[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference smurf::buildhelp

Title:USG buildhelp questions/answers
Moderator:SMURF::FILTER
Created:Mon Apr 26 1993
Last Modified:Mon Jan 20 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2763
Total number of notes:5802

2359.0. "Re: FORWARD: Build help question..." by AOSG::FILTER (Automatic Posting Software - mail to flume::puck) Mon Jun 24 1996 21:40

Date Of Receipt: 	24-JUN-1996 17:16:54.39
From: 	FLUME::"mario@zk3.dec.com" "Joe Mario DEC C  24-Jun-1996 1648"
To: 	John Flanagan - UNIX Systems Group <johnf@zk3.dec.com>
CC: 	buildhelp@zk3.dec.com, mario@zk3.dec.com, bagley@zk3.dec.com
Subj: 	Re: FORWARD: Build help question...
John and Dick: 	

	The real problem here is a broken build script in the v3.2
	pools.  The failure is being exposed because you're trying
	to build v3.2 on a v4.0 system.  The problem is not compiler
	related - both decc or acc would error on this.

	Here's a description of what the problem is:

 	 The compilation being done is using the host v4.0 compiler and then
	 mixing that compilation with header files from both the v3.2
	 backing tree and the 4.0 system.  Mixing header files like this
	 is definitely not compatible.  It's probably a good thing the
 	 compiler issued an error because if it succeeded there is no
	 guarantee what the build results would have been.

	 In general, using a newer version of the operating system to build
	 an older version is not guaranteed.   It happened to frequently
	 work in the V3.2c -> V3.2g streams because the development environment
	 on those platforms was relatively stable through those releases.  
	 It clearly won't work between v4.0 and v3.2x.

Joe
	   


> 
> I think this is related to the fact that he is doing the build on a V4.0
> machine.  kernel/src/mkdata is a part of the kernel which uses the host build
> compiler [i.e. the compiler on the machine hosting the sandbox], which means
> he is using the platinum compiler on 3.2 based code.
> 
> Buildhelpers/compiler helpers, advice on this one?
> 
> ------- Forwarded Message
> 
> Return-Path: bagley
> Received: from ralpha.zk3.dec.com by flume.zk3.dec.com; 
> (5.65v3.2/1.1.8.2/16Jan95-0946AM)
> 	id AA09347; Mon, 24 Jun 1996 16:48:40 -0400
> Received: from bwasted.zk3.dec.com by falpha.zk3.dec.com; 
> (5.65v3.2/1.1.8.2/20May95-1022AM)
> 	id AA19939; Mon, 24 Jun 1996 16:48:37 -0400
> Received: by fwasted.zk3.dec.com; (5.65v3.2/1.1.8.2/18Feb95-1123AM)
> 	id AA19184; Mon, 24 Jun 1996 16:48:16 -0400
> Date: Mon, 24 Jun 1996 16:48:16 -0400
> From: Dick Bagley USG <bagley>
> Message-Id: <9606242048.AA19184@fwasted.zk3.dec.com>
> To: odehelp
> Subject: Question regarding V4.0/V3.2 problem
> Cc: bagley
> 
> Hi -
> 
>   I'm running Digital UNIX V4.0 on my workstation, and am attempting to build
>   a v32de2supportos sandbox on it backed against v32de2supportos.bightly. 
> "build
>   setup" fails as shown below. "build setup" runs OK on another workstation
>   running V3.2C for a sandbox backed against v32de2supportos.nightly.
> 
>   Any idea what gives ? Looks like a compiler compatability problem. Don't I
>   inherit the compiler from the tools associated with the backing tree, or is
>   cc run locally from my workstation (which would explain it) ? A similar
>   problem must have existed during V4.0 development. Is there a simple
>   workaround ?
> 
>   Thanks for any help.
> 
> - - dick bagley
>   UHS_IO
>   1-0319
> 
> [ /kernel/src/mkdata ]
> makepath mkdata/. && cd mkdata &&  exec make 'RELEASE_OPTIONS=-idfile 
> usr/bin/cc   -EL -c  -DBSD44 -DMSG  -D__WCHAR_T_LEN=4 -Wf,-wchar32  -DMACH 
> -DCMU -DOSF  -
> Dalpha -D__alpha__ -D__alpha  -Dunix -D__unix__      -O2 -Olimit 5000    -DCM
U 
> -D_NO_PROTO -Dconst=      /usr/include  ../../../../../src/kernel/src/mkdata/
mk
> data.c
> cc: Error: /usr/include/string.h, line 164: In the declaration of "__R", a 
> function may not return a function type.
> extern char    * __R(basename) __((char  *));
> - -----------------^
> cc: Warning: /usr/include/string.h, line 164: In the declaration of "__R", a 
> function declarator has an identifier list but is not p
> art of a function definition.  Extraneous parameter names are ignored.
> extern char    * __R(basename) __((char  *));
> - -----------------^
> cc: Error: /usr/include/string.h, line 165: In the declaration of "__R", a 
> function may not return a function type.
> extern char    * __R(dirname) __((char  *));
> - -----------------^
> cc: Warning: /usr/include/string.h, line 165: In the declaration of "__R", a 
> function declarator has an identifier list but is not p
> art of a function definition.  Extraneous parameter names are ignored.
> extern char    * __R(dirname) __((char  *));
> - -----------------^
> cc: Error: /usr/include/stdlib.h, line 169: In the declaration of "__R", a 
> function may not return a function type.
> extern long     __R(a64l) __((const char *));
> - ----------------^
> cc: Warning: /usr/include/stdlib.h, line 169: In the declaration of "__R", a 
> function declarator has an identifier list but is not p
> art of a function definition.  Extraneous parameter names are ignored.
> extern long     __R(a64l) __((const char *));
> - ----------------^
> cc: Error: /usr/include/stdlib.h, line 170: In the declaration of "__R", a 
> function may not return a function type.
> extern char     * __R(l64a) __((long));
> - ------------------^
> cc: Warning: /usr/include/stdlib.h, line 170: In the declaration of "__R", a 
> function declarator has an identifier list but is not p
> art of a function definition.  Extraneous parameter names are ignored.
> extern char     * __R(l64a) __((long));
> - ------------------^
> cc: Error: /usr/include/stdlib.h, line 171: In the declaration of "__R", a 
> function may not return a function type.
> extern int      __R(ttyslot) __((void));
> - ----------------^
> cc: Warning: /usr/include/stdlib.h, line 171: In the declaration of "__R", a 
> function declarator has an identifier list but is not p
> art of a function definition.  Extraneous parameter names are ignored.
> extern int      __R(ttyslot) __((void));
> - ----------------^
> *** Exit 1
> Stop.
> *** Exit 1
> Stop.
> *** Exit 1
> Stop.
> *** Exit 1
> Stop.
> 
> ------- End of Forwarded Message
> 
> 
> -- 
> 
> 
>  ______________________________________________________________________
> 
>  John Flanagan	 		enet:    johnf@zk3.dec.com	
>  MS: ZKO3-3/W20			decnet:  flume::johnf
>  USG Release Management			 (603) 881-1719
>  110 Spitbrook Road 			 (DTN) 381-1719
>  Nashua, NH  
>  ______________________________________________________________________
> 
> 
> 



