frames "change kind" from Other to a new/future enum value? When or why would a developer check for "Other" type today?

Probably yes - With the speed and experimentation-centric approach of web standards, it's reasonable to expect new kinds of frames that we'll add well after they first appear. For example, Fenced Frames have been a proposal and prototype/experiment for years but are just getting to being a "standard". It's possible for something like this to change names, get dropped entirely, etc. that we wouldn't want to set in stone for our enum until it's actually complete.

Realistically, most use-cases will probably treat "Other" just like "IFrame" and be fine.

Frame type "Other" helps to provide a more complete/accurate frame tree structure while frame kinds updating.
There're other defined frame types like embed frame, object frame that developers might want to differentiate.

Have we written down some kind of a philosophy/strategy on API compat for WebView?

Please change Other -> Unknown, and update documentation to explain compat situation.

FrankC-msft · 2023-08-10T20:57:06Z

specs/FrameProcesssInfo.md

+the `FrameId` property. This property represents the unique identifier of
+the frame running in this `CoreWebView2` or `CoreWebView2Frame`.
+
+* We propose extending `CoreWebView2FrameInfo` to include `FrameId`, 


Since we're extending CoreWebView2ProcessInfo itself, results from existing GetProcessInfos() calls will start having an empty AssociatedFrameInfos property. GetProcessInfosWithDetails results may have empty results. Is it up the app developer to keep track of when the property is real vs inert?

Potential Option
GetProcessInfo: Light weight data
GetProcessInfoExtended(Async): Heavy data like AssociatedFrameInfos.

FrankC-msft · 2023-08-10T21:01:36Z

specs/FrameProcesssInfo.md

+    wil::com_ptr<ICoreWebView2FrameInfo> frameInfo);
+std::wstring FrameKindToString(COREWEBVIEW2_FRAME_KIND kind);
+
+// Display renderer process info with details which includes the list of 


right now "Details" specifically means AssociatedFrameInfos. If we add more new optional details, will they also be populated by calls to GetProcessInfosWithDetails, or would there be a new method to opt into even more specific information?
(I infer they'd be populated in calls to GetProcessInfosWithDetails, otherwise we'd name this method GetProcessInfosWithFramesInfo to avoid boxing out future method names.)

As a tangible example for when this comment is discussed - We can theoretically have renderer processes for service workers that don't have an active frame.

Considered we may provide more information for the processInfo. We'd like to extend CoreWebView2ProcessInfo to have more Associated...Info instead of adding more CoreWebView2Environment.GetProcessInfosWith...Info .

Returning the sparsely populated Infos is confusing. Could the associated info be a getter method? As in: CoreWebView2ProcessInfo.GetAssociatedFrameInfos()?

FrankC-msft · 2023-08-10T21:38:07Z

specs/FrameProcesssInfo.md

+  /// Note that `FrameId` may not be valid if `CoreWebView` has not done
+  /// any navigation. It's safe to get this value during or after the first
+  /// `ContentLoading` event. Otherwise, it could return the invalid frame Id 0.
+  [propget] HRESULT FrameId([out, retval] UINT32* id);


Will developers want the Frame instead of FrameId value? How often will developers want to use this id to lookup a CoreWebView2Frame, or is the common case to log this for diagnostics?

For memory monitoring cases, we expect developers not only want to acquire resource consumption for the frame tree, but also have chance to take further operations for the corresponding CoreWebView2/CoreWebView2Frame. Not sure how often it is, but it's a requirement for our existing customers.

No change: We will want to do this in the future but its complicated and not required for this spec now.

FrankC-msft · 2023-08-10T21:41:52Z

specs/FrameProcesssInfo.md

+
+# Description
+* We propose extending `CoreWebView2Environment` to include the 
+`GetProcessInfosWithDetails` API. This asynchronous call returns a 


Confirming, getting these frame details needs to be async? Is it expensive on the level of network or large-file access because of the multi-process model behind WV2?

It can be expensive on the level of network when navigating across pages with child frames.

In those cases, there could be a lot frame created/deleted message.

Discussion: yes, it's async because it can be slow

kmahone · 2023-08-14T22:47:04Z

specs/FrameProcesssInfo.md

+
+# Description
+* We propose extending `CoreWebView2Environment` to include the 
+`GetProcessInfosWithDetails` API. This asynchronous call returns a 


Inconsistent naming GetProcessInfosWithDetails here and GetProcessInfosWithDetailsAsync further down.

