#development
- w
witty-crayon-22786
06/21/2018, 6:24 PM@rough-minister-58256: ok, will comment - r
rough-minister-58256
06/21/2018, 6:24 PM👍 - r
rough-minister-58256
06/21/2018, 6:26 PMcommented re: weighting, but I’m open to ideas there. I’d just gone with the simplest strategy I could think of because I agree: probably, WAGNIa- 2
- 1
- r
rough-minister-58256
06/21/2018, 6:27 PMcompletely on board for a type-centric path there, but: 1) what could that look like 2) do we really need itw- 2
- 20
- r
rough-minister-58256
06/21/2018, 6:29 PMwill mull that over some more today - r
rough-minister-58256
06/21/2018, 6:30 PManother thought there was mapping @console_rules to discrete named execution phases, but I couldn’t think of a sane default ordering for that and it seemed like it wouldn’t allow much flexibility - w
witty-crayon-22786
06/21/2018, 6:33 PM@rough-minister-58256: random thought after reading some more about LSP and BSP servers: - w
witty-crayon-22786
06/21/2018, 6:33 PM"@foreground_rule" ...? - r
rough-minister-58256
06/21/2018, 6:33 PMas replacement for @console_rule? or variant? - w
witty-crayon-22786
06/21/2018, 6:33 PMyea. - w
witty-crayon-22786
06/21/2018, 6:33 PMreplacement. - w
witty-crayon-22786
06/21/2018, 6:33 PMjust a naming thing. - r
rough-minister-58256
06/21/2018, 6:33 PMin name alone, or functional? - r
rough-minister-58256
06/21/2018, 6:33 PMah - r
rough-minister-58256
06/21/2018, 6:34 PMmaybe - w
witty-crayon-22786
06/21/2018, 6:34 PMheh. doesn't matter. - w
witty-crayon-22786
06/21/2018, 6:35 PM@rough-minister-58256: i'm going to draft the release notes and then comment on the subtasks... sorry - r
rough-minister-58256
06/21/2018, 6:35 PMyeah, a code review point I suppose - r
rough-minister-58256
06/21/2018, 6:35 PMsg! - r
rough-minister-58256
06/21/2018, 6:36 PMmaybe if the distinction was
vs `@foreground_rule`; I sort of think@background_rule
vs@rule
is more descriptive, esp given e.g. historical@console_ruleConsoleTask - r
rough-minister-58256
06/21/2018, 6:36 PMbut prob just comes down to what feels right in the PR - r
rough-minister-58256
06/21/2018, 6:37 PMthe naming in HLD is not intended to be sealed, just fodder - a
aloof-angle-91616
06/21/2018, 6:55 PMi think a useful distinction (like i mentioned on the doc) is to be able to bind execution of some rule with a handle to some particular communication interface that the rule uses/"owns" (?), and a "console" vs "socket talking to an lsp server" both feel like they can conform to that representation without having to have a distinction between foreground and background, which seems like it could be a goal - w
witty-crayon-22786
06/21/2018, 6:56 PM...oo. yea, actually that's a very good point.👌 1 - a
aloof-angle-91616
06/21/2018, 6:56 PMthat's what i'm saying - w
witty-crayon-22786
06/21/2018, 6:57 PMleaning toward
, and then differentiating between "consoley" and "sockety" based on whether you request injection of the Console or a Socket@foreground_rule - a
aloof-angle-91616
06/21/2018, 6:58 PMyeah definitely a console vs a pipe vs a socket have different capabilities and no clue of the best way to represent that yet but "everything is a file" already - w
wide-energy-11069
06/21/2018, 6:59 PMdoes
imply async at all? or everything is async@background_rule - w
witty-crayon-22786
06/21/2018, 6:59 PM@wide-energy-11069: it's just `@rule`: and yes, it is 100% async👌 1 - a
aloof-angle-91616
06/21/2018, 7:00 PMand like it's a different model i think because you can't create a console -- you are invoked and have one console, and the rule would use it, but in the case of an lsp server, you would/could create one, which generates a new such handle -- no clue if this matters impl wisew- 2
- 7