Developer Notes - 2021-02 Ruleset Updates

[Moon Wizard]

My plan is to move all of these changes into the Test channel tonight; with the plan to release on Feb. 16.

Community ruleset and extension developers should start looking sooner than later, as this is a faster cycle than usual. I'm working to clear this set of changes out before any ruleset vision/lighting changes we may need to make.

Regards,
JPG

[SilentRuin]

My plan is to move all of these changes into the Test channel tonight; with the plan to release on Feb. 16.

Community ruleset and extension developers should start looking sooner than later, as this is a faster cycle than usual. I'm working to clear this set of changes out before any ruleset vision/lighting changes we may need to make.

Regards,
JPG

I assume that I can't make the changes and have them work with the current FGU version correct? And just to point out the obvious, as I track the forums pretty good and missed this because of the title not interesting me, you best be ready for a lot of false claims that it busted people. Because it will really be the extensions busting the game - not the update. And then... wowsers. Just be ready.

[Moon Wizard]

It's true for every release; and that's why the first thing when errors happen is that we tell people to turn off extensions.

Regards,
JPG

[SilentRuin]

It's true for every release; and that's why the first thing when errors happen is that we tell people to turn off extensions.

Regards,
JPG

Preaching to the choir. 100% agree. EXTENSION = RISK is my motto. Posting in a few discord chats as I know people don't track the forums that well. And this update will be different than normal releases which have never impacted me. I think this one will with some of the changes. But I'll find out tomorrow or the weekend when i try to figure out how to work on test version but still somehow flip back to the normal version for my Monday night game.

[lule]

Is Session a new API-class? Is it a replacement for User-class? Could you please post the functions and properties of Seesion?

Thank you.

Regards,

Lule

[deer_buster]

It's true for every release; and that's why the first thing when errors happen is that we tell people to turn off extensions.

Regards,
JPG

Just tossing an idea out here, but this would be an awesome time to think about extension self-validation vs a given ruleset version, and if it doesn't meet that validation criteria, it doesn't load and displays an error for the user. For a semantic versioning scheme (I assume this is the versioning scheme you are using?), you'd only require it if the major version number changed. Perhaps a warning if the minor version wasn't validated too, but that's only for backwards compatible changes anyway.

[SilentRuin]

Just tossing an idea out here, but this would be an awesome time to think about extension self-validation vs a given ruleset version, and if it doesn't meet that validation criteria, it doesn't load and displays an error for the user. For a semantic versioning scheme (I assume this is the versioning scheme you are using?), you'd only require it if the major version number changed. Perhaps a warning if the minor version wasn't validated too, but that's only for backwards compatible changes anyway.

You must be living under a rock the last month or so. The whole DMsG dumping extensions and SW coming up with FG Forge (extension delivery system) is already been beaten to death in the forums - now its just a waiting game. Rock removed, daylight exposed - don't blink!

[deer_buster]

You must be living under a rock the last month or so. The whole DMsG dumping extensions and SW coming up with FG Forge (extension delivery system) is already been beaten to death in the forums - now its just a waiting game. Rock removed, daylight exposed - don't blink!

No rock living, thanks for the concern. Having the Forge doesn't imply validation against ruleset changes...

[SilentRuin]

No rock living, thanks for the concern. Having the Forge doesn't imply validation against ruleset changes...

True - but that thread is more likely to be relevant on that concern - figured you just did not know it was out there. Plus validating against a ruleset is near impossible for anyone but the developer of the extension. To many variables - to much nuance and inter compatibility issues with other extensions. Just because an extension works with a ruleset does not mean they did not write it irresponsibility and have it stomp over lots of code and other extensions. Not something I can envision being done without quadrupling the size of their testing teams. And why should they be testing other peoples code anyway? That is a sure ticket to endless headaches. EXTENSIONS = RISK - you accept that - or should stay far far away from them. And no I'm not trying to give you a hard time - just my opinion on this matter. Don't mean to say you should not have your say on it.

[deer_buster]

True - but that thread is more likely to be relevant on that concern - figured you just did not know it was out there. Plus validating against a ruleset is near impossible for anyone but the developer of the extension. To many variables - to much nuance and inter compatibility issues with other extensions. Just because an extension works with a ruleset does not mean they did not write it irresponsibility and have it stomp over lots of code and other extensions. Not something I can envision being done without quadrupling the size of their testing teams. And why should they be testing other peoples code anyway? That is a sure ticket to endless headaches. EXTENSIONS = RISK - you accept that - or should stay far far away from them. And no I'm not trying to give you a hard time - just my opinion on this matter. Don't mean to say you should not have your say on it.

You didn't understand the concept then. The author of the extension would certify that they have checked the code against a specific major version, and when the client loads the extension, it would compare that ruleset's version against the certified version of the extension, and provide an error or warning depending upon whether it is a major or minor version violation. If it is an error, provide the option to skip loading that extension into memory at that point. At no point should any expectation be that FG, or any other program, would look at the code itself and try to validate it. As you said, too much variability.