[ClangImporter] Deduplicate clang buffers representing the same file #84904
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hnrklssn
requested review from
DougGregor,
Xazax-hun,
beccadax,
egorzhdan,
fahadnayyar,
ian-twilightcoder,
j-hui,
susmonteiro and
zoecarver
as code owners
|
@swift-ci please smoke test |
Xazax-hun
approved these changes
Oct 15, 2025
egorzhdan
approved these changes
Oct 15, 2025
j-hui
approved these changes
Oct 15, 2025
lib/Basic/SourceLoc.cpp
Outdated
| SourceManager::addNewSourceBuffer(std::unique_ptr<llvm::MemoryBuffer> Buffer) { | ||
| assert(Buffer); | ||
| StringRef BufIdentifier = Buffer->getBufferIdentifier(); | ||
| assert(BufIdentIDMap.find(BufIdentifier) == BufIdentIDMap.end()); |
There was a problem hiding this comment.
nit:
Suggested change
| assert(BufIdentIDMap.find(BufIdentifier) == BufIdentIDMap.end()); | |
| CONDITIONAL_ASSERT(BufIdentIDMap.find(BufIdentifier) == BufIdentIDMap.end()); |
There was a problem hiding this comment.
I removed this assert in the end. It was a bit optimistic, and the complete fix is a bit out of scope for this PR.
beccadax
approved these changes
Oct 15, 2025
Clang can end up with multiple file IDs representing the same file, but in different contexts. For example, file in the current module always have file IDs >0, while files in imported modules always have file IDs <-1. As we end up compiling multiple modules, the same file can be seen both as the "current" module, and a transitive import. We don't want multiple Swift buffer IDs for the same file, as it breaks things like -verify that compare buffer IDs. rdar://162661286
hnrklssn
force-pushed
the
dedup-clang-buffers
branch
from
c3fc710 to
696c08e
Compare
|
Removed the assert I added in |
|
@swift-ci please smoke test |
Virtual files from different modules can differ despite having identical names (e.g. `<module-imports>`. This fixes some diagnostics related test failures when these virtual files were (incorrectly) deduplicated and the module imports from a previous module were shown, by checking if the buffer has a corresponding clang::FileEntry.
|
@swift-ci please smoke test |
This would trigger when path prefixes are remapped, as the buffer identifier is the remapped path, not the path on the file system. There is no easy way to recover the original path, so drop the assert.
|
@swift-ci please smoke test |
|
@swift-ci please test macos platform |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Clang can end up with multiple file IDs representing the same file, but in different contexts. For example, files in the current module always have file IDs >0, while files in imported modules always have file IDs <-1. As we end up compiling multiple modules, the same file can be seen both as the "current" module, and a transitive import. We don't want multiple Swift buffer IDs for the same file, as it breaks things like -verify that compare buffer IDs.
rdar://162661286