T.RTitleUserPersonal
Name
DateLines
2359.1Re: FORWARD: Build help question...AOSG::FILTERAutomatic Posting Software - mail to flume::puckMon Jun 24 1996 22:42189
Date Of Receipt: 	24-JUN-1996 17:53:58.48
From: 	FLUME::"mario@zk3.dec.com" "Joe Mario DEC C  24-Jun-1996 1724"
To: 	bagley@zk3.dec.com
CC: 	johnf@zk3.dec.com, buildhelp@zk3.dec.com, mario@zk3.dec.com
Subj: 	Re: FORWARD: Build help question...
Dick: 	

	I tried your same setup (building the v3.2supportos mkdata from
	a V4.0 system) and it succeeded.  It still looks like something
	in your sandbox is causing the wrong header files to be picked
	up, but I'm not able to reproduce it.  Can you modify CFLAGS in 
	the mkdata Makefile and add both -v and -M to it?  The -M will 
	show where it's picking up the header files and the -v will show 
	the expansion of the 'genpath' include directories.

	You can send the output to me offline if this is cluttering up
	the buildhelp alias.

Joe


> John and Dick:
> 	The real problem here is a broken build script in the v3.2
> 	pools.  The failure is being exposed because you're trying
> 	to build v3.2 on a v4.0 system.  The problem is not compiler
> 	related - both decc or acc would error on this.
> 
> 	Here's a description of what the problem is:
> 
>  	 The compilation being done is using the host v4.0 compiler and then
> 	 mixing that compilation with header files from both the v3.2
> 	 backing tree and the 4.0 system.  Mixing header files like this
> 	 is definitely not compatible.  It's probably a good thing the
>  	 compiler issued an error because if it succeeded there is no
> 	 guarantee what the build results would have been.
> 
> 	 In general, using a newer version of the operating system to build
> 	 an older version is not guaranteed.   It happened to frequently
> 	 work in the V3.2c -> V3.2g streams because the development environment
> 	 on those platforms was relatively stable through those releases.  
> 	 It clearly won't work between v4.0 and v3.2x.
> 
> Joe
> 	   
> 
> 
> > 
> > I think this is related to the fact that he is doing the build on a V4.0
> > machine.  kernel/src/mkdata is a part of the kernel which uses the host bui
ld
> > compiler [i.e. the compiler on the machine hosting the sandbox], which mean
s
> > he is using the platinum compiler on 3.2 based code.
> > 
> > Buildhelpers/compiler helpers, advice on this one?
> > 
> > ------- Forwarded Message
> > 
> > Return-Path: bagley
> > Received: from ralpha.zk3.dec.com by flume.zk3.dec.com; 
> > (5.65v3.2/1.1.8.2/16Jan95-0946AM)
> > 	id AA09347; Mon, 24 Jun 1996 16:48:40 -0400
> > Received: from bwasted.zk3.dec.com by falpha.zk3.dec.com; 
> > (5.65v3.2/1.1.8.2/20May95-1022AM)
> > 	id AA19939; Mon, 24 Jun 1996 16:48:37 -0400
> > Received: by fwasted.zk3.dec.com; (5.65v3.2/1.1.8.2/18Feb95-1123AM)
> > 	id AA19184; Mon, 24 Jun 1996 16:48:16 -0400
> > Date: Mon, 24 Jun 1996 16:48:16 -0400
> > From: Dick Bagley USG <bagley>
> > Message-Id: <9606242048.AA19184@fwasted.zk3.dec.com>
> > To: odehelp
> > Subject: Question regarding V4.0/V3.2 problem
> > Cc: bagley
> > 
> > Hi -
> > 
> >   I'm running Digital UNIX V4.0 on my workstation, and am attempting to bui
ld
> >   a v32de2supportos sandbox on it backed against v32de2supportos.bightly. 
> > "build
> >   setup" fails as shown below. "build setup" runs OK on another workstation
> >   running V3.2C for a sandbox backed against v32de2supportos.nightly.
> > 
> >   Any idea what gives ? Looks like a compiler compatability problem. Don't 
I
> >   inherit the compiler from the tools associated with the backing tree, or 
is
