simplify check_cfg_arg impl #147570
simplify check_cfg_arg impl #147570
Conversation
|
Some changes occurred in cfg and check-cfg configuration cc @Urgau |
rustbot
added
S-waiting-on-review
|
rustbot has assigned @Mark-Simulacrum. Use |
This comment has been minimized.
This comment has been minimized.
|
r? Urgau |
rust-cloud-vms
bot
force-pushed
the
cleanup-bootstrap
branch
from
90bd404 to
33fa88d
Compare
This comment has been minimized.
This comment has been minimized.
rust-cloud-vms
bot
force-pushed
the
cleanup-bootstrap
branch
from
33fa88d to
78cacb9
Compare
| pub fn check_cfg_arg(name: &str, values: &[&str]) -> String { | ||
| if values.is_empty() { | ||
| format!("--check-cfg=cfg({name})") | ||
| } else { | ||
| format!("--check-cfg=cfg({name},values(\"{}\"))", values.join("\",\"")) | ||
| } | ||
| } |
There was a problem hiding this comment.
This is not technically correct, having an empty values() is valid and has the semantic of defining a cfg that has no expected values (and thus will always warn).
But it doesn't really make sense for bootstrap to be defining such cfgs.
There was a problem hiding this comment.
I read about this, but it seems really weird that values() makes the cfg name unusable(?) and what accepts an empty value is values(none()).
| const EXTRA_CHECK_CFGS: &[(Option<Mode>, &str, &[&str])] = &[ | ||
| (Some(Mode::Rustc), "bootstrap", &[]), | ||
| (Some(Mode::Codegen), "bootstrap", &[]), | ||
| (Some(Mode::ToolRustcPrivate), "bootstrap", &[]), | ||
| (Some(Mode::ToolStd), "bootstrap", &[]), | ||
| (Some(Mode::ToolRustcPrivate), "rust_analyzer", &[]), | ||
| (Some(Mode::ToolStd), "rust_analyzer", &[]), |
There was a problem hiding this comment.
What do you think about having &[Option<&str>] for representing the expected values? which would make the none() part explicit.
The generation part would then always be like that cfg(name, values(...)).
There was a problem hiding this comment.
That would allow &[None, None, None] which seems silly...
5 participants
No description provided.