Meeting minutes
<jgraham> RRSAgent: make minutes
Reminder to review the PR for screencasting API
github: w3c/
jdescottes: This is a reminder to review the PR from sasha. Sasha is on PTO at the moment, please raise questions and we will answer to the best of our ability
orkon: I've been trying to prototype this and there web content API has been quite hard to use and there is nothing about stitching things so would need to be invited
… but building it off web apis isn't doable
… are there restrictions to what format needs to be output or is that up to the implementation
jgraham: I don't know what web rtc has but that format is likely the best
… the other question has...
… with web apis there is an idea that you are calling it from a IDL context which assumes javascript etc
… we have a similar issue with the streams API
… but when I suggested using something not using JS on streams the editor told me that wouldn't work as it would need a way of resolving promises
… we would likely need to have a "specification realm" that is not a content realm and is something else
… and it spec "fiction". It's like running script but not exactly like running a script <gives an example>
… in Firefox this is easier and we know that doesn't mean much as in other browsers it could be a C++ type call
… that's my best idea but if we don't do that then we are in a difficult position
… it then becomes a very handwavy way doing "capture this and then output that with web rtc"
… and it becomes unclear what realms this should run in
orkon: thanks for the explaination. THis seems to work well for streams but I am not sure that works for screencast
… but I might not have fully read those spec. The main thing is that we don't depend of the side effects of javascript execution other than the algorithm itself
jgraham: I need to look to see if this would work. parts of it works well but as a whole maybe not. THere might be a possiblity is that this is very complicated. We also have a way in the future that we can stream the browser not the file what we are currently thinging about
Should modified request cookies be present in the RequestData for network events
github: w3c/
jdescottes: cookies on modified requests... I realised when serialsing the network events that the current spec that we put the old cookie data into future events
… this is because we are looping over the cookiejar and serialising what matches the network request
… or should we returned the modified. If we do the modified that we can't return all the values as we can't see them all
… or should we return the cookies in a different event later?
orkon: I've commented on the issue... but in general from puppeteer that we don't have a use case for cookies on a network request. we have a separate cookie api
… but I think we should the modified or a way to notify the user there is a change. I don't have a strong opinion on this either way
jdescottes: if there no other opinion then I will update the test accordingly
… on the client side, do you have opinions on this @jimevans?
jimevans: I will need to check my use case that prompted me to log the bug
… I can't remember what I was doing that caused the disconnect
emulation.setTextLayoutModeOverride
github: w3c/
orkon: there are a few PRs raised by Maksim. I wanted to check if there was anything that people wanted to do here?
whimboo: just checking is there a priority of PRs?
jgraham: I can see that Sasha has looked at this
jdescottes: it looks like some have been approved by sasha and we will ask her when back
<jgraham> RRSAgent: make minutes