Re: [lldb-dev] LLDB Future Directions BoF

2018-01-23 Thread Greg Clayton via lldb-dev 
I will work with the main contributors to LLDB in the near future and get 
something up on the http://lldb.llvm.org  web site that 
addresses these issues. Once all main contributors have signed off we will make 
the web page live for the public to find. This is a long overdue page that 
needs to be created. It will allow us to modify the document as times goes on 
and also provide the insights as to why I decided to write LLDB in the first 
place, where we are going and who contributes to which parts, and what parts of 
LLDB need work. So stay tuned and I will reply to this message when the web 
page goes up, hopefully in the near future.

Greg Clayton

> On Jan 22, 2018, at 5:40 PM, André Jansen Medeiros Villar via lldb-dev 
>  wrote:
> 
> Hi all,
> 
> I'm interested in understanding the why of LLDB, for that i'd love to be able 
> to read any meeting notes or slides of the BoF on the future of LLDB and if 
> possible make it available at [1] .
> 
> I'd also would like any indication of reading materials on the internals of a 
> debugger, preferabbly including something for begginers as the excelent 
> chapther for LLVM on the The Architecture of Open Source Applications book 
> [2].
> 
> And finally I'd ask why you chose to create LLDB when GDB was already there, 
> the idea of reusing LLVM components (the c++ parser and AST ?) was proven 
> benefical? Over the months of reading this list I've seen that there's 
> significant attrition on what LLVM and LLDB require of their components and 
> there are several cases of copying entire files with diverging changes 
> between each project, is there a simply reason on why or is it a matter of 
> needing more time to conciliate these disparate needs ?
> 
> Thx for any and all responses.
> 
> [1] http://llvm.org/devmtg/2017-10/#bof7 
> 
> [2] http://www.aosabook.org/en/llvm.html 
> ___
> lldb-dev mailing list
> lldb-dev@lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev

___
lldb-dev mailing list
lldb-dev@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev

[lldb-dev] [Bug 36048] Several C++ data formatter tests fail with -fmodules

2018-01-23 Thread via lldb-dev 
https://bugs.llvm.org/show_bug.cgi?id=36048

Adrian Prantl  changed:

   What|Removed |Added

   Assignee|lldb-dev@lists.llvm.org |apra...@apple.com
 Status|NEW |ASSIGNED

--- Comment #3 from Adrian Prantl  ---
Not sure, I'll have to investigate. Chances are 50/50 that this is a bug in
clang or in LLDB. The last -gmodules-related bug I fixed turned out to be a bug
in both :-)

-- 
You are receiving this mail because:
You are the assignee for the bug.___
lldb-dev mailing list
lldb-dev@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev

Re: [lldb-dev] [Release-testers] [llvm-dev] [6.0.0 Release] Release Candidate 1 tagged

2018-01-23 Thread Dimitry Andric via lldb-dev 
On 20 Jan 2018, at 13:32, Dimitry Andric via Release-testers 
 wrote:
> 
> On 19 Jan 2018, at 17:11, Hans Wennborg  wrote:
>> 
>> On Thu, Jan 18, 2018 at 7:27 PM, Dimitry Andric  wrote:
>>> On 18 Jan 2018, at 15:03, Jonas Hahnfeld  wrote:
 
 Am 2018-01-18 14:55, schrieb Dimitry Andric via llvm-dev:
> On 17 Jan 2018, at 18:53, Hans Wennborg via Release-testers
>  wrote:
>> Start your engines; 6.0.0-rc1 was just tagged.
>> I know there are still open blockers and it's early in the process in
>> a way, but I'd like to find out where we are. Please run the test
>> script, let me know the results, and upload binaries.
...

So, I finally managed to produce some tarballs, and uploaded them:

SHA256 (clang+llvm-6.0.0-rc1-amd64-unknown-freebsd10.tar.xz) = 
d93427fd2b8f5aa0d5278f1bd3020add07b8316ff8a512a203bf3c41639d7baf
SHA256 (clang+llvm-6.0.0-rc1-i386-unknown-freebsd10.tar.xz) = 
9f283235fb10242b9f79527c0d895260e6efa1c1bb761c57396490c7ccd5f5f0

This was an interesting round of testing, since I attempted to build the 'libs' 
part for the first time, and things like libcxx, polly etc are also included.  
I had to apply several patches, one of which was to be able to disable 
libc++abi, which we can't link against yet on FreeBSD.

During the build of the test-suite, I encountered several instances of link 
jobs failing due to the addition of -ldl, which does not exist on FreeBSD.  It 
is really a Linuxism, for which I will commit a few fixes.  In addition, some 
test-suite programs fail to link on FreeBSD 10.x, since it does not yet have 
__cpu_model in the base system's copy of libcompiler-rt.  I solved that with a 
local hack.

On amd64-freebsd10 there were 895 test failures:


Testing Time: 9904.87s

[...]
  Expected Passes: 44958
  Expected Failures  : 185
  Unsupported Tests  : 2938
  Unexpected Failures: 895

On i386-freebsd10 there were 1773 test failures:


Testing Time: 23871.89s

[...]
  Expected Passes: 42648
  Expected Failures  : 194
  Unsupported Tests  : 1953
  Unexpected Failures: 1773

Of these, a great many were in libc++, mostly due to 
-Wtautological-type-limit-compare errors, and "exception_ptr not yet 
implemented" errors.  Also all the dynamic address sanitizer tests failed due 
to -lpthread not being passed to the linker, but it is a problem I have not 
been able to solve.

If anybody is interested in the full error reports, I can upload them 
somewhere; just let me know.

-Dimitry



signature.asc
Description: Message signed with OpenPGP
___
lldb-dev mailing list
lldb-dev@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev