Permissive Sandbox

[OvermindDL1]

There was talk on IRC yesterday about wanting a way to allow the sandbox to not kill the game when a permission is denied for testing purposes, as well as documenting what is needed to supply the new access requests to the main Terasology set. I went ahead and looked in to it, however I found that the functionality is in gestalt, which as it being its own module that is depended on by Terasology and it missing the functionality I needed to add it to it first. Could be PR be looked over and make sure I did not miss anything? I created a new test file (copy of an existing one with appropriate changes made) and fixed all issues I found with it and it passes properly now with correct logging. If this works then could this be merged to main so the primary Terasology project can use it? I have a new branch ready for Terasology to take advantage of this new functionality via commandline switch (-permissiveSandbox, shouldn't those be two dashes for --permissiveSandbox since it is a whole word, but the rest use only one, so I followed the unusual style) and ready to PR it once the Gestalt PR is merged and I can run final tests to make sure it works in Terasology itself.

Gestalt Permissive Support PR: https://github.com/MovingBlocks/gestalt/pull/13

 

[Immortius]

(-permissiveSandbox, shouldn't those be two dashes for --permissiveSandbox since it is a whole word, but the rest use only one, so I followed the unusual style)

From what I can tell there are a plethora of standards for commandline arguments style. I assume you are referring to the gnu style, which uses -- for long argument names and - for short versions. We don't have short versions though nor support for combining short versions, so doesn't really matter (the standard is more for commands rather than applications really). Not that I would care if the decision was made to switch to -- either. But my experience is it is pretty common for applications to use a single dash to prefix all commands regardless of length (e.g. Java's keytool utility, Terraria).

As you may have seen I went a slightly different way with the gestalt-module changes (I didn't like that the ModuleSecurityManager and ModuleClassloader had to be modified for the change, so I restructured a little) and released as 3.0.0. So you should now be able to go ahead and make a branch for Terasology with this feature.

 

[OvermindDL1]

From what I can tell there are a plethora of standards for commandline arguments style. I assume you are referring to the gnu style, which uses -- for long argument names and - for short versions. We don't have short versions though nor support for combining short versions, so doesn't really matter (the standard is more for commands rather than applications really). Not that I would care if the decision was made to switch to -- either. But my experience is it is pretty common for applications to use a single dash to prefix all commands regardless of length (e.g. Java's keytool utility, Terraria).

Heh, true, I live on unix systems (SCO UNIX and Debian Linux), putting a single dash in front of multiple characters fills me with a large amount of anxiety without knowing what each character does, and words I instinctively put two dashes in to make sure things are parsed right, but that fails here. ^.^

As you may have seen I went a slightly different way with the gestalt-module changes (I didn't like that the ModuleSecurityManager and ModuleClassloader had to be modified for the change, so I restructured a little) and released as 3.0.0. So you should now be able to go ahead and make a branch for Terasology with this feature.

I saw, yours is a better design, I was going for the quickest change, especially as I was learning gestalt in the process. I will adapt my Terasology branch to handle that and PR it on my next day off (probably Friday) unless someone beats me to it.

Thanks much!