Mar 26 2020
Lit option --timeout partly solves the problem, so these changes are unnecessary.
Mar 20 2020
Mar 17 2020
Mar 16 2020
Mar 12 2020
platform.module-cache-directory should be fixed by rGfe74df01a909.
Thank you for details, I'm looking at these failures, however, I'm not able to debug on the macOS.
Mar 5 2020
Removed TestFixIts.py as those changes relate to the parent revision.
Mar 4 2020
Added --all property to settings clear + added a test
Mar 3 2020
Moved settings clearing to the separate patch D75537
Clear all settings during a test's setUp
As Pavel wrote, there are global properties that all debuggers share. That's why this approach doesn't work for me.
I'm going to add settings clear mode without arguments that will revert all the settings to their default values.
@jingham Is it Ok with you?
Feb 21 2020
I like this idea and ready to implement, but I'm not aware of the reasons why the current implementation was chosen. Suppose that initializing the Debugger for each test will slow down the test-suite significantly.
Feb 20 2020
Feb 18 2020
Closed by commit b624b7dfd087
Closed by commit 185ef697ef5c
Feb 14 2020
Just realized that BreakpointResolverScripted::CreateImplementationIfNeeded should work when m_breakpoint == nullptr.
GetBreakpoint with assertion is not suitable here, so I pass the breakpoint as a parameter.
Moved assertion into GetBreakpoint().
Feb 13 2020
Feb 10 2020
Jan 10 2020
Jan 9 2020
Stop iterating when reach the top-level test folder
Jan 6 2020
Increased number of lines of context.
Dec 26 2019
Dec 9 2019
Dec 5 2019
Nov 21 2019
Nov 12 2019
Nov 1 2019
I mean, when a process is not able to allocate memory, it cannot JIT, but still can interpret. Should the test report failure in this case?
Another problem of the test_ir_interpreter is that it fails even when JIT fails despite it should test the interpreter.
Oct 17 2019
Rebased on the current trunk.
Removed any ARC-specific logic from the ProcessGDBRemote.cpp.
Oct 15 2019
Build and installation completed successfully! LGTM, though it would be good if anyone tests this with Xcode.
Oct 14 2019
@hhb, please, take a look. Slightly changed your patch to make it work for Visual Studio.
Oct 12 2019
Oct 11 2019
Fixed by reverting the initial commit.
Oct 10 2019
Please see my comment D68719#1703785.
Oct 9 2019
Oct 8 2019
I have longed for this change, thank you!
Sep 26 2019
Sep 25 2019
Removed using UnixSignals::ShouldSuppress, just check if a signal is either SIGINT or SIGSTOP.
To Pavel: this problem is not only about the behavior of process attach, but for process launch too (and possibly something else). That's why I'm trying to move this logic out of specific commands.