When executing, ngcc writes a lock-file that is used to coordinate multiple concurrent instances of ngcc.
Previously, this file was written at `node_modules/@angular/compiler-cli/ngcc`, or similar depending upon the bundling of the package.
But this causes problems for setups where `node_modules` package directories are expected to be read-only.
Now, the lock-file is written as `.ngcc_lock_file` into the top of the `node_modules`, which is an acceptable place to store transient files.
This change should help to unblock use of tools like pnpm and lerna, which can use symlinks to readonly package directories.
PR Close#44228
Updates the lock file resolution logic in ngcc to work with ESM output.
The compiler-cli is now shipped in bundles, so the actual module resolution
needs to stay to keep the lock file path consistent regardless of where the
lock file code is bundled into. The ngcc integration test needs to be updated
though since the `ngcc` entry-point will always reside in the `bundles/` directory
now.
It has been considered using the top-level `package.json` of the compiler-cli
package, but that caused problems in tests down the line because the ngcc
tests only have the `@angular/compiler-cli/ngcc/...` targets linked into
the node modules. It's not worth changing this and reworking tests if ngcc
is going away in the future anyway (+ it has been like that before!).
PR Close#43431
Switches the compiler-cli usage of `__filename` to `import.meta.url`
when ESM bundles are generated. Unfortunately we cannot start using
only `import.meta` yet as we still build and run all code in Angular
in CommonJS module output for devmode tests.
This commit also fixes various instances where a jasmine spy was applied on
a namespace export that will break with ES module (and the interop for
CommonJS output). We fix these spies by using a default import.
PR Close#43431
ngcc currently dynamially loads the `Transformer` code. It does this
to avoid unnecessary parsing and loading of transformer-related code
if there is nothing to process (so-called noop case). Unfortunately
this dynamic require is not recognized by ESBuild. The import needs
to be discovered as otheriwse the transformer code would not be included
in the bundled package output of the CLI.
The ngcc code needs to use an async runtime import as it would work
in ES modules. This introduces async code into to the compililation
pipeline, breaking the `ngccMain` synchronous invocation feature.
To avoid this, we just move the dynamic require/async import to
the file top-level so that we do not break synchronous processing
which the CLI relies on. This has the downside of slowing-down
the noop case a little but I believe that should be mitigated
through bundling of ngcc anyway. In the future with full-ESM
we won't be able to get around this anyway (unless we remove the
sync variant of ngcc processing).
PR Close#43431
Now that `ReadonlyFileSystem` and `PathManipulation` interfaces are
available, this commit updates the compiler-cli to use these more
focussed interfaces.
PR Close#40281
The link to the "speeding-up-ngcc-compilation" URL does not exist,
it was removed shortly after it was added, but the link in the ngcc
error message was not updated.
Fixes#39837
PR Close#40285
This reverts commit 561c0f81a0.
The original commit provided a quick escape from an already terminal
situation by killing the process if the PID in the lockfile was not
found in the list of processes running on the current machine.
But this broke use-cases where the node_modules was being shared between
multiple machines (or more commonly Docker containers on the same actual
machine).
Fixes#38875
PR Close#39435
Previously an error thrown in the `analyzeFn` would cause
the ngcc process to exit immediately without removing the
lockfile, and potentially before the unlocker process had been
successfully spawned resulting in the lockfile being orphaned
and left behind.
Now we catch these errors and remove the lockfile as needed.
PR Close#37739
The `Logger` interface and its related classes are general purpose
and could be used by other tooling. Moving it into ngtsc is a more
suitable place from which to share it - similar to the FileSystem stuff.
PR Close#37114
Currently, if an ngcc process is killed in a manner that it doesn't clean
up its lock file (or is killed too quickly) the compiler reports that it
is waiting on the PID of a process that doesn't exist, and that it will
wait up to a maximum of N seconds. This PR updates the locking code to
additionally check if the process exists, and if it does not it will
immediately bail out, and print the location of the lock file so a user
may clean it up.
PR Close#37250
Previously the path to the unlocker process was being resolved by the
current file-system. In the case that this was a `MockFileSystemWindows`
on a non-Windows operating system, this resulted in an incorrect path
to the entry-point.
Now the path to the entry-point is hand-crafted to avoid being broken by
whatever FileSystem is in use.
PR Close#36989
When ngcc fails due to a timeout waiting for another process
to complete, it was not failing with a unique exit code, so that it
was not possible to know if the process can be restarted; compared to
ngcc failing for some more fatal reason.
Now if ngcc exits because of a timeout, the exit code will be 177.
PR Close#36838
When ngcc is having to pause and wait for another process
it provides a message to the user. This commit adds the extra
information about how to remove the lockfile if desired, since
this message is not shown if you Ctrl-C out of the process before
the timeout period ends.
PR Close#36838
The cached file-system was implemented to speed up ngcc
processing, but in reality most files are not accessed many times
and there is no noticeable degradation in speed by removing it.
Benchmarking `ngcc -l debug` for AIO on a local machine
gave a range of 196-236 seconds with the cache and 197-224
seconds without the cache.
Moreover, when running in parallel mode, ngcc has a separate
file cache for each process. This results in excess memory usage.
Notably the master process, which only does analysis of entry-points
holds on to up to 500Mb for AIO when using the cache compared to
only around 30Mb when not using the cache.
Finally, the file-system cache being incorrectly primed with file
contents before being processed has been the cause of a number
of bugs. For example https://github.com/angular/angular-cli/issues/16860#issuecomment-614694269.
PR Close#36687
The change in e041ac6f0d
to support sending unlocker process output to the main ngcc
console output prevents messages require that the main process
relinquishes the event-loop to allow the `stdout.on()` handler to
run. This results in none of the messages being written when ngcc
is run in `--no-async` mode, and some messages failing to be
written if the main process is killed (e.g. ctrl-C).
It appears that the problem with Windows and detached processes
is known - see https://github.com/nodejs/node/issues/3596#issuecomment-250890218.
But in the meantime, this commit is a workaround, where non-Windows
`inherit` the main process `stdout` while on Windows it reverts
to the async handler approach, which is better than nothing.
PR Close#36637
On Windows, the output of a detached process (such as the unlocker
process used by `LockFileWithChildProcess`) is not shown in the parent
process' stdout.
This commit addresses this by piping the spawned process' stdin/stdout
and manually writing to the parent process' stdout.
PR Close#36569
The current ngcc lock-file strategy spawns a new process in order to
capture a potential `SIGINT` and remove the lock-file. For more
information see #35861.
Previously, this unlocker process was spawned as soon as the `LockFile`
was instantiated in order to have it available as soon as possible
(given that spawning a process is an asynchronous operation). Since the
`LockFile` was instantiated and passed to the `Executor`, this meant
that an unlocker process was spawned for each cluster worker, when
running ngcc in parallel mode. These processes were not needed, since
the `LockFile` was not used in cluster workers, but we still had to pay
the overhead of each process' own memory and V8 instance.
(NOTE: This overhead was small compared to the memory consumed by ngcc's
normal operations, but still unnecessary.)
This commit avoids the extra processes by only spawning an unlocker
process when running on the cluster master process and not on worker
processes.
PR Close#36569
This version of `LockFile` creates an "unlocker" child-process that monitors
the main ngcc process and deletes the lock file if it exits unexpectedly.
This resolves the issue where the main process could not be killed by pressing
Ctrl-C at the terminal.
Fixes#35761
PR Close#35861
