Fixes #26677
Some packages (like supabase) declare bin entries that don't exist until
lifecycle scripts are run. For instance, the lifecycle script downloads
a binary file which serves as a bin entrypoint.
Unfortunately you can't just defer setting up the bin entries until
after lifecycle scripts have run, because the scripts may rely on them.
I looked into this, and PNPM just re-links bin entries after running
lifecycle scripts. I think that's about the best we can do as well.
Note that we'll only re-setup bin entries for packages whose lifecycle
scripts we run. This should limit the performance cost, as typically a
given project will not have many lifecycle scripts (and of those, many
of them probably don't have bin entries to set up).
Fixes https://github.com/denoland/deno/issues/24749
Runs a server that just returns the header tarball and checksum, and
sets the `NODEJS_ORG_MIRROR` env var so that `node-gyp` uses it instead
of `nodejs.org`
* cts support
* better cjs/cts type checking
* deno compile cjs/cts support
* More efficient detect cjs (going towards stabilization)
* Determination of whether .js, .ts, .jsx, or .tsx is cjs or esm is only
done after loading
* Support `import x = require(...);`
Co-authored-by: Bartek Iwańczuk <biwanczuk@gmail.com>
Fixes #25342.
Still not sure on the exact user agent to set (should it include
`node`?).
After this PR, here's the state of running some `create-*` packages
(just ones I could think of off the top of my head):
| package | prints/runs/suggests deno install | notes |
| ---------------- | ------------- | ------ |
| `create-next-app` | ❌ | falls back to npm, needs a PR
([code](c32e280209/packages/create-next-app/helpers/get-pkg-manager.ts (L3)))
| `sv create` | ❌ | uses `package-manager-detector`, needs a PR
([code](https://github.com/antfu-collective/package-manager-detector/tree/main))
| `create-qwik` | ✅ | runs `deno install` but suggests `deno start`
which doesn't work (should be `deno task start` or `deno run start`)
| `create-astro` | ✅ | runs `deno install` but suggests `npm run dev`
later in output, probably needs a PR
| `nuxi init` | ❌ | deno not an option in dialog, needs a PR
([code](f04e2e8944/src/commands/init.ts (L96-L102)))
| `create-react-app` | ❌ | uses npm
| `ng new` (`@angular/cli`) | ❌ | uses npm
| `create-vite` | ✅ | suggests working deno commands 🎉
| `create-solid` | ❌ | suggests npm commands, needs PR
It's possible that fixing `package-manager-detector` or other packages
might make some of these just work, but haven't looked too carefully at
each
Fixes https://github.com/denoland/deno/issues/26505
I'm not exactly sure how this case comes about (I tried to write tests
for it but couldn't manage to reproduce it), but what happens is the
parent filename ends up null, and we bail out of resolving the specifier
in package exports.
I've checked, and in node the parent filename is also null (so that's
not a bug on our part), but node continues to resolve even in that case.
So this PR should match node's behavior more closely than we currently
do.
Fixes https://github.com/denoland/deno/issues/25862.
npm only makes bin entries executable if they get linked into `.bin`, as
we did before this PR. So this PR actually deviates from npm, because
it's the only reasonable way to fix this that I can think of.
---
The reason this was broken in moment is the following:
Moment has dependencies on two typescript versions: 1.8 and 3.1
If you have two packages with conflicting bin entries (i.e. two
typescript versions which both have a bin entry `tsc`), in npm it is
non-deterministic and undefined which one will end up in `.bin`.
npm, due to implementation differences, chooses to put typescript 1.8
into the `.bin` directory, and so `node_modules/typescript/bin/tsc` ends
up getting marked executable. We, however, choose typescript 3.2, and so
we end up making `node_modules/typescript3/bin/tsc` executable.
As part of its tests, moment executes `node_modules/typescript/bin/tsc`.
Because we didn't make it executable, this fails.
Since the conflict resolution is undefined in npm, instead of trying to
match it, I think it makes more sense to just make bin entries
executable even if they aren't chosen in the case of a conflict.
Fixes #25813.
I initially tried doing this in `deno_semver`, where it's a cleaner
change, but that caused breakage in deno in places where we don't expect
a tag (see https://github.com/denoland/deno/issues/25857).
This does not fix wildcard requirements failing to choose pre-release
versions. That's a little more involved and I'll do a separate PR.