Discussion: the IDL projection has the "Async" suffix, the COM doesn't, which is pattern

vickiez · 2023-08-15T18:29:50Z

specs/FrameProcesssInfo.md

+  /// `CoreWebView2FrameInfo` objects obtained via `CoreWebView2.ProcessFailed` will
+  /// always have a `null` `ParentFrameInfo`. This property is also `null` for the  
+  /// top most document in the WebView2 which has no parent frame.
+  /// Note that this `ParentFrameInfo` could be out of date as it's a snapshot.   


How are developers expected to check if this frame info is out of date? Track whether the ICoreWebView2Frame with same ID was destroyed?

Developers can check if the frame info is up-to-date by ID, but it's only limited on CoreWebView2 and first sub layer Iframe CoreWebViewFrame. Developers will make their own strategies to call this API. For example, I've heard a customer will check it periodically.

Please document: does the snapshot correspond to when the call is made or when the call completes? And note you can use frame created/destroyed to know if the snapshot will correspond to the live data on CoreWebView2(Frame)

MikeHillberg · 2023-08-15T18:24:06Z

specs/FrameProcesssInfo.md

+`GetProcessInfosWithDetails` API. This asynchronous call returns a 
+snapshot collection of `ProcessInfo`s corresponding to all currently
+running processes associated with this `ICoreWebView2Environment` 
+except for crashpad process. This provide the same list of `ProcessInfo`s


Is "except for crashpad process" true for current GetProcessInfos? Not mentioned in the doc page

Consider: update public doc

@JosephJin0815 Can you open a DevEco bug to update the GetProcessInfos docs to include this note?

MikeHillberg · 2023-08-15T18:26:35Z

specs/FrameProcesssInfo.md

+        return L#kindValue;
+
+        KIND_ENTRY(COREWEBVIEW2_FRAME_KIND_MAIN_FRAME);
+        KIND_ENTRY(COREWEBVIEW2_FRAME_KIND_IFRAME);


Naming: I thought we'd decided at some point to consistently use "frame" rather than "iframe" because no one remembers framesets anyway.

No change. This is intentional. We have different kinds of frames one of which is an iframe.

Discussion: This is "iframe" to distinguish it from other kinds of frames like main frame (which isn't an iframe) and future such as fenced frame

MikeHillberg · 2023-08-15T18:36:10Z

specs/FrameProcesssInfo.md

+    wil::com_ptr<ICoreWebView2FrameInfo> frameInfo);
+std::wstring FrameKindToString(COREWEBVIEW2_FRAME_KIND kind);
+
+// Display renderer process info with details which includes the list of 


Returning the sparsely populated Infos is confusing. Could the associated info be a getter method? As in: CoreWebView2ProcessInfo.GetAssociatedFrameInfos()?

MikeHillberg · 2023-08-15T18:37:56Z

specs/FrameProcesssInfo.md

+                            CHECK_FAILURE(
+                                processInfo->QueryInterface(IID_PPV_ARGS(&processInfo2)));
+                            wil::com_ptr<ICoreWebView2FrameInfoCollection> frameInfoCollection;
+                            CHECK_FAILURE(processInfo2->get_AssociatedFrameInfos(


Or CoreWebView2ProcessInfo.GetAssociatedFrameInfos()?

MikeHillberg · 2023-08-15T18:41:25Z

specs/FrameProcesssInfo.md

+[v1_enum]
+typedef enum COREWEBVIEW2_FRAME_KIND {
+  /// Indicates that the frame is type of frame we don't differentiate yet.
+  COREWEBVIEW2_FRAME_KIND_OTHER,


Have we written down some kind of a philosophy/strategy on API compat for WebView?

MikeHillberg · 2023-08-15T18:44:47Z

specs/FrameProcesssInfo.md

+/// A continuation of the ICoreWebView2FrameInfo interface.
+[uuid(a7a7e150-e2ca-11ed-b5ea-0242ac120002), object, pointer_default(unique)]
+interface ICoreWebView2FrameInfo2 : ICoreWebView2FrameInfo {
+  /// This frame's parent frame's `FrameInfo`. `ParentFrameInfo` will only be


Suggested change

/// This frame's parent frame's `FrameInfo`. `ParentFrameInfo` will only be

/// This parent frame's `FrameInfo`. `ParentFrameInfo` will only be

MikeHillberg · 2023-08-15T18:46:25Z

specs/FrameProcesssInfo.md

+  [propget] HRESULT ParentFrameInfo([out, retval] ICoreWebView2FrameInfo** frameInfo);
+  /// The unique identifier of the frame associated with the current `FrameInfo`.
+  /// It's the same kind of ID as with the `FrameId` in `CoreWebView2` and via 
+  /// `CoreWebView2Frame`. `FrameId` will only be populated when obtained


Suggested change

/// `CoreWebView2Frame`. `FrameId` will only be populated when obtained

/// `CoreWebView2Frame`. `FrameId` will only be populated (non-zero) when obtained

Please fix.

MikeHillberg · 2023-08-15T18:48:20Z

specs/FrameProcesssInfo.md

+/// A continuation of the ICoreWebView2_17 interface.
+[uuid(ad712504-a66d-11ed-afa1-0242ac120002), object, pointer_default(unique)]
+interface ICoreWebView2_18 : ICoreWebView2_17 {
+  /// The unique identifier of the current frame. It's the same kind of ID as with 


What does "current frame" mean?

Please fix: say "main frame" instead.

Scottj1s · 2023-08-15T20:06:14Z

specs/FrameProcesssInfo.md

+also can be used to build the frame tree represented by `FrameInfo`s.
+
+# Examples
+C++


Any reason the sample isn't C++/WinRT?

david-risney · 2023-08-15T20:50:22Z

specs/FrameProcesssInfo.md

+        CHECK_FAILURE(frameInfo2->get_FrameId(&firstLevelFrameId));
+    }
+
+    result << L"{frame name:" << name << L" | frame Id:" << std::to_wstring(frameId)


david-risney · 2023-08-15T20:50:26Z

specs/FrameProcesssInfo.md

+std::wstring ProcessComponent::FrameKindToString(const COREWEBVIEW2_FRAME_KIND kind)
+{
+    switch (kind)
+    {                                                                


david-risney · 2023-08-15T20:52:36Z

specs/FrameProcesssInfo.md

+        return L#kindValue;
+
+        KIND_ENTRY(COREWEBVIEW2_FRAME_KIND_MAIN_FRAME);
+        KIND_ENTRY(COREWEBVIEW2_FRAME_KIND_IFRAME);


No change. This is intentional. We have different kinds of frames one of which is an iframe.

david-risney · 2023-08-15T21:03:51Z

specs/FrameProcesssInfo.md

+[v1_enum]
+typedef enum COREWEBVIEW2_FRAME_KIND {
+  /// Indicates that the frame is type of frame we don't differentiate yet.
+  COREWEBVIEW2_FRAME_KIND_OTHER,


Please change Other -> Unknown, and update documentation to explain compat situation.

david-risney · 2023-08-15T21:04:06Z

specs/FrameProcesssInfo.md

+[v1_enum]
+typedef enum COREWEBVIEW2_FRAME_KIND {
+  /// Indicates that the frame is type of frame we don't differentiate yet.
+  COREWEBVIEW2_FRAME_KIND_OTHER,


david-risney · 2023-08-15T21:05:33Z

specs/FrameProcesssInfo.md

+/// A continuation of the ICoreWebView2FrameInfo interface.
+[uuid(a7a7e150-e2ca-11ed-b5ea-0242ac120002), object, pointer_default(unique)]
+interface ICoreWebView2FrameInfo2 : ICoreWebView2FrameInfo {
+  /// This frame's parent frame's `FrameInfo`. `ParentFrameInfo` will only be


david-risney · 2023-08-15T21:08:20Z

specs/FrameProcesssInfo.md

+  /// `CoreWebView2FrameInfo` objects obtained via `CoreWebView2.ProcessFailed` will
+  /// always have a `null` `ParentFrameInfo`. This property is also `null` for the  
+  /// top most document in the WebView2 which has no parent frame.
+  /// Note that this `ParentFrameInfo` could be out of date as it's a snapshot.   


Please document: does the snapshot correspond to when the call is made or when the call completes? And note you can use frame created/destroyed to know if the snapshot will correspond to the live data on CoreWebView2(Frame)

david-risney · 2023-08-15T21:08:40Z

specs/FrameProcesssInfo.md

+  [propget] HRESULT ParentFrameInfo([out, retval] ICoreWebView2FrameInfo** frameInfo);
+  /// The unique identifier of the frame associated with the current `FrameInfo`.
+  /// It's the same kind of ID as with the `FrameId` in `CoreWebView2` and via 
+  /// `CoreWebView2Frame`. `FrameId` will only be populated when obtained


Please fix.