> >   cc run locally from my workstation (which would explain it) ? A similar
> >   problem must have existed during V4.0 development. Is there a simple
> >   workaround ?
> > 
> >   Thanks for any help.
> > 
> > - - dick bagley
> >   UHS_IO
> >   1-0319
> > 
> > [ /kernel/src/mkdata ]
> > makepath mkdata/. && cd mkdata &&  exec make 'RELEASE_OPTIONS=-idfile 
> > usr/bin/cc   -EL -c  -DBSD44 -DMSG  -D__WCHAR_T_LEN=4 -Wf,-wchar32  -DMACH 
> > -DCMU -DOSF  -
> > Dalpha -D__alpha__ -D__alpha  -Dunix -D__unix__      -O2 -Olimit 5000    -D
CM
> U 
> > -D_NO_PROTO -Dconst=      /usr/include  ../../../../../src/kernel/src/mkdat
a/
> mk
> > data.c
> > cc: Error: /usr/include/string.h, line 164: In the declaration of "__R", a 
> > function may not return a function type.
> > extern char    * __R(basename) __((char  *));
> > - -----------------^
> > cc: Warning: /usr/include/string.h, line 164: In the declaration of "__R", 
a 
> > function declarator has an identifier list but is not p
> > art of a function definition.  Extraneous parameter names are ignored.
> > extern char    * __R(basename) __((char  *));
> > - -----------------^
> > cc: Error: /usr/include/string.h, line 165: In the declaration of "__R", a 
> > function may not return a function type.
> > extern char    * __R(dirname) __((char  *));
> > - -----------------^
> > cc: Warning: /usr/include/string.h, line 165: In the declaration of "__R", 
a 
> > function declarator has an identifier list but is not p
> > art of a function definition.  Extraneous parameter names are ignored.
> > extern char    * __R(dirname) __((char  *));
> > - -----------------^
> > cc: Error: /usr/include/stdlib.h, line 169: In the declaration of "__R", a 
> > function may not return a function type.
> > extern long     __R(a64l) __((const char *));
> > - ----------------^
> > cc: Warning: /usr/include/stdlib.h, line 169: In the declaration of "__R", 
a 
> > function declarator has an identifier list but is not p
> > art of a function definition.  Extraneous parameter names are ignored.
> > extern long     __R(a64l) __((const char *));
> > - ----------------^
> > cc: Error: /usr/include/stdlib.h, line 170: In the declaration of "__R", a 
> > function may not return a function type.
> > extern char     * __R(l64a) __((long));
> > - ------------------^
> > cc: Warning: /usr/include/stdlib.h, line 170: In the declaration of "__R", 
a 
> > function declarator has an identifier list but is not p
> > art of a function definition.  Extraneous parameter names are ignored.
> > extern char     * __R(l64a) __((long));
> > - ------------------^
> > cc: Error: /usr/include/stdlib.h, line 171: In the declaration of "__R", a 
> > function may not return a function type.
> > extern int      __R(ttyslot) __((void));
> > - ----------------^
> > cc: Warning: /usr/include/stdlib.h, line 171: In the declaration of "__R", 
a 
> > function declarator has an identifier list but is not p
> > art of a function definition.  Extraneous parameter names are ignored.
> > extern int      __R(ttyslot) __((void));
> > - ----------------^
> > *** Exit 1
> > Stop.
> > *** Exit 1
> > Stop.
> > *** Exit 1
> > Stop.
> > *** Exit 1
> > Stop.
> > 
> > ------- End of Forwarded Message
> > 
> > 
> > -- 
> > 
> > 
> >  ______________________________________________________________________
> > 
> >  John Flanagan	 		enet:    johnf@zk3.dec.com	
> >  MS: ZKO3-3/W20			decnet:  flume::johnf
> >  USG Release Management			 (603) 881-1719
> >  110 Spitbrook Road 			 (DTN) 381-1719
> >  Nashua, NH  
> >  ______________________________________________________________________
> > 
> > 
> > 
> 
>