I am using nsight eclipse edition for mac to debug my cuda code. I was able to run cuda-gdb all fine from a terminal but was not able to use the debug function in Nsight. Whenever I started to debug, an error window pops up saying "Error in final launch sequence
Connection is shut down". My OS is Mac Lion (10.7.4).
I understand the big thing in Cuda 5.0 is the interactive debugging with eclipse but sadly I won't be able to use it unless I switch to Windows or Linux. If anybody has seen this or knows a workaround for this, please let me know and your help will be very much appreciated.
Thank you!
I'm afraid you've hit a known issue in CUDA 5.0 Toolkit. Due to some last-minute change, cuda-gdb (which is a shell script) is not entirely valid in Mac OS X. To remedy the problem:
Edit cuda-gdb shell script.
Make line #!/bin/sh the first line in this script.
Save the changes
We apologize for inconvenience. This (and many other) issues will be fixed in the upcoming CUDA 5.0 RC release.
I had this problem in linux recently.
Error in final launch sequence Connection is shut down which means nsight can not start debugger correctly.
first you have to make sure you can run cuda-gdb in terminal. just type cuda-gdb in terminal, if it says like cuda-gdb cannot find some lib in /usr/lib, just download or relink it. after you done this , rerun cuda-gdb in terminal, it should work fine. then you can use nsight debugging tool.
Related
we have created an assembler project for Cortex A9 that we would like to prepare the code for without the board. We, therefore, chose to use cortex-debug plugin and it seems that every 2-5 instructions, it takes over a second or two to get to the next step.
It seems that it happens during the "data-list-register-values" command issued to QEMU.
It works ok if I run same command from command line in GDB and also on Linux we didn't spot this problem. It happens only on windows and it seems only under VSCode.
Any advice or more info on this strange behaviour?
Thanks in advance,
regards.
I change my Hard drive and so I made a fresh Eclipse Corrosion install but I kept my Rust projets. Problem is I get an error when I launch my prog in debug mode. It seems to work in run mode but I didn't test this that much.
Error is
An internal error occurred during: "Launching my_prog".
Cannot read the array length because "launchConfigEnv" is null
Important note : it seems to be link with the use of an input file (debug configuration/common/input file)
I try many things, many installs, new projects, many, many things but nothing works (and I cannot figure out what launchConfigEnv is).
Any help welcome...
Edit : I install 2019.12 Eclipse and it works fine. I will work a little with this version and try out more configuration to shrink the problem.
As a workaround, add a single environment variable to the launch configuration, as follows:
Once that is done, re-launch it and it should work.
Note that in order to get rust debugging working in Eclipse I installed MSYS2 and the following MSYS2 packages.
pacman -S mingw-w64-x86_64-gcc
pacman -S mingw-w64-x86_64-gdb
pacman -S man
I followed some instructions in the Rust Corrosion Trouble Shooting guide to get a working GDB.
Reference: https://github.com/eclipse/corrosion/blob/master/documentation/Troubleshooting.md#debugging-on-windows
The Rust debugger command needs to be changed to rust-dbg.cmd, as follows:
However, the PATH environment wasn't working for me, so in the rust-gdb.cmd script I added this near the top:
SET PATH=%PATH%;F:\msys64\mingw64\bin;F:\Data\.cargo\bin
(Adjust as-per your installation of MSYS2 and Cargo/Rustup)
After which Eclipse was able to start a debug session and stop at a breakpoint.
I've been trying to install gdb to debug using eclipse on my mac with no luck, I followed both of those guides with no success:
https://www.youtube.com/watch?v=JdcpDOf5Sog
[second guide in other comment - can't add more than two links due to new user limitations]
I'm attaching a screenshot to demonstrate:
Gdb "terminated exit value: 0" eclipse on mac
Each time I hit the Debug button I get the same "treminated exit value: 0" error (the code runs perfectly when I hit the run button). Gdb does respond to terminal command "gdb"
I've been researching for quite long and it seems that no one has experienced the same problem I do. (Macbook pro 2016 10.12 Sierra)
I've heard from a friend that I might need to configure my Mac to "developer" mode, I have xcode up and running, but I have no idea what else should I do. Anyway I'm my friend could not guarantee that it is the cause of the problem.
I had issue with terminated exit value: 0 while using older gdb 7.11 because my 7.12.1 does not work on Sierra 10.12.4. There was no solution for my case.
Easy WA is to use LLDB as a debugger for now. Sample instructions here
Another LLDB instruction here
I'm following these instructions to remotely debug a python script http://pydev.org/manual_adv_remote_debugger.html
Using breakpoints and stepping work without a problem, but after resuming the script (without any other further breakpoints) the script executes the remaining instructions, and then hangs instead of exiting normally. Even closing the server from pydev does not help, I have to Ctrl+Break the script
I am sure I've done this a while ago with an older eclipse+pydev, and the script resumed and exited normally (and it's very important for me not to mess the rest of the execution).
I've tried this with both Eclipse Luna and Liclipse pydev_4.0.0.201504132356, pydev_4.0.0.201504092214 on Windows 7 x64, same behaviour
Is this an intended behavior or a bug? How could I make pydev let the script end normally?
I can't reproduce this here... please report that as a bug adding details on how to reproduce it (i.e.: ideally, create a project on github and add a step-by-step on how to reproduce it so that it can be fixed... without being able to reproduce it properly, it may be pretty hard to fix).
Also, please give details on the OS and Python version used in the report (https://sw-brainwy.rhcloud.com/tracker/PyDev/)
I have created an application in linux with GTK2 as GUI. It uses some linux-specific headers (e.g. arpa/inet.h) so to run under Windows I have to compile it with Cygwin. I downloaded the latest installer and choose to install GTK2 and its dependencies. My program compiled fine. But it needs X server to be running! I has old-style, ugly graphics and it doesn't open in a different window, like all Windows' applications do, but inside X server's window. Because of this it can't be portable. I found that guide, which is exactly what I need, but I get an error when I run "make" for GTK2 (undefined reference for _IID_IFilePersist, although I have uuid installed - also tried it with gtk2.20). Can you suggest what to do to build my application with cygwin? Or what do I need to install for the "_IID_IFilePersist" error? Thanks in advance!
There's prebuilt packages for windows that doesn't rely on X. http://gtk-win.sourceforge.net/home/index.php/Downloads
If you don't want X server to be running, then you're going to have to port the linux-specific parts of your code and compile with MinGW rather than Cygwin.