chore: remove private webContents.getId() API #13674
Conversation
1c6226b to
5d22f68
Compare
| @@ -210,9 +210,9 @@ const unwrapArgs = function (sender, contextId, args) { | |||
| return rendererFunctions.get(objectId) | |||
| } | |||
|
|
|||
| const webContentsId = sender.getId() | |||
There was a problem hiding this comment.
Should we use sender.id here given some people might be using affinity 🤔 Idk if remote even works properly with affinity but this stands out as being an incompat point
There was a problem hiding this comment.
The purpose of check here is to ensure that message is not sent to the old page, the WebContents is guaranteed to be the same one so there is no need to compare the sender.id, and messages sent to one WebContents won't be sent to other WebContents with same affinity.
The ID returned by sender.getId() is changed when renderer process is restarted, comparing the processId has the same behavior with the old code. And it is the reason we used sender.getId() instead of sender.id.
The check in main process however can not tell when a WebContents have multiple pages in it, and it is silently checked in the renderer process:
electron/lib/renderer/api/remote.js
Lines 279 to 285 in e90c4ab
|
Our app actually does use getId: https://github.com/minbrowser/min/search?q=getid. It doesn't look like it's too hard to migrate, but it might be nice to mention this in the release notes in case other apps are using this as well. |
5d22f68 to
befadd2
Compare
With #13603, there is no need to have the private
webContents.getId()API, which is very easy to be confused with the publicwebContents.idproperty.The
webContents.getId()API has never been documented, after doing a search in GitHub I did not found any app code that used the API, the only users were Electron itself and its forks. So this change will not break apps.