xiaoxiang781216 commented on issue #1407: URL: https://github.com/apache/incubator-nuttx/issues/1407#issuecomment-659823621
> Even though we merged the related PR, I'm actually a bit torn about this. > First, it's only the output that's redirected. An error should still be printed if configure.sh or sethost.sh fail. V=1 not only removes the redirection but also enables make debugging. Yes, the error still output but without any context information(one line error message). Since the build system inovke make recurisively into the subfolder, you can't even locate which Makefile hit the error. Here is an example @hartmannathan hit: https://groups.google.com/forum/#!searchin/nuttx/configure.sh|sort:date/nuttx/R72PpM_8alQ/oWPaNvaZEwAJ tools/configure.sh? I'm getting a make do_olddefconfig error: ``` $ tools/configure.sh -l tm4c1294-launchpad/nsh Copy files Select CONFIG_HOST_LINUX=y Refreshing... make: *** [do_olddefconfig] Error 1 Has anyone else experienced this? I'm continuing to investigate... ``` Nobody can locate the location with so little information. > The second thing is I still don't know if "configure.sh sim:nsh menuconfig" is well documentation use of configure.sh > As said above, the arguments after the board config are supposed to be make options not make targets. > Do you have any practical use of this? To hidden the configuration step, we create builld.sh script like this: ``` configure.sh $* make $* ``` User can invoke ```build.sh sim:nsh menuconfig``` to do configuration, it's work well but people report recently that the script hang if he/she run this command on a clear repo. After debuging, we found that it's because the output is redirected to /dev/null. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
