refactor(ses): Retool for module descriptors #2321
Conversation
kriskowal
force-pushed
the
kriskowal-ses-module-descriptors-refactor
branch
3 times, most recently
from
176f710 to
4d3a95b
Compare
kriskowal
force-pushed
the
kriskowal-ses-module-descriptors-refactor
branch
2 times, most recently
from
0a6fbb9 to
62fe754
Compare
kriskowal
force-pushed
the
kriskowal-ses-module-descriptors-refactor
branch
7 times, most recently
from
194959e to
2d7ace2
Compare
| // A (legacy) behavior: If we do not recognize the module descriptor as a | ||
| // module descriptor, we assume that it is a module source (record): |
There was a problem hiding this comment.
At some point, we’ll want to add options for strict portability and this case will become an exception in that mode. We would stop trying to differentiate module sources from module descriptors.
There was a problem hiding this comment.
I'm fine with the code here in this PR. But I don't understand the comment. Why not continue to differentiate module sources from module descriptors?
There was a problem hiding this comment.
We’ve discriminated all the module descriptors up to this point, but to preserve backward compatibility with import hooks that simply returned a module source (not a module source wrapped in a module descriptor), we have to fall through here.
We do have a clear path to deprecate this path, though, which I hope to get to separately.
There was a problem hiding this comment.
which I hope to get to separately
Worth filing an issue?
kriskowal
force-pushed
the
kriskowal-ses-module-descriptors-refactor
branch
7 times, most recently
from
caf80a9 to
db465e4
Compare
packages/ses/src/module-load.js
Outdated
| } | ||
| if (typeof aliasNamespace === 'string') { | ||
|
|
||
| if (typeof moduleDescriptor === 'string') { | ||
| // eslint-disable-next-line @endo/no-polymorphic-call | ||
| assert.fail( |
There was a problem hiding this comment.
On the function declaration, I'm currently getting the lint warning
131:11 warning Expected to return a value at the end of generator function 'loadWithoutErrorAnnotation' consistent-return
but if I change this line to
| assert.fail( | |
| throw assert.fail( |
and (see below)
There was a problem hiding this comment.
This presumably means that the type isn’t getting properly threaded, since fail and Fail return never. I’ll change this to throw assert.error since throwing at an inevitably throwing call feels weird.
There was a problem hiding this comment.
It felt weird for me the first 100 times. But our codebase now does that pervasively. In throw Fail... I read the Fail as "throw an error" and the throw as "Let static analysis know that we're throwing an error".
packages/ses/src/module-load.js
Outdated
| moduleLoads, | ||
| ); | ||
| } else { | ||
| Fail`module descriptor must be a string or object for specifier ${q( |
There was a problem hiding this comment.
(continued from above)
and modify this line to
| Fail`module descriptor must be a string or object for specifier ${q( | |
| throw Fail`module descriptor must be a string or object for specifier ${q( |
the warning goes away.
packages/ses/test/import.test.js
Outdated
| moduleMapHook(specifier) { | ||
| if (specifier === './index.js') { | ||
| return new StaticModuleRecord('export default 42'); | ||
| } | ||
| }, |
There was a problem hiding this comment.
I get 7 consistent return lint warnings in this file.
| moduleMapHook(specifier) { | |
| if (specifier === './index.js') { | |
| return new StaticModuleRecord('export default 42'); | |
| } | |
| }, | |
| moduleMapHook(specifier) { | |
| if (specifier === './index.js') { | |
| return new StaticModuleRecord('export default 42'); | |
| } | |
| return undefined; | |
| }, |
| // A (legacy) behavior: If we do not recognize the module descriptor as a | ||
| // module descriptor, we assume that it is a module source (record): |
There was a problem hiding this comment.
I'm fine with the code here in this PR. But I don't understand the comment. Why not continue to differentiate module sources from module descriptors?
kriskowal
force-pushed
the
kriskowal-ses-module-descriptors-refactor
branch
from
db465e4 to
5df0f1a
Compare
Refs: #400 #2251
Description
Pursuant to arriving at parity with XS, this change internally reörients the SES module loader around the concept of a “module descriptor”. This creates some clarity and removes some existing duplication of work for “alias” module descriptors, and prepares the material for a richer module descriptor schema to match those implemented by XS.
Security Considerations
No changes.
Scaling Considerations
No changes.
Documentation Considerations
No changes.
Testing Considerations
No changes.
Compatibility Considerations
No changes.
Upgrade Considerations
No changes.