fixproto crash

fixproto crash

[Mark Mitchell]

Folks --

  On an sgi-irix-6.5 build I'm seeing:

    Various warnings and error messages from fixproto are normal
    if [ -d include ] ; then true; else mkdir include; fi
    if [ -f include/fixed ] ; then true; \
    else \
      : This line works around a 'make' bug in BSDI 1.1.; \
      FIXPROTO_DEFINES="-D__EXTENSIONS__  -D_SGI_SOURCE -D_LANGUAGE_C_PLUS_PLUS"; export FIXPROTO_DEFINES; \
      if [ -d /usr/include ] ; then \
	/bin/sh ../../gcc/fixproto include include /usr/include; \
      else true; fi; \
      touch include/fixed; \
    fi
    fixproto: populating `include'
    ../../gcc/fixproto[251]: 277590 Bus error(coredump)

While I understand that `various warnings and error messages from
fixproto are normal', I would not expect `coredump' to be among them.

Interestingly, the build continues; only my eagle eyes :-) spotted
this possible problem.

Finally, I thought we'd cut over to fast-fixincludes, but I guess
that's still a configuration option?  Why haven't we thrown the
switch?

-- 
Mark Mitchell 			mark@markmitchell.com
Mark Mitchell Consulting	http://www.markmitchell.com

Re: fixproto crash

[Jeffrey A Law]

  In message < 199812152218.OAA23750@adsl-206-170-148-33.dsl.pacbell.net >you wri
te:
  >     fixproto: populating `include'
  >     ../../gcc/fixproto[251]: 277590 Bus error(coredump)
  > 
  > While I understand that `various warnings and error messages from
  > fixproto are normal', I would not expect `coredump' to be among them.
Yup. 

  > Interestingly, the build continues; only my eagle eyes :-) spotted
  > this possible problem.
Yea, this probably explains why I didn't "see" the problem recently -- now
to find that patch which was supposed to fix this bug.

  > Finally, I thought we'd cut over to fast-fixincludes, but I guess
  > that's still a configuration option?  Why haven't we thrown the
  > switch?
Still some issues to deal with.  Note that we'll still need fixproto, those
tools solve significantly different problems.

jeff

Re: fixproto crash

[Zack Weinberg]

On Thu, 17 Dec 1998 22:38:54 -0700, Jeffrey A Law wrote:
>
>  In message < 199812152218.OAA23750@adsl-206-170-148-33.dsl.pacbell.net >you wr
>i
>te:
>  >     fixproto: populating `include'
>  >     ../../gcc/fixproto[251]: 277590 Bus error(coredump)
>  > 
>  > While I understand that `various warnings and error messages from
>  > fixproto are normal', I would not expect `coredump' to be among them.
>Yup. 
>
>  > Interestingly, the build continues; only my eagle eyes :-) spotted
>  > this possible problem.
>Yea, this probably explains why I didn't "see" the problem recently -- now
>to find that patch which was supposed to fix this bug.

Would this be the #include_next in primary source file bug, which I
believe Dave B and I squashed?

1998-12-16  Zack Weinberg  <zack@rabi.phys.columbia.edu>

        * cpplib.c (do_include): Treat #include_next in the
        primary source file as #include plus warning.  Treat
        #include_next in a file included by absolute path as an
        error.  fp == CPP_NULL_BUFFER is a fatal inconsistency.

Unfortunately I don't have access to the platform in question (hpux,
right?) so can't confirm this.

zw