<div dir="ltr">Yea that part is clear, I just wonder what it's trying to parse. It sounds like an actual PDB file, because otherwise it shouldn't have made it that far in the code. But if it's an actual PDB file, it shouldn't be failing.<div><br></div><div>Anyway, the code definitely has a bug in that it doesn't correctly handle failure, it's just strange that it's failing at all. I'm about to check in a fix to better handle failure.</div></div><br><div class="gmail_quote"><div dir="ltr">On Fri, Oct 12, 2018 at 3:51 PM Jason Molenda <<a href="mailto:jmolenda@apple.com">jmolenda@apple.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Adrian suggests that parseFileHeaders is returning an error and that needs to be cleared?<br>
<br>
<br>
> On Oct 12, 2018, at 3:27 PM, Jason Molenda <<a href="mailto:jmolenda@apple.com" target="_blank">jmolenda@apple.com</a>> wrote:<br>
> <br>
> Ah, mystery solved. The lit test specifies REQUIRES Darwin. <br>
> <br>
>> On Oct 12, 2018, at 3:22 PM, Jason Molenda <<a href="mailto:jmolenda@apple.com" target="_blank">jmolenda@apple.com</a>> wrote:<br>
>> <br>
>> <br>
>> <br>
>>> On Oct 12, 2018, at 12:47 PM, Zachary Turner via lldb-commits <<a href="mailto:lldb-commits@lists.llvm.org" target="_blank">lldb-commits@lists.llvm.org</a>> wrote:<br>
>>> <br>
>>> Resubmit "Add SymbolFileNativePDB plugin."<br>
>>> <br>
>>> This was originally reverted due to some test failures on<br>
>>> Linux. Those problems turned out to require several additional<br>
>>> patches to lld and clang in order to fix, which have since been<br>
>>> submitted. This patch is resubmitted unchanged. All tests now<br>
>>> pass on both Linux and Windows.<br>
>> <br>
>> <br>
>> The lit test in lit/Modules/lc_build_version.yaml fails on Darwin systems, crashing here -<br>
>> <br>
>> 4 lldb-test 0x00000001008ebdab llvm::Error::fatalUncheckedError() const + 139<br>
>> 5 lldb-test 0x000000010084b307 llvm::Error::assertIsChecked() + 87 (Error.h:270)<br>
>> 6 lldb-test 0x000000010084b279 llvm::Error::~Error() + 25 (Error.h:230)<br>
>> 7 lldb-test 0x0000000100844725 llvm::Error::~Error() + 21 (Error.h:231)<br>
>> 8 lldb-test 0x0000000101b91e49 loadPDBFile(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, llvm::BumpPtrAllocatorImpl<llvm::MallocAllocator, 4096ul, 4096ul>&) + 1193 (SymbolFileNativePDB.cpp:73)<br>
>> 9 lldb-test 0x0000000101b9024e lldb_private::npdb::SymbolFileNativePDB::CalculateAbilities() + 542 (SymbolFileNativePDB.cpp:187)<br>
>> 10 lldb-test 0x0000000101060ceb lldb_private::SymbolFile::GetAbilities() + 43 (SymbolFile.h:87)<br>
>> 11 lldb-test 0x00000001010607fc lldb_private::SymbolFile::FindPlugin(lldb_private::ObjectFile*) + 764 (SymbolFile.cpp:58)<br>
>> <br>
>> <br>
>> Where the loadPDBFile line 73 is <br>
>> <br>
>> <br>
>> 59 static std::unique_ptr<PDBFile> loadPDBFile(std::string PdbPath,<br>
>> 60 llvm::BumpPtrAllocator &Allocator) {<br>
>> 61 llvm::ErrorOr<std::unique_ptr<llvm::MemoryBuffer>> ErrorOrBuffer =<br>
>> 62 llvm::MemoryBuffer::getFile(PdbPath, /*FileSize=*/-1,<br>
>> 63 /*RequiresNullTerminator=*/false);<br>
>> 64 if (!ErrorOrBuffer)<br>
>> 65 return nullptr;<br>
>> 66 std::unique_ptr<llvm::MemoryBuffer> Buffer = std::move(*ErrorOrBuffer);<br>
>> 67 <br>
>> 68 llvm::StringRef Path = Buffer->getBufferIdentifier();<br>
>> 69 auto Stream = llvm::make_unique<llvm::MemoryBufferByteStream>(<br>
>> 70 std::move(Buffer), llvm::support::little);<br>
>> 71 <br>
>> 72 auto File = llvm::make_unique<PDBFile>(Path, std::move(Stream), Allocator);<br>
>> 73 if (auto EC = File->parseFileHeaders())<br>
>> 74 return nullptr;<br>
>> 75 if (auto EC = File->parseStreamData())<br>
>> <br>
>> <br>
>> I'm not sure why this wouldn't fail on a linux/windows system. The point of the .yaml lit tests was that it is host independent?<br>
>> <br>
>> <br>
>> J<br>
> <br>
<br>
</blockquote></div>