Failed build of egcs-980411 on SGI/IRIX-6.5

Failed build of egcs-980411 on SGI/IRIX-6.5

[Martin Knoblauch]

Hi,

 trying to build egcs-980411 on the current IRIX-6.5
beta I failed twice at different stages:

% ../egcs-980411/config.guess
mips-sgi-irix6.5_BETA_1274425944

 For clarity, --host, --build and --target set
to mips-sgi-irix6.5.

a) when using "--enable-threads", the first stage
   of "gmake bootstrap" fails because the file
   "gthr-irix.h" is missing.

b) without "--enable-threads", the second stage
   fails with an "internal compiler error:

gmake[5]: Entering directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
./xgcc -B./ -O2   -DIN_GCC    -W -Wall -O2 -g -O2
-I./include   -g1  -DIN_LIBGCC2
-D__GCC_FLOAT_NOT_NEEDED -fexceptions    -I.
-I../../egcs-980411/gcc
-I../../egcs-980411/gcc/config \
  -c ../../egcs-980411/gcc/cp/tinfo.cc
../../egcs-980411/gcc/cp/tinfo.cc: In method `void *
__class_type_info::dcast(const class type_info &,
int, void *, const class type_info * = 0, void * = 0)
const':
../../egcs-980411/gcc/cp/tinfo.cc:80: warning:
comparison between signed and unsigned
include/typeinfo: At top level:
include/typeinfo:64: Internal compiler error.
include/typeinfo:64: Please submit a full bug report
to `egcs-bugs@cygnus.com'.
gmake[5]: *** [tinfo.o] Error 1
gmake[5]: Leaving directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
gmake[4]: *** [libgcc2.a] Error 1
gmake[4]: Leaving directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
gmake[3]: *** [stmp-multilib-sub] Error 2
gmake[3]: Leaving directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
gmake[2]: *** [stmp-multilib] Error 1
gmake[2]: Leaving directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
gmake[1]: *** [bootstrap] Error 2
gmake[1]: Leaving directory
`/mnts/d9/tmp/egcs-snapobjs/gcc'
gmake: *** [bootstrap] Error 2


Martin
===
------------------------------------------------------
Martin Knoblauch
email: knobi@rocketmail.com or knobi@sirius-cafe.de
www:   http://home.sirius-cafe.de/~knobi







_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Jim Wilson]

	a) when using "--enable-threads", the first stage
	   of "gmake bootstrap" fails because the file
	   "gthr-irix.h" is missing.

No one has contributed this file yet.  Someone who understands the irix
thread support will need to write the stub routines needed in this file.
If there is no such thing as `irix' threads, and we should instead be using
posix threads, or pthreads, or something else that is already supported,
then that would be easy to fix.

	b) without "--enable-threads", the second stage
	   fails with an "internal compiler error:

The C++ front end is broken.  Almost everyone has run into this.  Hopefully
it will be fixed in time for the next snapshot.

Jim

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Mark Mitchell]

>>>>> "Jim" == Jim Wilson <wilson@cygnus.com> writes:

    Jim> 	a) when using "--enable-threads", the first stage of
    Jim> "gmake bootstrap" fails because the file "gthr-irix.h" is
    Jim> missing.

    Jim> No one has contributed this file yet.  Someone who
    Jim> understands the irix thread support will need to write the
    Jim> stub routines needed in this file.  If there is no such thing
    Jim> as `irix' threads, and we should instead be using posix
    Jim> threads, or pthreads, or something else that is already
    Jim> supported, then that would be easy to fix.

    Jim> 	b) without "--enable-threads", the second stage fails
    Jim> with an "internal compiler error:

    Jim> The C++ front end is broken.  Almost everyone has run into
    Jim> this.  Hopefully it will be fixed in time for the next
    Jim> snapshot.

I have submitted patches that make the C++ front-end work, and I know
that they work on IRIX 6.4.  You should be able to get them from the
archives.  I haven't heard a peep from Jason lately, so I assume he is
either taking a well-deserved vacation, or else fighting some fire at
Cygnus. 

