plot.overlay_href does not return the layer

[lgsinnovations]

plot.overlay_href does not return the layer

[desean1625]

Would it make sense to change all the overlay functions to return a promise? This would allow you to write await functions for all the overlay methods.

[desean1625]

Or a non breaking change only return the promise on overlay_href.

[maihde]

@mrecachinas @desean1625 @sterre great news...I actually got this implemented last night. If users were treating the return value from overlay_bluefile and friends as an opaque value then there are no API breaks. If someone was ignoring the return value and hardcoding layer numbers into their push or reload calls then they will need to change their code. I'll submit a PR soon...but you can check out:

spectriclabs@3167fef
spectriclabs@52e8178
spectriclabs@a7d94bd

[maihde]

Here is a PR #89

[mrecachinas]

After discussing this with @maihde, it appears this is intentional because of the asynchronous nature of file-loading.

For the time being, this issue can be solved by modifying the docs.

The longer term solution might include returning the layer index of where the layer will be once the file is loaded.

[sterre]

Would it be helpful for overlay_href to take an optional callback argument, triggered when the layer is added sometime in the future? Or maybe SigPlot could trigger some other event when a layer resolves?

[mrecachinas]

overlay_href does take in an optional callback triggered when the layer is added and you have access to the layer index within the callback.

[sterre]

overlay_href does take in an optional callback triggered when the layer is added and you have access to the layer index within the callback.

Then clearly it was a really good idea and all of us who thought of it independently are Quality People. I do like the idea of being able to record immediately where the layer will be, for future access; but given that you won't know the layer is really there until after the callback fires, I don't see any reason to hurry.

[mrecachinas]

@sterre right -- and that's fair. This came up with bookkeeping within react-sigplot for the HrefLayer. I believe I've found a way around it (using the callback), but it is an open question.

[mrecachinas]

@desean1625 This would definitely be an API breaking change, but I agree it makes a lot of sense. For this big of a change, definitely going to want @maihde to weigh in.

[mrecachinas]

Given 45778ab, where multiple hrefs can now be plotted with overlay_href, should we return a list of promises (i.e., one promise per layer)? A promise where a parameter is a list of layers?

[mrecachinas]

Fixed in #89.