@salvo:
Congratulations on your progress.. It should be able to solve most (if not all) browser related issues.
Do note the remarks on github about the insecurity though.... you have to be careful from where the proxy can be reached (that can be from the outside world wide /mean/ internet).
Also note the remarks on cookies and password saving etc. (cleanup on exit by default but you are allowed to change it). Note that according to the documentation each individual client uses the same browserservice (and its settings). I would consider that a (security) flaw.
@Glinx:
Does this suggest that aros hosted could use the same host for browserservice? Perhaps irrelevant but curious
I fail to see why that would not be able to work. I am assuming the git site mentioning the requirement for using a separate machine have a few pro's in comparison to running it native.
- CEF is a beast, requires quite some cpu cycles. Combine it with the proxy and more or less constant image sending and you are using even more cycles.
- Using it as standalone allows you to use the service from different clients (seems to be a common use-case). Again cpu cycles and also memory hogging.
- using the service as standalone allows you to take some security measure that would be bothersome to do when located on the same host.. Separating client from service adds an additional layer of security and should be easier to maintain.
Running the service form a docker should be able to take care of most security issues (when properly setup) though.
Still no practical experience with this particular service but I do have some experience with using CEF for complete other goals. That more or less works in a similar fashion and is perfectly (cap)able to run on the same host.