first time running this script I will get the 'ref' image created, and it will be updated correctly. Now if I leave tev open and run the script again, I get a second 'ref' image in the image list, which remains black and the first 'ref' image I created gets updated.
There's two options for the behaviour here:
1) When create_image is called a second time with the duplicate name, the duplicate call is ignored. This would lead to the current behaviour, except that we wouldn't have empty images created every time we run the script
2) create_image could modify the requested name to be unique, and return the uniquified name over the socket. The calling script would then store the returned name and use it in subsequent update_image calls to fill in the new image, leaving the first-created one alone.
Option 2 is most similar to behaviour in common render catalogs (e.g. it, katana's monitor etc), but given that there isn't currently a mechanism for returning data to the client over the socket, it's probably simpler just to do 1) for now. I'm happy to do this and submit a PR if you're open to it.
Say I have a small python script that does something like:
first time running this script I will get the 'ref' image created, and it will be updated correctly. Now if I leave tev open and run the script again, I get a second 'ref' image in the image list, which remains black and the first 'ref' image I created gets updated.
There's two options for the behaviour here: 1) When
create_imageis called a second time with the duplicate name, the duplicate call is ignored. This would lead to the current behaviour, except that we wouldn't have empty images created every time we run the script 2)create_imagecould modify the requested name to be unique, and return the uniquified name over the socket. The calling script would then store the returned name and use it in subsequentupdate_imagecalls to fill in the new image, leaving the first-created one alone.Option 2 is most similar to behaviour in common render catalogs (e.g. it, katana's monitor etc), but given that there isn't currently a mechanism for returning data to the client over the socket, it's probably simpler just to do 1) for now. I'm happy to do this and submit a PR if you're open to it.