Open shans opened 6 years ago
@lindner says he looked at this previously a bit. @shans are you planning to look at this soon? Perhaps Paul can help, or if you have thoughts on approach, comment and we can discuss further.
This may be P2, I'm marking P1 as @sjmiles has it on his tracking issue https://github.com/PolymerLabs/arcs/issues/2262, but if it's not actually blocking landing shell 0_6_0 then it's probably P2.
So I don't forget, when the shell-0_6_0 encounters this problem it implements a workaround and emits this message:
Pinging for status update as this is marked P1.
I think this should probably be P2, or possibly even closed. I think what's happening is that the serialization code doesn't know what special fiddling the shell context code is doing, so it can't really do anything about the potential for duplicates.
We could allow duplicated importing (which would make this a P2 feature request, I think?) or just accept that this issue will resolve itself when we work out what to do with context (in which case this should be closed).
Sounds ok, although I am curious about the nature of special fiddling. :)
At this point my sense is that Shell isn't doing anything unexpected, so I would like to understand the bits that look iffy to you.
I'm marking it P2. Hopefully I can find time soon to take a quick look (and see if it's even still happening). If I find that time or not, I will aim to close this soonish.
Possibly only relevant to the context manifest, but sometimes the serialization code in arc.ts will directly include a manifest that the context already included.