embeddedml-developer Mailing List for Embedded ML

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

---------- Forwarded message ----------
From: intw6 <int...@gm...>
Date: Sun, Nov 29, 2009 at 3:32 PM
Subject: Re: [Embeddedml-developer] problem with embedded ml
To: Andrew Harris <and...@gm...>


hi,
that is no problem at all,
thanks again


On Sun, Nov 29, 2009 at 1:01 PM, Andrew Harris <and...@gm...> wrote:
>
> Would you mind if I forwarded our email discussion to the
> embeddedml-developer list?  That way others who searched for this
> might be able to find the help too.
>
> -andrew
>
> On Sun, Nov 29, 2009 at 12:31 PM, Andrew Harris <and...@gm...> wrote:
> > No problem.  I am glad you got it working.  Sorry for the extra work.
> > Let me warn you that when you are writing ml programs, and you have a
> > syntax error, the compiler is very terse, it doesn't even tell you a
> > line number.  This is quite annoying, but if you write small portions
> > of code and compile as you go, you shouldn't have too many troubles.
> > Note you don't have to run the program, just compile it.  That way any
> > syntax errors you have will be easier to track down.  Feel free to
> > email me with any questions.  Other errors, such as type check
> > failures are more descriptive.  It's only in the early parsing stage
> > that the errors aren't very useful.
> >
> > -andrew
> >
> > On Sun, Nov 29, 2009 at 12:25 PM, intw6 <int...@gm...> wrote:
> >> hello,
> >> after installing the new svn update
> >> i installed libc6-dev-i386 for ubuntu x64
> >> everything works now.
> >> thank you.
> >>
> >>
> >>
> >>
> >> On Sun, Nov 29, 2009 at 10:59 AM, Andrew Harris <and...@gm...>
> >> wrote:
> >>>
> >>> Hi intworld6,
> >>>
> >>>   I believe I have fixed that.  If you do an svn update in the root
> >>> directory (where compile.sml) is, you should get the changes that
> >>> support the cflags option.  Alternatively, you can download the
> >>> compiler again now and try it.  The new "cflags" option is applied
> >>> earlier, when each file is compiled.
> >>>
> >>> -andrew
> >>>
> >>> On Sun, Nov 29, 2009 at 10:55 AM, intw6 <int...@gm...> wrote:
> >>> > hi,
> >>> >
> >>> > the -cflags option didn't exist (i downloaded the svn version of
> >>> > embedded
> >>> > ml),
> >>> >
> >>> > instead, i tried -compileflags "-m32"
> >>> >
> >>> > it looks that -m32 option is only applied towards the end, and not with
> >>> > the
> >>> > dependencies.
> >>> >
> >>> > ar: creating libarchive.a
> >>> > ranlib libarchive.a
> >>> > gcc -m32 -O2 -L. -larchive  -o a.out
> >>> > /usr/bin/ld: skipping incompatible ./libarchive.a when searching for
> >>> > -larchive
> >>> > /usr/bin/ld: cannot find -larchive
> >>> > collect2: ld returned 1 exit status
> >>> > make: *** [a.out] Error 1
> >>> >
> >>> >
> >>> > i can use 32-bit ubuntu running in virtual box while embedded ml is
> >>> > upgraded
> >>> > to 64-bit
> >>> >
> >>> > thanks!
> >>> >
> >>> >
> >>> >
> >>> > On Sun, Nov 29, 2009 at 10:33 AM, Andrew Harris <and...@gm...>
> >>> > wrote:
> >>> >>
> >>> >> Hi intworld6,
> >>> >>
> >>> >>   OK, I have modified the compiler slightly to make it easier to
> >>> >> accept compiler flags.  If you update svn and recompile the compiler,
> >>> >> you should be able to use a new option on the compiler called "cflags"
> >>> >> that will allow you to pass the -m32 option to the compiler when the C
> >>> >> files are compiled.  This should compile the code with 32-bit
> >>> >> instructions.  You can trying compiling sandmark.uml as follows:
> >>> >>
> >>> >> ../mlc.exe -cbackend sandmark.uml -cflags "-m32"
> >>> >>
> >>> >> and then just cd output/ and make as you did before.  Let me know how
