pass in a raw pixelBuffer to the open command #86
Conversation
|
motivation for using open: jpg and png buffers are already accepted as input. I think it might be confusing to use a separate function to create images based on raw buffers. Possible code optimization might be creating a decoder in c which doesn't transform the raw input buffers, but only does the validation of the buffer size & parameters. Right now this happens in javascript. However, I don't think there's much of a speed benefit of doing that in c, it would rather be for code structuring reasons. |
|
I got rid of the transparency option, as this can be auto-detected based on the given width, height and buffer size. |
| callback(err, err ? null : new Image(pixelsBuf, width, height, trans)); | ||
| }); | ||
| if(typeof type === 'object') { | ||
| if(!type.width && !type.height) { |
There was a problem hiding this comment.
This condition can be dropped. The next two ones are sufficient.
| if(!type.height) { | ||
| throw Error("Missing height"); | ||
| } | ||
| if(typeof type.width !== "number") { |
There was a problem hiding this comment.
width and height need to be positive integers.
Can replace both of the following conditions
if(!type.width) {
throw Error("Missing width");
}
if(typeof type.height !== "number") {
throw Error("Height must be numeric");
}With:
if (!(type.width === parseInt(type.width) && type.width > 0)){
throw Error("Width must be a positive integer");
}This will cover both cases where type.width is undefined, and not of the correct type.
|
I'm wondering if the section for validating the Then do something like: // ...
{
name: 'type',
types: ['string', 'rawBufferProperties'],
optional: true
}
// ... |
…o custom decree type
|
improved validation based on your feedback & created a custom decree type to validate the width and height (required positive integer). |
pass in a raw pixelBuffer to the open command
|
Great! I've merged this, and am going to push a few more changes, in case you want to take a look. Thanks a lot for this PR! |
|
cool :-) |
Referencing my previous pull request #84 - but now on the 0.0.6 branch