Meta: Update Explainer for Isolated Realms #292

leobalter · 2021-03-27T01:54:33Z

No description provided.

explainer.md

mhofman · 2021-03-27T16:04:24Z

explainer.md

-realm.globalThis.top;
-realm.globalThis.location;
+// Custom properties can be added to the Realm
+// The resolution callback is used as the current API can't wrap promises


It's not quite clear to me what this means.

yes, I wrote this while modifying the example. It makes no sense so I removed it.

mhofman · 2021-03-27T16:50:03Z

explainer.md

-    });
+```javascript
+export default function() {
+  const someRealmIntrinsicsNeededForWrappers = extractIntrinsicsFromGlobal(customGlobalThis);


What is customGlobalThis at this point? Why is it not simply globalThis since this executes in the realm.

I need to review this example with @caridy, it seems like the approach would be different now and there is more to be changed here.

yeah, let's chat about it

Co-authored-by: Mathieu Hofman <86499+mhofman@users.noreply.github.com>

littledan

I'm happy about this direction for the Realms API. Thanks for making these changes.

littledan · 2021-03-31T19:03:32Z

explainer.md

@@ -153,22 +158,23 @@ Instances of Realm Objects and their Global Objects have their lifeline to their

 ### <a name='Evaluation'></a>Evaluation

-The Realms API does not introduce a new way to evaluate code, it is subject to the existing evaluation mechanisms such as the [Content-Security-Policy (CSP)](https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Security-Policy).
+The Realms API does not introduce a new way to evaluate code. Any code is subject to the existing evaluation mechanisms such as the [Content-Security-Policy (CSP)](https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Security-Policy).


Nit: I guess the current API (unlike a previous draft which was just based on globalThis.eval) does create multiple new evaluation mechanisms, it just fits into the existing CSP schema. So, you might want to tweak the wording.

This new mechanism is basically indirect eval, or just mirrors indirect eval, with the restriction that evaluate won't accept non-string values as argument and won't be able to return Non Callable Objects.

I'm tweaking this out for clarification but I believe one could still debate saying this isn't a new evaluation mechanism.

I agree with you; I was making a very narrow comment about the wording. The new wording looks great. This is very much analogous to things that CSP already controls.

littledan · 2021-03-31T19:34:41Z

explainer.md

+```javascript
+export default function() {
+  const someRealmIntrinsicsNeededForWrappers = extractIntrinsicsFromGlobal(customGlobalThis);
+  Object.defineProperties({


Maybe missing the first argument here?

littledan · 2021-03-31T19:37:06Z

explainer.md

- may have host defined properties in addition to the properties defined in this specification. This may include a property whose value is the global object itself.
-
-```
+Yes! It only creates a new copy of the built-ins from ECMAScript.


Hmm, maybe we should link to the issue where we're discussing which globals to include?

I haven't discussed that yet but I believe this new API model leads me to insist more in just providing a copy of the intrinsics. I hope this API resolves that matter.

caridy · 2021-04-05T18:16:08Z

README.md

@@ -50,6 +50,7 @@ See some examples [in the Explainer file](explainer.md).

 ## <a name='History'></a>History

+* we moved on from the exposed globalThis model to a lean isolated realms API (see #289 and #291)


isolated realms => Realm with Callable Boundary or something like that...

we chatted about this with Shu, and I believe we should avoid using the term isolated realm all together.

Another alternative is also "Callable Realms"

I find the names that @caridy is suggesting confusing, though avoiding the word "isolated" could be useful to avoid other kinds of confusion. Maybe we could avoid giving this version a name for now, until we're really confident in one, since it's awkward to change names later.

Suggested change

* we moved on from the exposed globalThis model to a lean isolated realms API (see #289 and #291)

* we moved on from the exposed globalThis model to a revisited format of Realms with a Callable Boundary API (see #289 and #291)

How about this? We can't explain what this is in a single line, but we can further explain the function auto wrapping creates this callable boundary.

caridy · 2021-04-05T18:20:29Z

explainer.md

@@ -107,7 +112,7 @@ These programs must currently contend for the global shared resources, specifica

 Attempting to solve these problems with existing DOM APIs will require to implement an asynchronous communication protocol, which is often a deal-breaker for many use cases. It usually just adds complexity for cases where a same-process Realm is sufficient. It's also very important that values can be immediately shared. Other communications require data to be serialized before it's sent back and forth.

-__The primary goal of this proposal is to provide a proper mechanism to control the execution of a program, providing a new global object, a new set of intrinsics, no default access to the incubator realm's global object, a separate module graph and synchronous communication with the incubator realm__.
+__The primary goal of this proposal is to provide a proper mechanism to control the execution of a program, providing a new global object, a new set of intrinsics, no immediate access to objects cross-realms, a separate module graph and synchronous communication between both realms, while not forcing fingerprints in the constructed realm__.


I also continue having a hard time understanding the finderprints issues with the previous proposal. can you elaborate more?

I'm skeptical that this proposal changes the fingerprinting surface. I think that it'd be best to remove this claim, since it could cause more of these negative reactions about this proposal's authors claiming too much.

I also continue having a hard time understanding the finderprints issues with the previous proposal. can you elaborate more?

That's just about identifying identity from other realms through objects that would then be accessed by globalThis. Of course any framework would handle that, and that's not any new complication (iframe allows that already).

I'll remove this part.

littledan · 2021-04-05T19:13:45Z

README.md

@@ -50,6 +50,7 @@ See some examples [in the Explainer file](explainer.md).

