Experimental Physics and
Industrial Control System

"Hill, Jeff" <[email protected]> · Tue, 14 May 2013 23:58:18 +0000

> I'm using Eclipse CDT, and I'm not sure if it doesn't have this problem, or if I
> didn't try the right test. I used an IOC app, typed "findThisError." on a random
> line in a source file, compiled, double clicked on the error message
> ("../drvSis3350.cpp:361: error: 'findThisError' was not declared in this scope"),
> and it took me to the right place.

This issue will of course be very specific to both the compiler and the IDE, and problems might be worse when you are working several levels down in base, IDK.

I currently see this issue with the code blocks IDE, and in the past also with certain debuggers although the issue with debuggers was some time back, and I don't remember the specifics at this point. As I recall, the EPICS build system works as-is with Visual C++ as long as one change some search path settings, but with some border case problem remaining with certain files. Visual C++ has been changing over the years so this may be less of a problem with the newer versions.

Jeff

> -----Original Message-----
> From: Brown, Garth [mailto:[email protected]]
> Sent: Tuesday, May 14, 2013 4:02 PM
> To: Hill, Jeff; Tech Talk ([email protected])
> Subject: RE: using the EPICS build system with an IDE
> 
> Hi Jeff,
> I'm using Eclipse CDT, and I'm not sure if it doesn't have this problem, or if I
> didn't try the right test. I used an IOC app, typed "findThisError." on a random
> line in a source file, compiled, double clicked on the error message
> ("../drvSis3350.cpp:361: error: 'findThisError' was not declared in this scope"),
> and it took me to the right place.
> 
> Garth
> 
> -----Original Message-----
> From: [email protected] [mailto:tech-talk-
> [email protected]] On Behalf Of Hill, Jeff
> Sent: Tuesday, May 14, 2013 11:38 AM
> To: Tech Talk ([email protected])
> Subject: RE: using the EPICS build system with an IDE
> 
> 
> Furthermore, this simple patch to configure/RULES_BUILD appears to
> produce IDE compatible compiler output, which appears to be an
> improvement, from my perspective.
> 
> @@ -210,15 +210,15 @@
> 
>  %$(OBJ): %.c
>  	@$(RM) $@
> -	$(COMPILE.c) -c $<
> +	$(COMPILE.c) -c $(realpath $<)
> 
>  %$(OBJ): %.cc
>  	@$(RM) $@
> -	$(COMPILE.cpp) -c $<
> +	$(COMPILE.cpp) -c $(realpath $<)
> 
>  %$(OBJ): %.cpp
>  	@$(RM) $@
> -	$(COMPILE.cpp) -c $<
> +	$(COMPILE.cpp) -c $(realpath $<)
> 
> 
> > -----Original Message-----
> > From: [email protected] [mailto:tech-talk-
> > [email protected]] On Behalf Of Hill, Jeff
> > Sent: Tuesday, May 14, 2013 12:09 PM
> > To: Tech Talk ([email protected])
> > Subject: using the EPICS build system with an IDE
> >
> > Hi All,
> >
> > I have noticed over the years that the EPICS build system isn't
> > particularly helpful when using an IDE because it doesn't generate
> > full source file paths on the compiler command line, and instead
> > subdirectory dependent relative paths are used. This seems to cause
> > gcc to produce also relative source file paths in its compiler
> > messages which can interfere with clicking on the error message and
> > vectoring the focus to the appropriate source file in the IDE. It
> > seems that gcc could easily provide an option to force generation of
> > full (not
> > relative) source file paths in its output no matter what the format is
> > of its input file paths might be, but AFAIKT, it doesn't.
> >
> > However, maybe this post suggests a solution that could be applied
> > generically within the internals of the EPICS build system to produce
> > full path names in the compilers input while continuing to navigate
> > the build system with relative paths.
> >
> > http://gcc.gnu.org/ml/gcc-help/2008-03/msg00084.html
> >
> > This would result in rather more verbose build message output, but
> > would improve productivity when using an IDE from my perspective. I
> > suppose that we might also provide a configure option which determines
> > the type of compiler file path output that gets generated depending on
> > a particular user's needs, if that was important.
> >
> > Comments?
> >
> > Jeff
> 

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System