    Jim> Jim

-- 
Mark Mitchell <mmitchell@usa.net>
http://home.earthlink.net/~mbmitchell
Consulting Services Available

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Jeffrey A Law]

  In message < 199804171426.HAA27689@mail.earthlink.net >you write:
  > I have submitted patches that make the C++ front-end work, and I know
  > that they work on IRIX 6.4.  You should be able to get them from the
  > archives.  I haven't heard a peep from Jason lately, so I assume he is
  > either taking a well-deserved vacation, or else fighting some fire at
  > Cygnus. 
I don't remember seeing them, and I can't find them in my "outstanding
patches" folder.

I might have just missed it though.

It's probably worth resending.

jeff

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Mark Mitchell]

>>>>> "Jeffrey" == Jeffrey A Law <law@hurl.cygnus.com> writes:

    Jeffrey>   In message
    Jeffrey> < 199804171426.HAA27689@mail.earthlink.net >you write:
    >> I have submitted patches that make the C++ front-end work, and
    >> I know that they work on IRIX 6.4.  You should be able to get
    >> them from the archives.  I haven't heard a peep from Jason
    >> lately, so I assume he is either taking a well-deserved
    >> vacation, or else fighting some fire at Cygnus.
    Jeffrey> I don't remember seeing them, and I can't find them in my
    Jeffrey> "outstanding patches" folder.

    Jeffrey> I might have just missed it though.

    Jeffrey> It's probably worth resending.

It's a moot point now; they're in the tree.

    Jeffrey> jeff

-- 
Mark Mitchell <mmitchell@usa.net>
http://home.earthlink.net/~mbmitchell
Consulting Services Available

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Martin Knoblauch]

---Jeffrey A Law <law@hurl.cygnus.com> wrote:
>
> 
>   In message
< 19980417071059.23991.rocketmail@web2.rocketmail.com >you
write:
>   >  starting with IRIX-6.5 there will be POSIX
threads
>   > available by default. 6.2, 6.3 and 6.4 need them
>   > to be installed separately.
>   > 
>   >  Can "configure" check whether they are available
>   > and do "the right thing". Maybe I can find the
>   > time to :-) to implement it.
> Yes it can if you write the necessary code.
> 
> We already handle targets where thread support is
conditionally
> available (hpux targets for example).  It's not
that hard to add.
> 
> 
> jeff
> 


  seems like a good starting point. I will have a
look at it.

Martin
===
------------------------------------------------------
Martin Knoblauch
email: knobi@rocketmail.com or knobi@sirius-cafe.de
www:   http://home.sirius-cafe.de/~knobi

_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Martin Knoblauch]

---Jim Wilson <wilson@cygnus.com> wrote:
>
> 	a) when using "--enable-threads", the first stage
> 	   of "gmake bootstrap" fails because the file
> 	   "gthr-irix.h" is missing.
> 
> No one has contributed this file yet.  Someone who
understands the irix
> thread support will need to write the stub routines
needed in this file.
> If there is no such thing as `irix' threads, and we
should instead be using
> posix threads, or pthreads, or something else that
is already supported,
> then that would be easy to fix.
> 
Jim,

 starting with IRIX-6.5 there will be POSIX threads
available by default. 6.2, 6.3 and 6.4 need them
to be installed separately.

 Can "configure" check whether they are available
and do "the right thing". Maybe I can find the
time to :-) to implement it.

Martin
===
------------------------------------------------------
Martin Knoblauch
email: knobi@rocketmail.com or knobi@sirius-cafe.de
www:   http://home.sirius-cafe.de/~knobi

_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: Failed build of egcs-980411 on SGI/IRIX-6.5

[Jeffrey A Law]

  In message < 19980417071059.23991.rocketmail@web2.rocketmail.com >you write:
  >  starting with IRIX-6.5 there will be POSIX threads
  > available by default. 6.2, 6.3 and 6.4 need them
  > to be installed separately.
  > 
  >  Can "configure" check whether they are available
  > and do "the right thing". Maybe I can find the
  > time to :-) to implement it.
Yes it can if you write the necessary code.

We already handle targets where thread support is conditionally
available (hpux targets for example).  It's not that hard to add.


jeff