Partially addresses https://github.com/denoland/deno/issues/25648.
This allows packages that use `crossws` to be installed with `deno
install`. `crossws` specifies an optional peer dependency on
`uWebSockets`, but `uWebSockets` is not on npm (it is used with `git:`
or `github:` specifiers). Previously we would error on this, now we
don't error on non-existent optional peer dependencies.
implement require(esm) using `op_import_sync` from deno_core.
possible future changes:
- cts and mts
- replace Deno.core.evalContext to optimize esm syntax detection
Fixes: https://github.com/denoland/deno/issues/25487
Turns out we only virtualized it so one could have a `Console` property,
and the other one not. We can just make this `console.Console` available
everywhere.
PrismJS uses `WorkerGlobalScope` and `self` for detecting browser's Web
Worker context:
59e5a34713/prism.js (L11)
Now the detection logic above is broken when it's imported from Deno's
Web Worker context because we only hide `self` (Prism assumes when
`WorkerGlobalScope` is available, `self` is also available).
This change fixes the above by also hiding `WorkerGlobalScope` global in
Node compat mode.
closes #25008
This change improves the Node.js compatibility in managed npm resolution
mode by disabling the discovery of `node_modules` when the
main specifier is inside of `DENO_DIR`.
closes #22732
closes #24589
NPM inserts a default install script when a package has a `binding.gyp`
file.
It's possible, however, for the package to exclude the `binding.gyp`
file when they publish, and in this case the install script will never
succeed for a user of the package.
This happens with `fsevents`, for instance. They don't include the
`binding.gyp` file in their published tarball, but the default install
script appears in the manifest served by `npm`.
This causes us to warn that `fsevents` has an install script, but when
you try to run it it fails due to `binding.gyp` not existing.
Adds support for running npm package lifecycle scripts, opted into via a
new `--allow-scripts` flag.
With this PR, when running `deno cache` (or `DENO_FUTURE=1 deno
install`) you can specify the `--allow-scripts=pkg1,pkg2` flag to run
lifecycle scripts attached to the given packages.
Note at the moment this only works when `nodeModulesDir` is true (using
the local resolver).
When a package with un-run lifecycle scripts is encountered, we emit a
warning suggesting things may not work and to try running lifecycle
scripts. Additionally, if a package script implicitly requires
`node-gyp` and it's not found on the system, we emit a warning.
Extra things in this PR:
- Extracted out bits of `task.rs` into a separate module for reuse
- Added a couple fields to `process.config` in order to support
`node-gyp` (it relies on a few variables being there)
- Drive by fix to downloading new npm packages to test registry
---
TODO:
- [x] validation for allow-scripts args (make sure it looks like an npm
package)
- [x] make allow-scripts matching smarter
- [ ] figure out what issues this closes
---
Review notes:
- This adds a bunch of deps to our test registry due to using
`node-gyp`, so it's pretty noisy
Adds much better support for the unstable Deno workspaces as well as
support for npm workspaces. npm workspaces is still lacking in that we
only install packages into the root node_modules folder. We'll make it
smarter over time in order for it to figure out when to add node_modules
folders within packages.
This includes a breaking change in config file resolution where we stop
searching for config files on the first found package.json unless it's
in a workspace. For the previous behaviour, the root deno.json needs to
be updated to be a workspace by adding `"workspace":
["./path-to-pkg-json-folder-goes-here"]`. See details in
https://github.com/denoland/deno_config/pull/66
Closes #24340
Closes #24159
Closes #24161
Closes #22020
Closes #18546
Closes #16106
Closes #24160
Previously various reads of files in `node_modules` would error on
invalid UTF-8. These were cases involving:
- reading package.json from Rust
- reading package.json from JS
- reading CommonJS files from JS
- reading CommonJS files from Rust (for ESM translation)
- reading ESM files from Rust
Fixes #24012.
In the case of multiple packages providing a binary with a same name, we
were basically leaving the results undefined (since we set up things in
parallel, and whichever got set up first won). In addition, we were
warning about these cases, even though it's a situation that's expected
to occur.
Instead, in the case of a collision in the binary names, we prefer the
binary provided by the package with the least depth in the dependency
tree.
While I was at it, I also took moved more code to `bin_entries.rs` since
it was starting to get a bit cluttered.
