feat: Always return capture size #1202
Conversation
| if (!returnImage) { | ||
| mobileScannerCallback(barcodeMap, null, null, null) | ||
| mobileScannerCallback( | ||
| barcodeMap, |
There was a problem hiding this comment.
When not returning the created bitmap, use the available dimensions from the input image, which are the same.
|
|
||
| /// Whether scanned barcodes should contain the image | ||
| /// that is embedded into the barcode. | ||
| /// Whether the [BarcodeCapture.image] bytes should be provided. |
There was a problem hiding this comment.
This wording was incorrect, so I adjusted it. There is no support for returning images that are embedded in the center of some specific QR codes. (i.e. little square logo's)
| /// The size of the input [image]. | ||
| /// | ||
| /// If [image] is null, this will be [Size.zero]. | ||
| /// The size of the camera input [image]. |
There was a problem hiding this comment.
Adjusted the wording to match the new behavior.
| controller.add( | ||
| BarcodeCapture( | ||
| barcodes: [result.toBarcode], | ||
| size: videoSize, |
There was a problem hiding this comment.
We already had this information in the getter implementation, so I just forwarded it. This returns Size.zero if the camera output is not available.
| // The image dimensions are always provided. | ||
| // The image bytes are only non-null when `returnImage` is true. | ||
| "image" to mapOf( | ||
| "bytes" to image, |
There was a problem hiding this comment.
The width / height were nullable, which could have been changed. However, the Dart code already handles null, so I just left it as-is, to reduce the diff here.
| final JSArray<JSString>? facingModes = capabilities.facingModeNullable; | ||
|
|
||
| // TODO: this is an empty array on MacOS Chrome, where there is no facing mode, but one, user facing camera. | ||
| // We might be able to add a workaround, using the label of the video track. |
There was a problem hiding this comment.
I'm going to revisit the broken check for the transform on the web in the future. Just left a breadcrumb for myself here.
| } | ||
| // The image dimensions are always provided. | ||
| // The image bytes are only non-null when `returnImage` is true. | ||
| let imageData: [String: Any?] = [ |
There was a problem hiding this comment.
Making the bytes conditional started giving a warning with nullable Any, so I moved it to a separate dict here to fix that.
|
|
||
| // The image dimensions are always provided. | ||
| // The image bytes are only non-null when `returnImage` is true. | ||
| let imageData: [String: Any?] = [ |
There was a problem hiding this comment.
Like on iOS, I moved the dict to a separate var to fix a warning.
|
|
||
| DispatchQueue.main.async { | ||
| if (!MobileScannerPlugin.returnImage) { | ||
| guard let image = cgImage else { |
There was a problem hiding this comment.
Using a guard-let to avoid ! in the cgImage below, makes things look better
navaronbracke
changed the title
navaronbracke
changed the title
This PR updates the behavior of
returnImageto only refer to the bytes of the output image, not the size.The size can always be returned, to allow users to do calculations with it (for example, together with the
Barcode.size)Part of #1183