This is the mail archive of the gcc-help@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]

Re: Error when compiling gcc 4.3.1




Ian Lance Taylor-3 wrote:
> 
> Jerry Adair <Jerry.Adair@sas.com> writes:
> 
>> I am encountering a somewhat puzzling error when attempting to build gcc
>> on an HP Itanium box (ia64-hp-hpux11.31).  I've done a google search to
>> see what I could find WRT others having encountered this error too, but
>> nothing resulted, at least that was/is directly applicable. 
>> Mysteriously, the error is telling me that there is no rule to make all
>> (and I didn't find a Makefile in the gcc subdirectory) in my build tree! 
>> I ran the following configure command:
>>
>> ../gcc-4.3.1/configure --prefix /u/jeadai/GCC --with-gmp=/u/jeadai/GCC
>> --with-mpfr=/u/jeadai/GCC
>>
>> Where gcc-4.3.1 is where the source is located and gcc-4.3.1-BuildTree
>> being where I am actually doing the compile.  I then issued the make
>> bootstrap command, which built some of gcc, but when execution flowed
>> into the gcc subdirectory, I got the strange error followed by an early
>> termination of the make bootstrap, thus:
> 
> 
> The gcc build now requires GNU make version 3.80 or later.  Are you
> using GNU make?
> 
> Ian
> 
> 

Hi all,

Ian and I have exchanged several emails regarding the error I am
encountering but we can't seem to pinpoint the trouble.  Has anyone else
ever enountered this as well?  It seems quite strange that the bootstrap
build is being stopped because it doesn't have a rule to make 'all' in the
gcc subdirectory.  I have been referring to "The Definitive Guide to GCC",
by the way.  The context for my situation is that I'm attempting to build on
an HP Itanium box, thus:

HP-UX <name deleted> B.11.31 U ia64 3394250849 unlimited-user license

This box ha no gcc utilities installed on it whatsoever, so I am attempting
to build gcc from scratch.  I have already successfully built and installed
some components of GCC, such as libtoolize, autoconf, automake, GNU make,
etc.  They work as expected.

Here is some additional information regarding things that I have done/tried
during my effort to resolve the problem:

Hmmmmm...well my bin directory is first in my path actually and when I issue
the command as:

GCC/BuildTree> /users/jeadai/bin/make bootstrap

This removes all ambiguity about which make I am executing, "telling Unix"
to run the make found in my bin directory.  Also doing a which produces the
following:

[146] GCC/BuildTree> which make
/users/jeadai/bin/make
[147] GCC/BuildTree> make --version
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for ia64-hp-hpux11.31 [158] GCC/BuildTree>

So the version seems correct and either way it's found, I'm covered...

---------------
A few observations that may be contributing factors:

1) This machine does not already have gcc installed on it, thus when the
make bootstrap comes to the point where a compiler is needed, it uses the
compiler native to the machine (ia64-hp-hpux11.31, which means the aCC
compiler).  An example is the fixincludes stuff:
make[3]: Entering directory
`/r/ge.unx.sas.com/vol/vol4/u42/jeadai/GCC/bt/build-ia64-hp-hpux11.31/fixincludes'
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/fixincl.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/fixtests.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/fixfixes.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/server.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/procopen.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/fixlib.c
cc -Wp,-H256000 -c -g   -DHAVE_CONFIG_H -I. -I../../../gcc-4.3.1/fixincludes
-I../include -I../../../gcc-4.3.1/fixincludes/../include
../../../gcc-4.3.1/fixincludes/fixopts.c
cc -Wp,-H256000 -g  -o fixincl fixincl.o fixtests.o fixfixes.o server.o
procopen.o fixlib.o fixopts.o ../libiberty/libiberty.a echo timestamp >
full-stamp

Where cc means this:
[158] jeadai/GCC> which cc
/usr/bin/cc
[159] jeadai/GCC> cc --version
cc: HP C/aC++ B3910B A.06.22 [Nov 14 2008] [160] jeadai/GCC>

2) I did not download the entire gcc suite, but rather only the core and the
c++ compilers, thus these were the only two tarballs that I put onto my
machine:

gcc-g++-4.3.1.tar*
gcc-core-4.3.1.tar*

I extracted the core first.  Both are in my /u/jeadai/GCC directory, which
thus results in the extraction putting the files into the same directory,
/u/jeadai/GCC/gcc-4.3.1/, which the book states is correct.

3) I am attempting to build on a HPUX-IA64 machine, but the files actually
exist on a mounted disk that resides elsewhere, ergo the files that result
from downloading and building aren't stored locally to the HPUX-IA64 machine
but rather on a mounted drive.  However the execution of the build (of
course) happens on the machine itself.  Is it permissible to run the build
on your host machine (this will be a native compiler) and yet store all of
the files on a remotely mounted disk?

4) I have deleted everything, untarred everything again and tried the
configuration and build.  Same results.

5) On the disk where all of the source and build files are being stored, I
have been allocated 15 G of space, which I am not using much of outside of
this gcc effort.  'Seems like it should be enough disk space, but I could be
missing something.  I have double-checked to see if I'm taking up more space
than I think I am, but alas no.  Even after cleaning up what little I didn't
need, I still get the same error, so it's not an issue of not having enough
disk space.

6) I have not yet installed the binutils component, but from what I read in
the book, it isn't required.  Could this be the problem? (I don't think so,
but I'm grasping for straws at the moment...)

Thank you for reading, I'd much appreciate any clues you could provide to
steer me in the right direction to solve this problem and complete the build
of gcc.

Jerry
-- 
View this message in context: http://www.nabble.com/Error-when-compiling-gcc-4.3.1-tp22253787p22374491.html
Sent from the gcc - Help mailing list archive at Nabble.com.


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