Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Unit tests: pass. 61175 tests passed, 0 failed and 729 were skipped.
clang-tidy: fail. Please fix clang-tidy findings.
clang-format: fail. Please format your changes with clang-format by running git-clang-format HEAD^ or applying this patch.
Build artifacts: diff.json, clang-tidy.txt, clang-format.patch, CMakeCache.txt, console-log.txt, test-results.xml
llvm/utils/gdb-scripts/prettyprinters.py | ||
---|---|---|
322 ↗ | (On Diff #236535) | Note: this change is from https://reviews.llvm.org/D72136, which I will merge first. |
Unit tests: pass. 61175 tests passed, 0 failed and 729 were skipped.
clang-tidy: fail. Please fix clang-tidy findings.
clang-format: fail. Please format your changes with clang-format by running git-clang-format HEAD^ or applying this patch.
Build artifacts: diff.json, clang-tidy.txt, clang-format.patch, CMakeCache.txt, console-log.txt, test-results.xml
This is cute. We should do something similar with the LLDB pretty printers.
- What do you think about putting this into an llvm subdirectory, to make clear that this is a test of llvm/utils rather than generic debug info as all the other tests?
- We are gradually moving away from test_debuginfo.pl in favor of dexter. Have you looked at whether this could be implemented using dexter?
- What do you think about putting this into an llvm subdirectory, to make clear that this is a test of llvm/utils rather than generic debug info as all the other tests?
Do you mean something like debuginfo-tests/llvm-prettyprinters/? Or 'llvm/tests/prettyprinter'?
- We are gradually moving away from test_debuginfo.pl in favor of dexter. Have you looked at whether this could be implemented using dexter?
I didn't look any further when I saw that dexter doesn't support GDB (according to the readme).
Not sure if we need to use a debugger abstraction because the pretty printers aren't standardized and therefore the tests will be pretty debugger specific.
If you would like to move away from test_debuginfo.pl, maybe I should split the C++ code and gdb commands+CHECKS into two different files manually. It's not perfect FileCheck methodology, but for a single test in a subdirectory it should be manageable.
What about debuginfo-tests/llvm-prettyprinters/gdb/ and then we'll create an lldb subdirectory next to later?
I didn't look any further when I saw that dexter doesn't support GDB (according to the readme).
CC @jmorse
I forgot about that. I believe there exists a binding but it hasn't been upstreamed yet.
Unit tests: pass. 61745 tests passed, 0 failed and 780 were skipped.
clang-tidy: fail. Please fix clang-tidy findings.
clang-format: fail. Please format your changes with clang-format by running git-clang-format HEAD^ or applying this patch.
Build artifacts: diff.json, clang-tidy.txt, clang-format.patch, CMakeCache.txt, console-log.txt, test-results.xml
Unit tests: pass. 61745 tests passed, 0 failed and 780 were skipped.
clang-tidy: fail. Please fix clang-tidy findings.
clang-format: fail. Please format your changes with clang-format by running git-clang-format HEAD^ or applying this patch.
Build artifacts: diff.json, clang-tidy.txt, clang-format.patch, CMakeCache.txt, console-log.txt, test-results.xml
(Heading through my email backlog),
Adrian wrote:
We are gradually moving away from test_debuginfo.pl in favor of dexter. Have you looked at whether this could be implemented using dexter?
Christian wrote:
Not sure if we need to use a debugger abstraction because the pretty printers aren't standardized and therefore the tests will be pretty debugger specific.
I agree with Christian, while Dexter is good for ensuring that information is read accurately by different DWARF/CodeView consumers, how that's presented to the user is open to choices in the debugger application. IMHO, YMMV.
debuginfo-tests/CMakeLists.txt | ||
---|---|---|
5–9 | Might need to change the way this is built, @tbosch identified a bug where this test fails on a non-debug build of LLVM, because it doesn't explicitly request debug info, so it gets the build defaults. (either that, or the test should detect this situation - use a REQUIRES or similar to not run in the absence of debug info) |
debuginfo-tests/CMakeLists.txt | ||
---|---|---|
5–9 | https://reviews.llvm.org/D79897 should fix it. |
Might need to change the way this is built, @tbosch identified a bug where this test fails on a non-debug build of LLVM, because it doesn't explicitly request debug info, so it gets the build defaults. (either that, or the test should detect this situation - use a REQUIRES or similar to not run in the absence of debug info)
https://bugs.llvm.org/show_bug.cgi?id=45652