Stream: t-compiler/major changes

Topic: create windows target group compiler-team#293


triagebot (May 15 2020 at 14:32, on Zulip):

A new proposal has been announced #293. It will be brought up at the next meeting.

Daniel Silverstone (May 20 2020 at 09:53, on Zulip):

May I please ask that consideration be given to extending this to the full rust-lang ecosystem, e.g. the various dev-tools often have Windows specific issues, particuarly rustup.

simulacrum (May 20 2020 at 11:01, on Zulip):

Yeah, I think we wouldn't at least initially expect to extend it beyond the rust teams but anyone within that set should feel free to ask on "official" questions imo

Daniel Silverstone (May 20 2020 at 12:24, on Zulip):

OK, that's good to know. Especially if you end up with anyone properly Windowsy who might be able to help with strange behaviours under anti-virus etc (though I appreciate that's not the goal of the compiler group per-se) -- OOI is there any potential for Microsoft to sponsor someone into this kind of role?

nikomatsakis (May 20 2020 at 13:19, on Zulip):

So @XAMPPRocky brought up the question of whether the name "target group" ought to be changed, and I think a particular concern was with the question of "what targets need groups" and so forth. (My take is that eventually we should have groups for every "target" we support, but I would prefer to go step by step vs trying to do something exhaustive to start.) They proposed "experts group", but I am somewhat wary of this term "expert".

nikomatsakis (May 20 2020 at 13:19, on Zulip):

I do think we want to extend this concept to cover areas of the compiler -- e.g., the type checker. Maybe we should just call it a "group" :)

nikomatsakis (May 20 2020 at 13:19, on Zulip):

With "area groups" being the general term or something like that.

nikomatsakis (May 20 2020 at 13:20, on Zulip):

@XAMPPRocky I am reminded of the question of what distinguishes a "project group" from a "working group" -- these feel like "working groups" to me, in that they have indefinite duration and are scoped not to some particular task but to a "domain" or "area".

nikomatsakis (May 20 2020 at 13:21, on Zulip):

So maybe it's just that the compiler has Working Groups for various parts of the compiler as well as compilation targets and other areas of expertise.

nikomatsakis (May 20 2020 at 13:22, on Zulip):

Daniel Silverstone said:

OK, that's good to know. Especially if you end up with anyone properly Windowsy who might be able to help with strange behaviours under anti-virus etc (though I appreciate that's not the goal of the compiler group per-se) -- OOI is there any potential for Microsoft to sponsor someone into this kind of role?

@Ryan Levick mentioned that they've found a number of candidates of people to get involved

Daniel Silverstone (May 20 2020 at 13:25, on Zulip):

nikomatsakis said:

Ryan Levick mentioned that they've found a number of candidates of people to get involved

That's fantastic.

XAMPPRocky (May 20 2020 at 13:25, on Zulip):

I was just about suggest "domain" is pretty good term for it that doesn't have gate-keepy connotations that expert does. My other concern around "target" is that I think there are some groups like a Windows group would have value to teams outside of the compiler team.

nikomatsakis (May 20 2020 at 13:25, on Zulip):

Yes, @Daniel Silverstone raised this point and I agree.

nikomatsakis (May 20 2020 at 13:26, on Zulip):

This doesn't really need to be scoped to the compiler per se

nikomatsakis (May 20 2020 at 13:34, on Zulip):

OK I reworked the proposal somewhat

nikomatsakis (May 20 2020 at 13:35, on Zulip):

I just used the term "working group for a target"

nikomatsakis (May 20 2020 at 13:35, on Zulip):

and not "target group"

Ryan Levick (May 20 2020 at 14:34, on Zulip):

We have folks joining with MSVC experience but also people from Windows dev tooling who won't necessarily be compiler experts but will be Windows experts. I hope this won't just be confined to compiler work.
The only concern I've had expressed is that for now this should be seen as individuals contributing. They _might_ be doing so during work time with blessing from their managers but Microsoft as entity does not wish to formally support this work at this time. Does that make sense?

nikomatsakis (May 20 2020 at 14:41, on Zulip):

Yes, that is my expectation.

nikomatsakis (May 20 2020 at 14:42, on Zulip):

(Individuals, this isn't an official MS sponsorship of any kind.)

nikomatsakis (May 20 2020 at 14:43, on Zulip):

In general I wouldn't really expect companies to "officially sponsor" in that sense when it comes to development work like this, we're always dealing with individuals (who may get paid to participate, but we're participating with them as individuals).

nikomatsakis (May 21 2020 at 18:09, on Zulip):

@pnkfelix care to second this MCP?

pnkfelix (May 21 2020 at 20:46, on Zulip):

let me read it properly first

pnkfelix (May 21 2020 at 20:48, on Zulip):

nikomatsakis said:

I just used the term "working group for a target"

is this change just in the hackmd write up? I don't see it in compiler-team#293 itself ...?

pnkfelix (May 21 2020 at 20:49, on Zulip):

(it seems like that must be indeed what happened.)

pnkfelix (May 21 2020 at 20:49, on Zulip):

probably best to update the issue as well, since some people might never read the writeup

pnkfelix (May 21 2020 at 20:49, on Zulip):

(perhaps this is all the more reason to change this format to be PR's instead of issues with inline text, though that then adds a barrier to entry for filing MCP's)

pnkfelix (May 21 2020 at 20:50, on Zulip):

@nikomatsakis you want me to try to update the issue text accordingly?

nikomatsakis (May 21 2020 at 20:50, on Zulip):

Ah, yeah, sure!

nikomatsakis (May 21 2020 at 20:50, on Zulip):

Forgot about the issue text indeed

Last update: May 07 2021 at 07:00UTC