image: RegisterImage can't be taught to detect Targa images #6691
Labels
Comments
I don't have data, but I feel like TARGA images are relatively rare (and the format was designed in a time when images were typically files on disk instead of streamed over the network), so while there would be a benefit to this, I don't think that the benefits outweight the complexity costs. You are of course free to implement your own Decode(io.ReadSeeker) function in your own package, or if you know that you're dealing with a .tga file, you could call targa.Decode directly. Labels changed: removed priority-someday, packagechange. Status changed to WorkingAsIntended. |
This issue was closed.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
by jeff.allen:
Targa files end with "54 52 55 45 56 49 53 49 4f 4e 2d 58 46 49 4c 45 2e 00". The detection mechanism in package images can't handle this, because it only looks at the beginning of files. A solution might be to provide a new registration function like: package image func RegisterDetector(name string, detect func(io.ReaderSeeker) bool, decode func(io.Reader) (Image, error), decodeConfig func(io.Reader) (Config, error)) RegisterFormat could be reimplemented as a call to RegisterDetector that constructs the correct lambda to do the string matching on the front of the file. The docs should encourage continued use of RegisterFormat, because it only requires an io.Reader, not the more {complex,rare,memory-hungry} io.ReaderSeeker. As a backwards compatible addition to the API, this could be done in Go 1.x, but I wouldn't be shocked if you just marked this as a wishlist item for Go 2.The text was updated successfully, but these errors were encountered: