Article View: gnu.gcc.help
Article #51768gcc fixincludes problem under RISCos 4.51
From: Paul Bickerstaff
Date: Tue, 13 Apr 1993 20:30
Date: Tue, 13 Apr 1993 20:30
99 lines
5214 bytes
5214 bytes
I have recently "built" and installed gcc-2.3.3 on a Mips Magnum 3000
under RISCos 4.51. There appear to be some problems, especially with
the gcc includes files. My first serious attempt to use gcc (a build of
world wide web in xmosaic) results in
gcc -c HTAccess.c
In file included from HTAccess.c:11:
HTUtils.h:47: stdlib.h: No such file or directory
In file included from /bsd43/usr/include/sys/file.h:37, from tcp.h:311,
from HTA
ccess.c:19:
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/bsd43/sys/type
s.h:130:
warning: `NBBY' redefined
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/sys/param.h:12
3: warni
ng: this is the location of the previous definition
In file included from tcp.h:316, from HTAccess.c:19:
/bsd43/usr/include/sys/dir.h:137: warning: `NULL' redefined
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/stdio.h:80:
warning: t
his is the location of the previous definition
*** Error code 1
I had assumed that fixincludes would put the include files in reasonable
(i.e. standard -- whatever that means to gcc) places. But a look at the README
shows that that is not what it does and it has simply gone through *all* the
Mips include files (not just those for bsd) and "fixed" the non-ansi stuff.
So I linked the <"fixed">/posix/stdlib.h to where programs will want to
look for it
and tried again. It continues
In file included from tcp.h:303, from HTAccess.c:19:
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/sys/types.h:63
: redefi
nition of `size_t'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/stdlib.h:28: `size_t'
previously declared here
In file included from /bsd43/usr/include/sys/file.h:37, from tcp.h:311,
from HTA
ccess.c:19:
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/bsd43/sys/types.h:54:
conflicting types for `physadr'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/sys/types.h:30
: previo
us declaration of `physadr'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/bsd43/sys/types.h:61:
conflicting types for `label_t'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/sys/types.h:54
: previo
us declaration of `label_t'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/bsd43/sys/types.h:86:
conflicting types for `size_t'
/usr/local/lib/gcc-lib/mips-mips-riscos4bsd/2.3.3/include/sys/types.h:63
: previo
us declaration of `size_t'
*** Error code 1
I'm not sure where to begin with this. It looks like the fixincludes
stage of my build
has gone all wrong but it is a fully automated step. Surely it should
have only done
the bsd43 includes but since it did do both why, if it fixed both, are
there conflicting
types? If I can force fixincludes to only fix the bsd43 includes what
happens to
stdlib.h in posix? Am I left without it? Is it created for me?
Any help would be much appreciated. Somebody out there must have a *fully*
functional gcc on a RISCos 4.5 box.
Here are a few other details on what I did:
gcc-2.3.3 is yet another package where the stuff supplied for Mips machines
does not build as provided. The configure step creates a makefile that does
not work! (From the above you will see that I built with mips-mips-riscos4bsd,
other valid combinations not working either.) I proceeded with fixing a CC
flag somewhere to include -systype bsd43. (How the hell can something as
simple as this not get included in the otherwise nice configure?)
The build then went smoothly apart from a weird file __________EBEB_
not being created in several libgcc libraries. (Anybody know what this
file is and is it important?) I (otherwise) successfully (I think :-) ) built
the stage 3 compiler and it passed the compare test.
The fixincludes step seemed to go on forever and warned many times
about redefinitions and something being the place of the previous
definition. (Sorry, didn't keep a log of the build but it would be too
long to post anyway.) Since I didn't understand I didn't care :-(
and went ahead and installed.
Can anybody tell me what I should have done.
Thanks
Paul Bickerstaff Internet: pbickers@tamaluit.phys.uidaho.edu
Physics Dept., Univ. of Idaho Phone: (208) 885 6809
Moscow ID 83843-2341, USA FAX: (208) 885 6173
Technical Details
Message-ID:
<C5FvnM.6Dx@moscow.uidaho.edu>Path:
arc1-ca.pugleaf.net!gmd.de!newsserver.jvnc.net!howland.reston.ans.net!zaphod.mps.ohio-state.edu!uwm.edu!ogicse!news.u.washington.edu!serval!moscow.uidaho.edu!tamaluit!pbickersShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 11318c,6cb41000595f5be,start X-Google-Attributes: gid11318c,public X-Google-Thread: 1051a8,6cb41000595f5be,start X-Google-Attributes: gid1051a8,public X-Google-ArrivalTime: 1993-04-13 17:03:52 PST Path: arc1-ca.pugleaf.net!gmd.de!newsserver.jvnc.net!howland.reston.ans.net!zaphod.mps.ohio-state.edu!uwm.edu!ogicse!news.u.washington.edu!serval!moscow.uidaho.edu!tamaluit!pbickers From: pbickers@tamaluit.phys.uidaho.edu (Paul Bickerstaff) Newsgroups: gnu.gcc.help,comp.sys.mips Subject: gcc fixincludes problem under RISCos 4.51 Message-ID: <C5FvnM.6Dx@moscow.uidaho.edu> Date: 13 Apr 93 20:30:57 GMT Article-I.D.: moscow.C5FvnM.6Dx Sender: news@moscow.uidaho.edu Reply-To: pbickers@tamaluit.phys.uidaho.edu (Paul Bickerstaff) Organization: phys dept Lines: 99 Nntp-Posting-Host: tamaluit.phys.uidaho.edu