-
Notifications
You must be signed in to change notification settings - Fork 1.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
LLDB debugger hangs (using 100% cpu) at launch if std::map is uninitialised #5805
Comments
@zeeepw Thanks for the very detailed issue. I will need to validate if this occurs in |
Hello, has there been any progress pls? |
I'm having what looks like the same problem, on a big codebase (Apache Arrow). As soon as I get to a stack frame that contains more complex objects (that use STL collections in their implementation, the debugger freezes and does not respond to commands. Looking at things under the hood, VSCode is waiting forever on a reply to the following: This is making C++ debugging completely unusable for me. |
I've been experiencing this issue when debugging C++ program in macOS. When this happens, I have to turn to Xcode which works perfectly. So I would have to guess lldb works fine and it is a lldb-mi specific issue.
|
Can confirm this is still happening on: |
HI Is there any update on this ? I am writing a simple C++ program for maps and not able to debug at all!! |
I just can not believe this issue still hasn't been addressed. I've been running into this for more than one year. And my workaround is to not use lldb-mi at all. |
just ran into it. Cannot believe it is still unresolved |
I've got this same issue. |
It happens to me a lot. I restart my MacBook and sometimes it fixes it. Other times it does not help at all. |
Happens to me with unordered_map, sometimes even if it's initialized. I just need an IDE with a debugger and a CMake extension for OSX. Any workarounds or recommendations? |
Can confirm this is still happening on: When I uncomment the declaration of multi_map (example C) and set a breakpoint on the first line of main() or have "stopAtEntry": true then lldb-mi goes to high CPU and I have to kill it and the debugged process manually. I could also reproduce the high cpu hang without a breakpoint on the first line of main() if I attached to the process and put the code in a function called after a sleep with a breakpoint at the start of the function. |
I also ran into a similar issue. |
Same here |
I have a similar problem and I noticed the local section in the variables window in the Visual Code debugger has a spinner during the hang. I hid the variables window and the debugger no longer hangs. Hopefully this works for someone else because this was driving me insane. |
@csloan-gmail It works, thanks. Not a solution but this allows to actually debug stuff, at least. |
@csloan-gmail Thanks! That works for me if in addition I'm also super careful with my mouse to not accidentally hover over a map. |
Thanks @csloan-gmail same problem here. Workaround (not looking at variables) at least allows me to move forward. |
Thanks, it solved the issues. |
Have the same problem on MacOS, hiding variable window is a partial workaround for now. |
I have the same experience both on LLDB and on Perl Debugger, using VSCode on Mac, so maybe the problem is not in LLDB but in VSCode?
Cheers
FC
… On 13 Jan 2023, at 12:19, Bas du Pré ***@***.***> wrote:
Have the same problem on MacOS, hiding variable window is a partial workaround for now.
—
Reply to this email directly, view it on GitHub <#5805 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AIAC4OXZV7VOQR65BCZPUPDWSFB4FANCNFSM4PBTBPLA>.
You are receiving this because you commented.
|
The CodeLLDB extension works without hanging (or hogging the CPU, or using all the memory) on Apple M1 with the variables window open, so it can be done. |
This issue has been closed as lower priority. We're sorry if this issue still impacts you but unfortunately we're not able to address this. We will accept a pull request from the community if it's applicable for this issue. |
Same here. I can't debug at all if my HandleMouseEvent(SDL_mouse) ) is in the call stack. Works perfectly fine everywhere else. Using Kubuntu 23.04, CodeLLDB |
this is still happening! |
1 similar comment
this is still happening! |
Type: Debugger
Describe the bug
Similar log to #860
When using
include <map>
andstd::map
/std::multimap
without initialisation the debugger hangs at launch and is stuck at-var-create - - "map_variable_name" --thread 1 --frame 0
; seems like no garbage value can be createdSee example A below
However this does not happen, if before the (one)
std::map
declaration there is anothercontainer
such asstd::vector
or
std::unordered_map
declared, even uninitialised.See example B below
And the problem re-surfaces if more than two
<map> contrainers
are used.See example C below
Therefore this seems to strangely happen only to std::map containers; a few other
stl containers
andstd::vector
alone uninitialised can launch fine and be seen in the local variables windowPlease could you investigate? many thanks in advance
To Reproduce
Steps to reproduce the behavior:
EXAMPLE A
EXAMPLE B
EXAMPLE C
- stuck atmulti_map
Full log for example A
log.txt
cpu usage when hanging
The text was updated successfully, but these errors were encountered: