Quuxplusone / LLVMBugzillaTest

0 stars 0 forks source link

Assertion failed: (OldTInfo && "substituting function without type source info") #34893

Open Quuxplusone opened 6 years ago

Quuxplusone commented 6 years ago
Bugzilla Link PR35920
Status NEW
Importance P normal
Reported by Vedant Kumar (vsk@apple.com)
Reported on 2018-01-11 17:18:18 -0800
Last modified on 2018-01-11 17:24:31 -0800
Version unspecified
Hardware PC All
CC aprantl@apple.com, ditaliano@apple.com, llvm-bugs@lists.llvm.org, zturner@google.com
Fixed by commit(s)
Attachments
Blocks
Blocked by
See also
Bot failure: http://lab.llvm.org:8080/green/view/LLDB/job/lldb-xcode/4101/

Command invoked: test/dotest.py --executable
/Users/buildslave/jenkins/workspace/lldb-xcode/lldb/build/Release/lldb -C
/Users/buildslave/jenkins/workspace/lldb-xcode/host-compiler/bin/clang --arch
x86_64 --session-file-format fm --rerun-all-issues --env TERM=vt100 -s 2018-01-
11-09_34_09 --results-port 59606 -S fm --inferior -p
TestFunctionTemplateParameterPack.py /Users/buildslave/jenkins/workspace/lldb-
xcode/lldb/packages/Python/lldbsuite/test --event-add-entries worker_index=1:int

Configuration: arch=x86_64 compiler=/Users/buildslave/jenkins/workspace/lldb-
xcode/host-compiler/bin/clang-7.0
----------------------------------------------------------------------
Collected 3 tests

Assertion failed: (OldTInfo && "substituting function without type source
info"), function SubstFunctionType, file
/Users/buildslave/jenkins/workspace/lldb-
xcode/lldb/llvm/tools/clang/lib/Sema/SemaTemplateInstantiateDecl.cpp, line 3418.

[TestFunctionTemplateParameterPack.py FAILED]
Quuxplusone commented 6 years ago

This assertion failure has been biting us sporadically for weeks (possibly longer).

I'm planning on marking this test skipped (it's already XFAIL'd, but assertion failures are unexpected failures).

I'm also prototyping a change to the DWARF AST reconstructor which errors out when we can't create a well-formed FunctionTemplateDecl. We'll see how that goes: https://github.com/apple/swift-lldb/pull/276

Quuxplusone commented 6 years ago

FWIW, this seems to reproduce almost deterministically on the swift-lldb upstream-with-swift branch, at least for me. It might be the case for you as well.

Quuxplusone commented 6 years ago

Marked the test skipped in r322340. @davide yes this reproduces deterministically for me locally.