Portabilize Stdlib #3393
Open
Portabilize Stdlib #3393
+4,432
−1,799
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
tdelvecchio-jsc
force-pushed
the
tdelvecchio-portablize-stdlib
branch
from
569c742 to
8096435
Compare
tdelvecchio-jsc
force-pushed
the
tdelvecchio-portablize-stdlib
branch
from
ce744e4 to
b14c2da
Compare
tdelvecchio-jsc
commented
Jan 2, 2025
tdelvecchio-jsc
commented
Jan 2, 2025
| [@@alert unsynchronized_access | ||
| "GC parameters are a mutable global state." | ||
| ] | ||
| (** Return the current values of the GC parameters in a [control] record. *) | ||
|
|
||
| external set : control -> unit = "caml_gc_set" | ||
| external set : control -> unit @@ nonportable = "caml_gc_set" |
There was a problem hiding this comment.
Should these be alerted?
There was a problem hiding this comment.
Yeah, I think we can remove the alert now. Feel free to do it here or in another PR.
There was a problem hiding this comment.
Sorry, I meant should these be alerted as [@@alert unsafe_multidomain]? They are unsafe w.r.t. portability/contention, but there's no "safe" version that's easily writable.
There was a problem hiding this comment.
Why is it unsafe if annotated with nonportable?
There was a problem hiding this comment.
Continuing discussion in Slack, but basically these functions use atomic_load_relaxed/atomic_store_relaxed, which I think are unsafe in general without some kind of fencing (though I really don't understand this stuff well).
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
tdelvecchio-jsc
force-pushed
the
tdelvecchio-portablize-stdlib
branch
from
0b26357 to
529cbf5
Compare
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
Replace the [@@unsafe_allow_any_kind_in_{impl,intf}] attributes with a new
[@@unsafe_allow_any_mode_crossing] attribute. This is different in that it:
1. Works on the type declaration, not the inclusion check, so is more powerful -
it can be used to illegally mode cross types defined in the same module, or
illegally mode cross non-abstract types in interfaces. The latter especially
is necessary to fully subsume -allow-illegal-crossing in stdlib
2. Only allows changing the modal bounds of a kind, not the layout - it's
unclear that anyone should ever want to unsafely change the layout of a kind;
I personally can't think of any sound reason to do so.
Some [past discussion][0] on the specific syntax for this attribute suggested
specifying the bounds directly on the attribute, rather than using the jkind
annotation, but I and others feel reasonably strongly that this way of doing
things is a better design for the sake of consistency.
[0]: #3385 (comment)
Signed-off-by: Thomas Del Vecchio <[email protected]>
Signed-off-by: Thomas Del Vecchio <[email protected]>
|
Does the tree build with this PR? |
|
Could you move the |
| @@ -28,6 +30,8 @@ type raw_data = nativeint (* @since 4.12 *) | |||
| external repr : 'a -> t = "%obj_magic" | |||
| external obj : t -> 'a = "%obj_magic" | |||
| external magic : 'a -> 'b = "%obj_magic" | |||
| external magic_portable : 'a -> 'a @ portable = "%identity" | |||
There was a problem hiding this comment.
I think those should have alerts, even if Obj.magic does not have one.
This commit comes from #3424 . Updated PR description to mention this. |
Working on it, but not yet; there's a few bugs I'm working with @riaqn on. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR updates stdlib in three main ways:
portablewhen possible.[@alert unsafe_multidomain "Use [...]."].Safesubmodules when necessary to redefine unsafe functions from (2) with necessary mode annotations.One place in particular that needed some work is
Domain, which needed a rework of theDLSmodule to fit more in line with the capsule model.Portable versions of functors in
Set,Map,Hashtbl, andEphemeronare also created, which takeportableinput modules and produceportableoutput modules.Tests are updated to disable the "unsafe_multidomain" alert rather than moving tests to the new "safe" versions.
This PR currently depends on #3391 and #3424 .
Changes should be reviewed as a whole, not by commit.