> >>> >> it goes.  Also if it doesn't work, please send me the output of the
> >>> >> make.
> >>> >>
> >>> >> thanks,
> >>> >> -andrew
> >>> >>
> >>> >> On Sun, Nov 29, 2009 at 10:11 AM, Andrew Harris <and...@gm...>
> >>> >> wrote:
> >>> >> > Hi intworld6,
> >>> >> >
> >>> >> >   This is likely caused by gcc compiling 64-bit instructions.  The
> >>> >> > runtime for embeddedml currently requires 32-bit instructions.  Give
> >>> >> > me a bit to think of a good way to resolve this.
> >>> >> >
> >>> >> > -andrew
> >>> >> >
> >>> >> > On Sun, Nov 29, 2009 at 9:35 AM, intw6 <int...@gm...> wrote:
> >>> >> >> hello,
> >>> >> >>
> >>> >> >> the sandmark example now compiles,
> >>> >> >> but a.out gives a segmentation fault error.
> >>> >> >>
> >>> >> >> thank you
> >>> >> >>
> >>> >> >>
> >>> >> >> On Sun, Nov 29, 2009 at 9:08 AM, Andrew Harris
> >>> >> >> <and...@gm...>
> >>> >> >> wrote:
> >>> >> >>>
> >>> >> >>> Hi intworld6,
> >>> >> >>>
> >>> >> >>>   The first thing I would try would be to make sure there is no
> >>> >> >>> output/ directory in the directory in which you're trying to
> >>> >> >>> compile
> >>> >> >>> sandmark.uml.  Before trying the compilation, try deleting that
> >>> >> >>> directory with rm -rf output/
> >>> >> >>>
> >>> >> >>>   The output/ directory is where the intermediate C files and
> >>> >> >>> makefiles are stored.  If that directory already exists, the
> >>> >> >>> compilation will fail.  Let me know if that doesn't help.  And of
> >>> >> >>> course, make sure to type that line carefully!
> >>> >> >>>
> >>> >> >>> -andrew
> >>> >> >>>
> >>> >> >>> On Sun, Nov 29, 2009 at 8:36 AM, intw6 <int...@gm...> wrote:
> >>> >> >>> > hello,
> >>> >> >>> >
> >>> >> >>> > there is no additional output after "Compilation failed".
> >>> >> >>> >
> >>> >> >>> > thank you
> >>> >> >>> >
> >>> >> >>> > On Sun, Nov 29, 2009 at 7:46 AM, Andrew Harris
> >>> >> >>> > <and...@gm...>
> >>> >> >>> > wrote:
> >>> >> >>> >>
> >>> >> >>> >> Hi intworld6, is there any additional text that is output after
> >>> >> >>> >> the
> >>> >> >>> >> words
> >>> >> >>> >> "Compilation failed.?
> >>> >> >>> >>
> >>> >> >>> >> thanks,
> >>> >> >>> >> -andrew
> >>> >> >>> >>
> >>> >> >>> >> On Nov 28, 2009, at 3:52 PM, intw6 <int...@gm...> wrote:
> >>> >> >>> >>
> >>> >> >>> >>> hello,
> >>> >> >>> >>> i am using ubuntu 9.04 amd64 and mlton to compile embeddedml
> >>> >> >>> >>> it compiles with no problem, but i receive an error when
> >>> >> >>> >>> attempting to
> >>> >> >>> >>> run one of the tests
> >>> >> >>> >>>
> >>> >> >>> >>> following output:
> >>> >> >>> >>>
> >>> >> >>> >>> ../mlc.exe -cbackend sandmark.uml
> >>> >> >>> >>> Parsing...
> >>> >> >>> >>> Elaborating...
> >>> >> >>> >>> Optimizing IL...
> >>> >> >>> >>> (1387640) (122300) (46850) (33310) (18205) (18000) (17800)
> >>> >> >>> >>> (14300)
> >>> >> >>> >>> (14100) (12600) (7800) (7800) (7800) (7800) (7800) (7800)
> >>> >> >>> >>> (7800)
> >>> >> >>> >>> (7800)
> >>> >> >>> >>> (7800) (7800) (7800) (7800) (5200) (0)
> >>> >> >>> >>> CPS converting...
> >>> >> >>> >>> Optimizing CPS...
> >>> >> >>> >>> (infinite) (infinite) (10) (150) (0)
> >>> >> >>> >>> Closure converting...
> >>> >> >>> >>> Alloc converting...
> >>> >> >>> >>> Backend conversion
> >>> >> >>> >>> to C...
> >>> >> >>> >>> Assembling...
> >>> >> >>> >>>
> >>> >> >>> >>> Compilation failed.
> >>> >> >>> >>>

2008	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct (1)	Nov	Dec
2009	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov (3)	Dec

embeddedml-developer Mailing List for Embedded ML

embeddedml-developer — List for persons interested in embedded ML development.