Re: [PATCH 2/3][GCC] GNAT/testsuite: Pass the `ada' option to target compilation

"Maciej W. Rozycki" <[email protected]>
Newsgroups gmane.comp.gcc.patches,gmane.comp.sysutils.dejagnu.general
Message-ID <[email protected]>
On Wed, 15 May 2019, Jacob Bachmeyer wrote:

> > Index: gcc/gcc/testsuite/lib/gnat.exp
> > ===================================================================
> > --- gcc.orig/gcc/testsuite/lib/gnat.exp
> > +++ gcc/gcc/testsuite/lib/gnat.exp
> > @@ -167,6 +167,8 @@ proc gnat_target_compile { source dest t
> >  	set options [concat "additional_flags=$TOOL_OPTIONS" $options]
> >      }
> >  
> > +    set options [concat "{ada}" $options]
> > +
> >      return [target_compile $source $dest $type $options]
> >  }
> >   
> Your Tcl syntax looks suspicious to me.  Is there a reason for "ada" to 
> be in both double quotes and braces?

 Most existing `options' elements are lists, as shown by:

clone_output "options: $options\n"

placed at the top of `default_target_compile' (leading paths stripped 
here):

options: {ada} {additional_flags=-fno-diagnostics-show-caret -fno-diagnostics-show-line-numbers -fdiagnostics-color=never --sysroot=.../sysroot} {additional_flags=-gnatJ -c -u} {compiler=.../gcc/gnatmake --GCC=.../gcc/xgcc --GNATBIND=.../gcc/gnatbind --GNATLINK=.../gcc/gnatlink -cargs -B.../gcc -largs --GCC=.../gcc/xgcc\ -B.../gcc\ -march=rv64imafdc\ -mabi=lp64d -margs --RTS=.../riscv64-linux-gnu/lib64/lp64d/libada -q -f} timeout=300

so I did this for consistency, although in reality it doesn't matter, not 
at least for `default_target_compile', and either approach would work.

 We are not consistent here in `gnat_target_compile' anyway, as you can 
see from the two existing `concat' invocations, and also the `timeout=300' 
element.

> Perhaps {lappend options ada} might be simpler?  Is placing ada at the 
> beginning of the list important?

 It can't be last because we override the default compiler otherwise
selected by this option in `default_target_compile', and then options 
passed in may override it further.  Overall I felt it to be safer if we 
placed the compiler type selection first rather than somewhere in the 
middle.

 I hope it clears your concerns.

  Maciej
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.