Previously in the template type-checking engine, it was assumed that every
input file would have an associated type-checking shim. The type check block
code for all components in the input file would be generated into this shim.
This is fine for whole-program type checking operations, but to support the
language service's requirements for low latency, it would be ideal to be
able to check a single component in isolation, especially if the component
is declared along with many others in a single file.
This commit removes the assumption that the file/shim mapping is 1:1, and
introduces the concept of component-to-shim mapping. Any
`TypeCheckingProgramStrategy` must provide such a mapping.
To achieve this:
* type checking record information is now split into file-level data as
well as per-shim data.
* components are now assigned a stable `TemplateId` which is unique to the
file in which they're declared.
PR Close#38105
This optimization builds on a lot of prior work to finally make type-
checking of templates incremental.
Incrementality requires two main components:
- the ability to reuse work from a prior compilation.
- the ability to know when changes in the current program invalidate that
prior work.
Prior to this commit, on every type-checking pass the compiler would
generate new .ngtypecheck files for each original input file in the program.
1. (Build #1 main program): empty .ngtypecheck files generated for each
original input file.
2. (Build #1 type-check program): .ngtypecheck contents overridden for those
which have corresponding components that need type-checked.
3. (Build #2 main program): throw away old .ngtypecheck files and generate
new empty ones.
4. (Build #2 type-check program): same as step 2.
With this commit, the `IncrementalDriver` now tracks template type-checking
_metadata_ for each input file. The metadata contains information about
source mappings for generated type-checking code, as well as some
diagnostics which were discovered at type-check analysis time. The actual
type-checking code is stored in the TypeScript AST for type-checking files,
which is now re-used between programs as follows:
1. (Build #1 main program): empty .ngtypecheck files generated for each
original input file.
2. (Build #1 type-check program): .ngtypecheck contents overridden for those
which have corresponding components that need type-checked, and the
metadata registered in the `IncrementalDriver`.
3. (Build #2 main program): The `TypeCheckShimGenerator` now reuses _all_
.ngtypecheck `ts.SourceFile` shims from build #1's type-check program in
the construction of build #2's main program. Some of the contents of
these files might be stale (if a component's template changed, for
example), but wholesale reuse here prevents unnecessary changes in the
contents of the program at this point and makes TypeScript's job a lot
easier.
4. (Build #2 type-check program): For those input files which have not
"logically changed" (meaning components within are semantically the same
as they were before), the compiler will re-use the type-check file
metadata from build #1, and _not_ generate a new .ngtypecheck shim.
For components which have logically changed or where the previous
.ngtypecheck contents cannot otherwise be reused, code generation happens
as before.
PR Close#36211
In #33551, a bug in `ngc --watch` mode was fixed so that a component is
recompiled when its template file is changed. Due to insufficient
normalization of files paths, this fix did not have the desired effect
on Windows.
Fixes#32869
PR Close#34015
Previously, the compiler performed an incremental build by analyzing and
resolving all classes in the program (even unchanged ones) and then using
the dependency graph information to determine which .js files were stale and
needed to be re-emitted. This algorithm produced "correct" rebuilds, but the
cost of re-analyzing the entire program turned out to be higher than
anticipated, especially for component-heavy compilations.
To achieve performant rebuilds, it is necessary to reuse previous analysis
results if possible. Doing this safely requires knowing when prior work is
viable and when it is stale and needs to be re-done.
The new algorithm implemented by this commit is such:
1) Each incremental build starts with knowledge of the last known good
dependency graph and analysis results from the last successful build,
plus of course information about the set of files changed.
2) The previous dependency graph's information is used to determine the
set of source files which have "logically" changed. A source file is
considered logically changed if it or any of its dependencies have
physically changed (on disk) since the last successful compilation. Any
logically unchanged dependencies have their dependency information copied
over to the new dependency graph.
3) During the `TraitCompiler`'s loop to consider all source files in the
program, if a source file is logically unchanged then its previous
analyses are "adopted" (and their 'register' steps are run). If the file
is logically changed, then it is re-analyzed as usual.
4) Then, incremental build proceeds as before, with the new dependency graph
being used to determine the set of files which require re-emitting.
This analysis reuse avoids template parsing operations in many circumstances
and significantly reduces the time it takes ngtsc to rebuild a large
application.
Future work will increase performance even more, by tackling a variety of
other opportunities to reuse or avoid work.
PR Close#34288
