Namespaced Attributes

[kaalleen]

This is to solve issues #201 and #480

[lexelby]

That's so clever, using the attribute namespace transition as a way to handle the underlay default changing!

What happens if they had enabled underlay previously though? I think this would set it to false?

[kaalleen]

Ops, so right. I've corrected it.

If no other "embroider_" attribute has been set on the object, fill_underlay will switch it's value from False to True - which, I think it won't be too bad. Also the fill_underlay param will be set on other non-fill-elements. I don't think it's much of a problem though...?!?

[lexelby]

Ah, that makes much more sense, thank you!

This is a pretty good time to look over all the other params. Any more defaults we should change?

[lexelby]

Also the fill_underlay param will be set on other non-fill-elements. I don't think it's much of a problem though...?!?

Nope, shouldn't be a problem at all.

[lexelby]

This looks great so far!

I wonder, do we really need INKSTITCH_ATTRIBS? I'm worried that we'll forget to add new attributes to the list.

I think maybe you created that so that we're not constantly calling addNS over and over again. That may be a good use for a function with the @cache decorator instead?

[kaalleen]

Hm, I don't know. I don't think it is possible to forget adding them ... you will get an error if you do and test it only once :)

At first I was annoyed by having to add all the attributes into a list and thought that there must be a better solution for it (like the one you suggested). On the other hand I thought it delivers a good overview of the attributes we have and also we keep the tags all in the same place. I haven't tested your suggested solution yet. Maybe it's the way to go, but I don't think it's all bad to have the ATTRIBS list.

I wouldn't know about any other default setting that need a change.
Until now we have no complaints...

[lexelby]

Good argument! You've won me over :)

[kaalleen]

Yay, I'll merge it then :)