Forum Replies Created
-
Posts
-
Hi,
Normally VisualGDB should run the ‘chmod’ command automatically. Does it also happen with an empty “hello, world” project on the same board?
Hi,
This will be fixed in VisualGDB 5.0 Beta 2. As a workaround, use the Go To Definition (F12) command instead.There are 2 ways of achieving this:
An undocumented way that may be broken in the further versions of GCC would look like this:
extern const int nErasedFlag asm("0x0807F800");A classical “supported” approach would be similar to what CooCox version uses: put the variable into a separate section and modify the linker script to place this section at a given address.
You can read more about this here: http://sysprogs.com/w/forums/topic/bootloader/
Hi,
You can try further reducing the size of your binary by switching the libc to newlib-nano and experimenting with other libc options on the Device Settings page.
Also feel free to check out the new IntelliSense engine of VisualGDB 5.0. It is designed to handle GCC-specific code and supports refactoring and Code Map.
Most likely your daemon tries to launch some other executable before handling your command. VisualGDB “follows” that launch and ends the debug session when it exits.
Please try running the set follow-fork-mode command manually via the GDB Session window just before you expect your binary to run the child job.
You can also enable GDB logging in the Quick Debug window and examine the log for any additional clues.
Hi,
Please forward your activation key to our support so that we can deactivate it. Note that there is some problem with your current email address. We regularly get “delivery failed” reports from your email provider. Please contact us from an alternate email.
Hi,
Currently it is not possible to create the project without a board directly, however you can copy the .vgdbsettings file and the Makefile (with all .mak files) from an existing project. This will transfer your settings to the new project. Please also note that we will be adding support for creating projects without connection to a board in the final release of VisualGDB 5.0.
Hi,
Thanks for reporting this. Looks like there is a bug in our semihosting handler that triggers when used with the texane/ST-Link tool. Please try the following build: http://visualgdb.com/tmp/VisualGDB-5.0.5.355.msi
Hi,
We have looked into it and will be getting some ESP8266 boards to investigate the complexity of adding a BSP for it. Feel free to follow us on Twitter (@Sysprogs) to stay updated.
Update: here you go.
Hi,
The clang version we use has the c++1y, not c++14 syntax. We will add an alias to the final release, though.
The Include directories and Linker script are managed by GNU Make and do not have access to VisualGDB variables. You can simply use relative paths to specify something relatively to the project directory or create a per-user variable (Custom edition and higher) and select to propagate it to GNU Make.
This is true. Can you inspect the contents of View->Output->VisualGDB Launcher Output after an unsuccessful debugging session? It should normally show something like this:
VisualGDB: Executing predebug actions VisualGDB: Deploying e:\PROJECTS\TEMP\test\test/Debug/test to /tmp/vgtest/test on raspberrypi... VisualGDB: Launching gdbserver VisualGDB: Run "gdbserver :2000 "$(DeployPath)" " in directory "$(DeployDir)" on pi@raspberrypi (SSH) VisualGDB: Launching custom program console VisualGDB: Launching gdb VisualGDB: Executing postdebug actions
Then check that:
1) The file on the Windows machine exists and is readable under your account
2) The remote file does not exist, the remote directory does and the device has enough free space
If none of that provides clues, can you reproduce it with a “Hello, World” project? If yes, can you create an image of your SD card and send it to us so that we could reproduce it on our side?
Please run ‘dmesg | tail’ to see debug messages from the kernel. Are there any details regarding kgdboe issues? Have you tried following the manual kgdboe tutorial?
