sorry, please ignore the previous mail. i've just realized that i need
to build binutils first. 

terribly sorry,
-yw

On Fri, 2003-04-25 at 17:47, Y.W. Law wrote:
> hi all,
> 
> (1) while compiling from the CVS, i've encountered this problem:
> 
> config/msp430/libgcc.S: Assembler messages:
> config/msp430/libgcc.S:11: Error: junk `(r1)' after expression
> config/msp430/libgcc.S:11: Error: too many memory references for `sub'
> config/msp430/libgcc.S:12: Error: no such instruction: `subc 4(r1),r13'
> config/msp430/libgcc.S:13: Error: no such instruction: `subc 6(r1),r14'
> config/msp430/libgcc.S:14: Error: no such instruction: `subc 8(r1),r15'
> config/msp430/libgcc.S:16: Error: no such instruction: `tst r15'
> config/msp430/libgcc.S:16: Error: invalid character '<' in mnemonic
> config/msp430/libgcc.S:19: Error: suffix or operands invalid for `mov'
> config/msp430/libgcc.S:22: Error: no such instruction: `bis r12,r14'
> config/msp430/libgcc.S:22: Error: no such instruction: `check if zero'
> config/msp430/libgcc.S:23: Error: no such instruction: `bis r13,r15'
> config/msp430/libgcc.S:24: Error: no such instruction: `bis r14,r15'
> config/msp430/libgcc.S:25: Error: no such instruction: `tst r15'
> config/msp430/libgcc.S:26: Error: suffix or operands invalid for `je'
> config/msp430/libgcc.S:26: Error: junk `or or105ng all nibbles10' after
> expression
> config/msp430/libgcc.S:26: Error: suffix or operands invalid for `test'
> config/msp430/libgcc.S:27: Error: suffix or operands invalid for `mov'
> config/msp430/libgcc.S:30: Error: suffix or operands invalid for `mov'
> make[2]: *** [libgcc/./_cmpdi2.o] Error 1
> 
> i'm using the gcc-3.3 branch (of MSPGCC) and gcc-3.2.2 source, but
> gcc-3.2 source gives the same error. since i have actually compiled it
> successfully once, i am wondering if any CVS commit has produced the
> error.
> 
> (2) which branch is actually the latest? the only thing i know for sure
> is that gcc-3.1 is not being maintained
> 
> (3) i am also wondering why createlinks_gcc.py has hardcoded the branch
> to gcc-3.0.
> 
> (4) the script createlinks_gcc.py actually does not work for me. i have
> looked at the source code, understood it and modified it to the essense
> of what the original createlinks_gcc.py is purported to achieve. (see
> below). logically the error should not be due to my modified version of
> createlinks_gcc.py.
> 
> 
> thank you,
> yw
> 
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 
> #!/usr/bin/env python
> 
> # this script is for un*x/linux users. it creates hardlinks from this
> # directory to a target source tree.
> # this allows to check out the patch to a separate directory rather than
> # mixing the original gcc/binutils sources with this patch directly.
> 
> # [email protected]
> # $Id: createlinks_gcc.py,v 1.4 2002/01/20 03:24:47 cliechti Exp $
> 
> import sys,os, string, stat
> VERSION = string.split("$Revision: 1.4 $")[1]   #extract CVS version
> 
> #this is the name of the directory of the gcc sources
> #it starts with "./"
> versiondir = './gcc-3.3/'
> 
> #first scan for files in the source directory
> #leave out CVS dirs
> files = []
> dirs = []
> def visitor(arg, dirname, names):
>     #dirname = dirname[len(versiondir):]
>     for name in names:
>         if name and name[-3:] != 'CVS' and (dirname[-3:] != 'CVS'):
>             path = os.path.join(dirname,name)
>             if os.path.isdir(os.path.abspath(path)):                
>                 path = path[len(versiondir):]
>                 print "Added dir ", path
>                 dirs.append(path)
>             else:                
>                 path = path[len(versiondir):]
>                 print "Added file ", path
>                 files.append(path)
> 
> os.path.walk(versiondir, visitor, None)
> 
> #option(s) for the ln command
> lnopts = '-s' #'-s' #to be used when symbolic links should be
> genereated, empty string for hardlinks
> 
> if len(sys.argv) < 2:
>         print """USAGE: %s targetdir
>         version: %s
> 
>         This script replaces the files that are in the gcc source with
>         links to files in this directory.
>         
>         Please specify the target directory, where the gcc sources are
>         located, e.g. /home/user/src/gcc-3.2.2
> 
>         Warning: the original files in the target directory tree are
>         deleted!!
> 
>         The following files will be modifed in the target directory:
> """ % (sys.argv[0], VERSION)
> 
>         for f in files:
>                 print "\t\t%s" % os.path.join(versiondir, f)
> 
>         print """\n\tscroll up if you don't see the help message ;-)"""
>         sys.exit(1)
> 
> #make paths absolute so that the links are absolute too
> source = os.path.abspath(versiondir)
> target = os.path.abspath(sys.argv[1]);
> 
> if not os.path.exists(target):
>     print "The directory %s does not exist." % target
>     sys.exit(1)
> 
> print "Source dir: %s" % source
> print "Target dir: %s" % target
> 
> if os.path.abspath(source) == os.path.abspath(target):
>         print "No, no! Don't specify the source directory as target."
>         sys.exit(1)
> 
> for d in dirs:        
>         tgt = os.path.join(target, d)
>         if not os.path.exists(tgt):
>                 print "Creating directory: ", tgt
>                 cmd = "mkdir %s " % (tgt)
>                 os.system(cmd) 
>         else: 
>                 print "Directory %s already exists" % tgt
> 
> for f in files:
>         src = os.path.join(source, f) 
>         tgt = os.path.join(target, f) 
>         if os.path.exists(tgt):
>             if os.path.samefile(src, tgt):
>                 print "%s already linked to %s" % (tgt, src)
>                 continue
>             else:
>                 print "Removing file: ", tgt  
>                 os.system("rm %s" % tgt)
>         cmd = "ln %s %s %s" % (lnopts, src, tgt)
>         print "  ", cmd
>         os.system(cmd) 
> 
> print "Done"
> 
> 
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Mspgcc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/mspgcc-users


Reply via email to