david-risney · 2023-08-15T21:09:54Z

specs/FrameProcesssInfo.md

+/// A continuation of the ICoreWebView2_17 interface.
+[uuid(ad712504-a66d-11ed-afa1-0242ac120002), object, pointer_default(unique)]
+interface ICoreWebView2_18 : ICoreWebView2_17 {
+  /// The unique identifier of the current frame. It's the same kind of ID as with 


Please fix: say "main frame" instead.

david-risney · 2023-08-15T21:12:28Z

specs/FrameProcesssInfo.md

+  /// Note that `FrameId` may not be valid if `CoreWebView` has not done
+  /// any navigation. It's safe to get this value during or after the first
+  /// `ContentLoading` event. Otherwise, it could return the invalid frame Id 0.
+  [propget] HRESULT FrameId([out, retval] UINT32* id);


No change: We will want to do this in the future but its complicated and not required for this spec now.

Wangsong Jin and others added 18 commits May 9, 2023 18:09

Add Frame Process Info API

d4b72f0

Update FrameProcessInfo.md

bf09264

Update FrameProcessInfo.md

72a19b7

Update FrameProcessInfo.md

b4edd29

Update FrameProcessInfo.md

81c2b03

Update FrameProcessInfo.md

81f887f

Update FrameProcessInfo.md

4816275

Update FrameProcessInfo.md

c435b13

Update FrameProcessInfo.md

0983948

Update FrameProcessInfo.md

970d45d

Update FrameProcessInfo.md

f65b6ee

Update FrameProcessInfo.md

29d0d74

Update FrameProcessInfo.md

9cb6da2

Update FrameProcessInfo.md

11fa6fa

Update FrameProcessInfo.md

0d7b32a

Update FrameProcessInfo.md

79ad4f3

Update FrameProcessInfo.md

854c25a

Merge pull request #3473 from MicrosoftEdge/FrameProcessInfo-draft

e38fd80

Add Spec for FrameProcessInfo.md

JosephJin0815 added the API Proposal Review WebView2 API Proposal for review. label Jul 31, 2023

JosephJin0815 requested a review from david-risney July 31, 2023 17:20

oldnewthing reviewed Aug 10, 2023

View reviewed changes

FrankC-msft reviewed Aug 10, 2023

View reviewed changes

kmahone reviewed Aug 14, 2023

View reviewed changes

vickiez reviewed Aug 15, 2023

View reviewed changes

MikeHillberg reviewed Aug 15, 2023

View reviewed changes

Scottj1s reviewed Aug 15, 2023

View reviewed changes

david-risney requested changes Aug 15, 2023

View reviewed changes

Update FrameProcessInfo.md

a228c57

david-risney approved these changes Sep 15, 2023

View reviewed changes

JosephJin0815 added the review completed WebView2 API Proposal that's been reviewed and now needs final update and push label Sep 15, 2023

JosephJin0815 mentioned this pull request Sep 15, 2023

API Review: Frame Process Info MicrosoftEdge/WebView2Announcements#75

Open

	std::wstring name = ((std::wstring)(nameRaw.get())).empty() ? L"none" : nameRaw.get();
	std::wstring name = nameRaw.get()[0] ? nameRaw.get() : L"none";

	result << L"{frame name:" << name << L" \| frame Id:" << std::to_wstring(frameId)
	result << L"{frame name:" << name << L" \| frame Id:" << frameId

	a host application to understand the overall performance impact(memory, CPU
	a host application to understand the overall performance impact (memory, CPU

	/// This frame's parent frame's `FrameInfo`. `ParentFrameInfo` will only be
	/// This parent frame's `FrameInfo`. `ParentFrameInfo` will only be

	/// `CoreWebView2Frame`. `FrameId` will only be populated when obtained
	/// `CoreWebView2Frame`. `FrameId` will only be populated (non-zero) when obtained

API Review: Frame process info #3675

Are you sure you want to change the base?

API Review: Frame process info #3675

Conversation

JosephJin0815 commented Jul 31, 2023

Choose a reason for hiding this comment

JosephJin0815 Aug 11, 2023 • edited

Choose a reason for hiding this comment

Scottj1s Aug 15, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JosephJin0815 Aug 11, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kmahone Aug 14, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JosephJin0815 Aug 11, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JosephJin0815 Aug 11, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

JosephJin0815 Aug 11, 2023 •

edited

Scottj1s Aug 15, 2023 •

edited

JosephJin0815 Aug 11, 2023 •

edited

kmahone Aug 14, 2023 •

edited

JosephJin0815 Aug 11, 2023 •

edited

JosephJin0815 Aug 11, 2023 •

edited