- User Since
- Sep 4 2014, 1:00 AM (291 w, 5 d)
Remove the now useless code in TypeSummary.cpp, and change the C++
implementation of IsNilReference() to return true soemtimes. This
breaks other tests though...
This looks like what I'd like to implement, but unfortunately it breaks other tests. Some C tests start printing null pointers as nullptr too. I suppose this is because the expression evaluator is always in C++ mode. Is there a way to get the origin type/language of a variable through a ValueObject?
Implement null C++ printing at the ValueObjectPrinter level
Wed, Apr 1
I was checking whether there is a way to catch null pointer before a type summary is even chosen. And found out that such logic exists, but only for Objective-C. Languages have a IsNilReference() virtual method that can tell whether a ValueObject is a nil reference. It's only implemented for ObjectiveC and value objects that test positive for this method are displayed as "nil" is the generic code of ValueObjectPrinter.cpp. I can see a few options:
- I can rework the patch to affect only the libc++ string formatter which was the issue I was trying to solve in the first place
- I can implement IsNilReference() for C++ and change ValueObjectPrinter to print C++ "nil references" as "nullptr".
- I do the above and add another customization point in the Language plugins which tells ValueObjectPrinter out to display "nil" references, so that there's no Language specific code in ValueObjectPrinter.
Tue, Mar 31
I haven't tested the libstdc++ part, would be great if someone could confirm this works.
I agree host shell command should run with the host environment.
Fri, Mar 27
I think this won't work with Python3, were out and err will be already "decoded" str objects
Thu, Mar 26
Thanks for fixing this Pavel!
Mon, Mar 23
Fri, Mar 20
Simplify command definition
Add a new target command to dump the launch environment
Make env-vars override unset-env-vars
Make helper function name more descriptive.
Add tests and update m_launch_info when undet-vars change.
Thu, Mar 19
This needs some tests, but I wanted to put it out there for comments.
I skipped the test for embedded in 8758d02074be7b80b804fad19e8b7de6ebd43c31. I'll abandon this for now
Wed, Mar 18
Tue, Mar 17
Sun, Mar 15
Sat, Mar 14
Thu, Mar 12
This is somewhat useless without D76009 because the default LaunchInfo doesn't get populated with the environment without it. I'll wait for a resolution there before landing this.
I put up a "fix" for the inherit-env issue mentioned here: https://reviews.llvm.org/D76105
It is mostly orthogonal to this patch as Jim showed the behavior today is already broken.
Wed, Mar 11
Remove unrelated change commited by accident
Mar 5 2020
Mar 2 2020
Feb 26 2020
Feb 7 2020
I think we're on the same page then.
Remove the fallback code and assert that StopInfo returns a non-empty string.
Feb 6 2020
Feb 5 2020
Jan 28 2020
This generally looks good. I'm still not fond of registering this in the Target itself. But I don't have a immediately better idea as we don't have a C language runtime.
Jan 24 2020
Jan 23 2020
Jan 21 2020
Jan 20 2020
use python formatting in docs
Jan 17 2020
Jan 3 2020
Jan 2 2020
Dec 17 2019
Nov 21 2019
Nov 19 2019
Sorry that I haven't reviewed the patch, but there's something I'd like to point out before anyone invests a lot of time into plugin holes in our current template support code.