Originally, the ivy_switch mechanism used Bazel genrules to conditionally
compile one TS file or another depending on whether ngc or ngtsc was the
selected compiler. This was done because we wanted to avoid importing
certain modules (and thus pulling them into the build) if Ivy was on or
off. This mechanism had a major drawback: ivy_switch became a bottleneck
in the import graph, as it both imports from many places in the codebase
and is imported by many modules in the codebase. This frequently resulted
in cyclic imports which caused issues both with TS and Closure compilation.
It turns out ngcc needs both code paths in the bundle to perform the switch
during its operation anyway, so import switching was later abandoned. This
means that there's no real reason why the ivy_switch mechanism needed to
operate at the Bazel level, and for the ivy_switch file to be a bottleneck.
This commit removes the Bazel-level ivy_switch mechanism, and introduces
an additional TypeScript transform in ngtsc (and the pass-through tsc
compiler used for testing JIT) to perform the same operation that ngcc
does, and flip the switch during ngtsc compilation. This allows the
ivy_switch file to be removed, and the individual switches to be located
directly next to their consumers in the codebase, greatly mitigating the
circular import issues and making the mechanism much easier to use.
As part of this commit, the tag for marking switched variables was changed
from __PRE_NGCC__ to __PRE_R3__, since it's no longer just ngcc which
flips these tags. Most variables were renamed from R3_* to SWITCH_* as well,
since they're referenced mostly in render2 code.
Test strategy: existing test coverage is more than sufficient - if this
didn't work correctly it would break the hello world and todo apps.
PR Close#26550
Bazel has a restriction that a single output (eg. a compiled version of
//packages/common) can only be produced by a single rule. This precludes
the Angular repo from having multiple rules that build the same code. And
the complexity of having a single rule produce multiple outputs (eg. an
ngc-compiled version of //packages/common and an Ivy-enabled version) is
too high.
Additionally, the Angular repo has lots of existing tests which could be
executed as-is under Ivy. Such testing is very valuable, and it would be
nice to share not only the code, but the dependency graph / build config
as well.
Thus, this change introduces a --define flag 'compile' with three potential
values. When --define=compile=X is set, the entire build system runs in a
particular mode - the behavior of all existing targets is controlled by
the flag. This allows us to reuse our entire build structure for testing
in a variety of different manners. The flag has three possible settings:
* legacy (the default): the traditional View Engine (ngc) build
* local: runs the prototype ngtsc compiler, which does not rely on global
analysis
* jit: runs ngtsc in a mode which executes tsickle, but excludes the
Angular related transforms, which approximates the behavior of plain
tsc. This allows the main packages such as common to be tested with
the JIT compiler.
Additionally, the ivy_ng_module() rule still exists and runs ngc in a mode
where Ivy-compiled output is produced from global analysis information, as
a stopgap while ngtsc is being developed.
PR Close#24056
This commit adds a mechanism by which the @angular/core annotations
for @Component, @Injectable, and @NgModule become decorators which,
when executed at runtime, trigger just-in-time compilation of their
associated types. The activation of these decorators is configured
by the ivy_switch mechanism, ensuring that the Ivy JIT engine does
not get included in Angular bundles unless specifically requested.
PR Close#23833
Alex Rickabaugh
Kara Erickson
Ben Lesh
Miško Hevery