[symbolizer] Treat invalid address as addr2line does

[sepavloff]

Summary

llvm-symbolizer echoed input if it was not recognized as a valid address.
This behavior was extended to llvm-addr2line as well. GNU addr2line in
this case optputs "??:0". This difference prevents implementation of
symbol+offset lookup available in the recent versions of GNU binutils.
In that case a string that is not an address may be a symbol.

This change make reaction of llvm-addr2line on unrecognized input closer
to GNU addr2line.

[MaskRay]

Thanks. I have verified that GNU addr2line has the ??:0 behavior.

[sepavloff]

Thanks!

[jhenderson]

Mostly looks good to me, but please see inline comments.

llvm/test/tools/llvm-symbolizer/input-base.test

29–30 ↗

(On Diff #509213)

These checks should be deleted, right?

llvm/test/tools/llvm-symbolizer/invalid-input-address.test

25

If I'm not mistaken, the different behaviour is controlled by the --output-style option and llvm-addr2line just happens to have the right default value for this to work. Is it worth explicitly testing llvm-symbolizer --output-style=GNU for this case?

[sepavloff]

Add tests for llvm-symbolizer in GNU output mode

[sepavloff]

llvm/test/tools/llvm-symbolizer/input-base.test

29–30 ↗

(On Diff #509213)

These are used in test on line 11.

llvm/test/tools/llvm-symbolizer/invalid-input-address.test

25

Added such tests (lines 17-20).

[jhenderson]

Thanks for the new cases, but their output actually surprised me. I'd expect GNU output style to render invalid inputs as ??:0 (and equivalently, llvm-addr2line with non-GNU output to echo the input). I'm not too fussed by it, but I would have thought that this comes under "output style" really.

[sepavloff]

Committed in: https://reviews.llvm.org/rG27c4777f41d2.

llvm/test/tools/llvm-symbolizer/invalid-input-address.test

 # Use address that can't fit in a 64-bit number. Show that llvm-symbolizer
 # simply echoes it as per other malformed input addresses.
 RUN: llvm-symbolizer --obj=addr.exe 0x10000000000000000 | FileCheck --check-prefix=LARGE-ADDR %s
 
 LARGE-ADDR-NOT: {{.}}
 LARGE-ADDR:     0x10000000000000000
 LARGE-ADDR-NOT: {{.}}
 
 RUN: echo '"some text"' '"some text2"' > %t.rsp
 RUN: echo -e 'some text\nsome text2\n' > %t.inp
 
 # Test bad input address values, via stdin, command line and response file.
 RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe < %t.inp | FileCheck --check-prefix=BAD-INPUT %s
 RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe "some text" "some text2" | FileCheck --check-prefix=BAD-INPUT %s
 RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe @%t.rsp | FileCheck --check-prefix=BAD-INPUT %s
 
+# Test bad input in GNU-compatible output mode.
+RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe --output-style=GNU < %t.inp | FileCheck --check-prefix=GNU-BAD-INPUT %s
+RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe --output-style=GNU "some text" "some text2" | FileCheck --check-prefix=GNU-BAD-INPUT %s
+RUN: llvm-symbolizer --obj=%p/Inputs/addr.exe --output-style=GNU @%t.rsp | FileCheck --check-prefix=GNU-BAD-INPUT %s
+
 BAD-INPUT:      some text
 BAD-INPUT-NEXT: some text2
 
 # Test bad input address values for the GNU-compatible version.
 RUN: llvm-addr2line --obj=%p/Inputs/addr.exe < %t.inp | FileCheck --check-prefix=GNU-BAD-INPUT %s
 RUN: llvm-addr2line --obj=%p/Inputs/addr.exe "some text" "some text2" | FileCheck --check-prefix=GNU-BAD-INPUT %s
 RUN: llvm-addr2line --obj=%p/Inputs/addr.exe @%t.rsp | FileCheck --check-prefix=GNU-BAD-INPUT %s
 
+# Test bad input address values for the GNU-compatible version if output mode is LLVM.
+RUN: llvm-addr2line --obj=%p/Inputs/addr.exe --output-style=LLVM < %t.inp | FileCheck --check-prefix=BAD-INPUT %s
+RUN: llvm-addr2line --obj=%p/Inputs/addr.exe --output-style=LLVM "some text" "some text2" | FileCheck --check-prefix=BAD-INPUT %s
+RUN: llvm-addr2line --obj=%p/Inputs/addr.exe --output-style=LLVM @%t.rsp | FileCheck --check-prefix=BAD-INPUT %s
+
 GNU-BAD-INPUT:      ??:0
 GNU-BAD-INPUT-NEXT: ??:0