[Lldb-commits] [lldb] 3e03873 - [lldb][Test] Fix TestFrameFormatNameWithArgs.test on Windows/Linux

[Zequan Wu via lldb-commits]

That's a known issue for the NativePDB plugin:
https://github.com/llvm/llvm-project/issues/51933. Because currently the
plugin internally looks up function names by full names. Even if you try to
set a breakpoint with a full function name, the plugin itself only receives
the basename.

On Wed, Nov 2, 2022 at 7:03 AM Michael Buch <michaelbuch12 at gmail.com> wrote:

> > For example the type `const char *` also has a
> summary provider, and I'd hope that this feature does not depend on the
> specific way in which that summary is computed.
>
> Yea, it would be nice to have this on Linux. Will try to use a type whose
> format is compatible across platforms
>
> Am Mi., 2. Nov. 2022 um 06:54 Uhr schrieb Pavel Labath <pavel at labath.sk>:
>
>> On 01/11/2022 06:59, Michael Buch via lldb-commits wrote:
>> >
>> > Author: Michael Buch
>> > Date: 2022-10-31T22:59:16-07:00
>> > New Revision: 3e03873e363b5aa90e4488da63a6de0648d11aba
>> >
>> > URL:
>> https://github.com/llvm/llvm-project/commit/3e03873e363b5aa90e4488da63a6de0648d11aba
>> > DIFF:
>> https://github.com/llvm/llvm-project/commit/3e03873e363b5aa90e4488da63a6de0648d11aba.diff
>> >
>> > LOG: [lldb][Test] Fix TestFrameFormatNameWithArgs.test on Windows/Linux
>> >
>> > * Windows doesn't support setting these breakpoints by basename
>> > * On Linux std::function arguments aren't formatted as such
>>
>> The windows issue seems problematic (Zequan might be interested in
>> that), but the non-pretty-printing of std::function is most likely
>> caused by our lack of a pretty-printer for libstdc++'s (default for
>> linux) std::function. It doesn't seem ideal that a fairly generic test
>> would depend on the existence of a (very complicated) pretty printer.
>>
>> Is there a specific edge case that this particular check was trying to
>> hit? Could it be moved into a separate test case, or ideally replaced by
>> a something simpler? For example the type `const char *` also has a
>> summary provider, and I'd hope that this feature does not depend on the
>> specific way in which that summary is computed.
>>
>> >
>> > Added:
>> >
>> >
>> > Modified:
>> >      lldb/test/Shell/Settings/TestFrameFormatNameWithArgs.test
>> >
>> > Removed:
>> >
>> >
>> >
>> >
>> ################################################################################
>> > diff  --git a/lldb/test/Shell/Settings/TestFrameFormatNameWithArgs.test
>> b/lldb/test/Shell/Settings/TestFrameFormatNameWithArgs.test
>> > index ab16c656624d8..d990114f57845 100644
>> > --- a/lldb/test/Shell/Settings/TestFrameFormatNameWithArgs.test
>> > +++ b/lldb/test/Shell/Settings/TestFrameFormatNameWithArgs.test
>> > @@ -1,3 +1,4 @@
>> > +# REQUIRES: system-darwin
>> >   # RUN: %clangxx_host -g -O0 %S/Inputs/names.cpp -std=c++17 -o %t.out
>> >   # RUN: %lldb -b -s %s %t.out | FileCheck %s
>> >   settings set -f frame-format "frame ${function.name-with-args}\n"
>> >
>> >
>> >
>> > _______________________________________________
>> > lldb-commits mailing list
>> > lldb-commits at lists.llvm.org
>> > https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20221102/b2d4a31f/attachment-0001.html>