 ## <a name='History'></a>History

+* we moved on from the exposed globalThis model to a lean isolated realms API (see #289 and #291)


I find the names that @caridy is suggesting confusing, though avoiding the word "isolated" could be useful to avoid other kinds of confusion. Maybe we could avoid giving this version a name for now, until we're really confident in one, since it's awkward to change names later.

littledan · 2021-04-05T19:15:58Z

explainer.md

@@ -268,18 +270,35 @@ A big codebase tend to evolve slowly and soon becomes legacy code. Old code vs n

 Modifying code to resolve a conflict (e.g.: global variables) is non-trivial, specially in big codebases.

-The Realms API can provide a _lightweight_ mechanism to preserve the integrity of the intrinsics.0 Therefore, it could isolate libraries or logical pieces of the codebase per Realm.
+The Realms API can provide a _lightweight_ mechanism to preserve the integrity of the intrinsics. Therefore, it could isolate libraries or logical pieces of the codebase per Realm.

 ### <a name='Templatelibraries'></a>Template libraries


I still think that this example should be removed. It will likely initially be too expensive to instantiate this many Realms; we don't want to be telling developers that this is expected to be cheap.

I can remove, but it doesn't require many Realms. You could freeze a realm to operate the templates inside.

littledan · 2021-04-05T19:18:18Z

explainer.md

-a1 instanceof globalTwo.Array; // yield false
-a2 instanceof globalOne.Array; // yield false
-```
+This code is **not** possible with the Realms API! Non-primitive values are not transfered cross-realms using the Realms API.

 ## <a name='Example:NodesvmobjectsvsRealms'></a>Example: Node's vm objects vs Realms

 If you're using node's `vm` module today to "evaluate" javascript code in a different realm, you can replace it with a new Realm, e.g.:


This is no longer quite true, given the restrictions. You probably want to explain that.

246e41b

It's not much but it provides some honest information. WDYT, @littledan?

Co-authored-by: Caridy Patiño <caridy@gmail.com>

It's not part of our primary goals anyway

leobalter · 2021-04-06T16:40:57Z

@caridy @littledan I believe I addressed the latest feedback. PTAL

leobalter · 2021-04-12T19:28:31Z

I'll be glad to address any new feedback in a follow up PR. I'm merging this now to avoid confusion before the TC39 meeting.

Meta: Update Explainer for Isolated Realms

88dd197

leobalter requested review from caridy and littledan March 27, 2021 01:54

mhofman reviewed Mar 27, 2021

View reviewed changes

leobalter and others added 3 commits March 29, 2021 10:32

Update explainer.md

c9718a8

Co-authored-by: Mathieu Hofman <86499+mhofman@users.noreply.github.com>

Update explainer.md

380b797

Co-authored-by: Mathieu Hofman <86499+mhofman@users.noreply.github.com>

fixup, remove unclear comment

a9dcc69

leobalter mentioned this pull request Mar 29, 2021

RFC: Realm.prototype.import #294

Closed

littledan approved these changes Mar 31, 2021

View reviewed changes

leobalter added 3 commits April 1, 2021 12:08

Address review feedback

c9bb8d3

Add Link to discussions

6220f8e

Add link to compartments

a83b493

caridy reviewed Apr 5, 2021

View reviewed changes

littledan reviewed Apr 5, 2021

View reviewed changes

leobalter and others added 4 commits April 6, 2021 08:17

Update explainer.md

0f671db

Co-authored-by: Caridy Patiño <caridy@gmail.com>

Remove sentence about fingerprints

7b88d54

It's not part of our primary goals anyway

Slightly expand comparison with Node vm

246e41b

Remove templates from use cases

20535f0

leobalter merged commit 7187708 into main Apr 12, 2021

leobalter deleted the leo/explainer-isolated-realms branch April 12, 2021 19:29

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Meta: Update Explainer for Isolated Realms #292

Meta: Update Explainer for Isolated Realms #292

leobalter commented Mar 27, 2021

mhofman Mar 27, 2021

leobalter Mar 29, 2021

mhofman Mar 27, 2021

leobalter Mar 29, 2021

caridy Mar 29, 2021

littledan left a comment

littledan Mar 31, 2021

leobalter Mar 31, 2021

littledan Apr 5, 2021 •

edited

Loading

littledan Mar 31, 2021

littledan Mar 31, 2021

leobalter Mar 31, 2021

leobalter Apr 5, 2021

caridy Apr 5, 2021

caridy Apr 5, 2021

littledan Apr 5, 2021

leobalter Apr 6, 2021

This comment was marked as resolved.

caridy Apr 5, 2021

littledan Apr 5, 2021 •

edited

Loading

leobalter Apr 6, 2021

leobalter Apr 6, 2021

littledan Apr 5, 2021

littledan Apr 5, 2021

leobalter Apr 6, 2021

littledan Apr 5, 2021

leobalter Apr 6, 2021

leobalter commented Apr 6, 2021

leobalter commented Apr 12, 2021

		@@ -50,6 +50,7 @@ See some examples [in the Explainer file](explainer.md).

		## <a name='History'></a>History

		* we moved on from the exposed globalThis model to a lean isolated realms API (see #289 and #291)

Meta: Update Explainer for Isolated Realms #292

Meta: Update Explainer for Isolated Realms #292

Conversation

leobalter commented Mar 27, 2021

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

littledan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

littledan Apr 5, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

This comment was marked as resolved.

Choose a reason for hiding this comment

littledan Apr 5, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

leobalter commented Apr 6, 2021

leobalter commented Apr 12, 2021

littledan Apr 5, 2021 •

edited

Loading

littledan Apr 5, 2021 •

edited

Loading