This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug ada/53677] New: GNAT build didn't stop after a fatal error


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53677

             Bug #: 53677
           Summary: GNAT build didn't stop after a fatal error
    Classification: Unclassified
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ada
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: tschwinge@gcc.gnu.org


Working on a GNU/Hurd for for GCC Ada/GNAT, I stumbled over the following
issue (which I believe not to be specific to the GNU/Hurd port).

I botched up an edit of gcc/ada/s-oscons-tmplt.c (mismatched parentheses;
I tried adding the __GNU__ conditional next to the __linux__ one).  In
the build log I could see the following error message:

    [...]/gcc/ada/s-oscons-tmplt.c:83:5: error: missing ')' in expression
     #if (defined (__linux__) || defined (__GNU__) && !defined (_XOPEN_SOURCE)
         ^
    make -C rts \
                    CC=[...]

However -- as this snippet also shows -- the GNAT build didn't stop at
this point, which I believe it should have, as this is a "hard" syntax
error.


And what is perhaps even more bewildering is that despite this fatal
error, the build successfully continued up to the very end, and even the
make check run produced totally sensible results.  I understand that
s-oscons-tmplt.c is used for generating the package System.OS_Constants,
but how can this have been done successfully if building that file
failed?


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]