For now, force the clang-based actions to skip the end of the pipeline. This restores the previous behavior of the frontend, but may not be desirable in the long run. For example, one may want to dump clang stats after running an -emit-pch job, but that is impossible without forcing the end of the pipeline to be more tolerant of ObjCHeader/modulemap-only inputs.
rdar://68587228
Try to impose a simple structure that splits performing actions from the
pre and post-pipeline conditions. Wherever actions would take more than
a simple return, split them into functions. Refine functions that
perform effects to return status codes when they fail. Finally,
delineate functions that need semantic analysis from those that do not.
Overall this should be NFC.
To help solving rdar://67079780, this change allows swift-driver to configure scanner using additional
arguments passed down via the batch input JSON file for each module under scanning.
This scanning mode allows swift-driver to query module dependencies in a batch
and in a more granular way. In short term, it could help solve a problem that
clang module dependencies may vary if target triple changes. In a longer term,
we could break a holistic dependencies graph into smaller pieces for better caching
and reusing.
This change doesn't include the implementation of using the specified scanner
arguments to set up Clang dependencies scanner. It will come in later commits.
For the issue mentioned in rdar://67079780, swift-driver needs to run clang dependencies
scanner multiple times with different target triples for a Swift target. This patch adds
a new scanning action to generate the JSON file for a given clang module to accommodate
this requirement.
Resolves: rdar://problem/67269210
Based on measurements, it seemed to save about 0.2 ms - 0.4 ms for a module
with a large number of imports from the SDK. That's insignificant compared to
the compile time for a typical module, which usually takes somewhere between
a few seconds and a few minutes.
- deduplicate the logic to compute the resource folder
- install headers and module files in shared and static resource folders
- forward -static flag when calling swiftc with -print-target-info
Seperating out TBD file generation uncovered a bug
in the frontend where we could fail to notice that
we emitted an error during the end-of-pipeline
actions and returned a zero exit code. Tweak the
logic to detect this case.
We need to traverse the module dependency graph and track which modules expose
which other modules' ABIs, while making sure that we don't hit a loop while
trawling through Clang (sub)modules.
Fixes rdar://64993153.
This flag no longer does anything now that the unified statistics
reporting infrastructure exists. It is better to use
-driver-time-compilation to see a bird's eye view of timing statistics
for frontend jobs, and -stats-output-dir to see a down-and-dirty view of
everything including performance counters.
A couple of clients are iterating over the result,
so switch to a vector to ensure we don't
accidentally introduce any non-determinism.
(cherry picked from commit 56929fdaaa)
With an inverted pipeline, IRGen needs to be able
to compute the linker directives itself, so sink
it down such that it can be computed by the
`IRGenDescriptor`.
(cherry picked from commit c0a2ea7d0e)
