Re: [Gcc4cli-devel] Binutils/as patch

[Andrea C. O. <and...@gm...>]

ok.

On Tue, Mar 23, 2010 at 2:24 PM, Thierry Lafage <thi...@in...> wrote:
> Yes, sorry.
> Here is the complete patch with CSObject.cs cleanup.
>
> Sorry for the "noise",
>
> Thierry Lafage.
>
>
>
> Andrea Carlo Ornstein a écrit :
>>
>> No, I wasn't clear :)
>> Something like this:
>>
>>       -o)
>>  -          PARAM="$PARAM $1$2";
>> +          PARAM="$PARAM -output $2";
>>           shift
>>           shift
>>           ;;
>> and remove the sed at the end
>>
>> the script keeps the same input options
>> but there is no need of the sed at the end, we just rewrite the option
>> when we recognize it
>>
>> On Tue, Mar 23, 2010 at 2:06 PM, Thierry Lafage <thi...@in...>
>> wrote:
>>
>>>
>>> Andrea Carlo Ornstein a écrit :
>>>
>>>>
>>>> Hi,
>>>>
>>>> I would just use -output in the switch and remove the sed at the end
>>>> (keeping the extra arguments '-dll -quiet')
>>>> The hack of no space after -o was there just to simplify the sed at
>>>> the end of the script, now it is not needed.
>>>>
>>>>
>>>
>>> Ok, but I'll have to modify gcc (cil32.h) to make it pass -output to 'as'
>>> instead of -o. Do you really want me to introduce such a version
>>> dependency?
>>> (personnally, I don't mind).
>>>
>>>
>>>>
>>>> If you remove that part, you should cleanup also
>>>>
>>>>
>>>> http://clibinutils.svn.sourceforge.net/viewvc/clibinutils/mono-based-binutils/trunk/tools/gcc4cli/utils/CObject.cs?revision=12&view=markup
>>>> There is a hack to fixup an error on varargs methods generated by pnet
>>>> ilasm.
>>>> just remove fix_DOTGNU_Errors()
>>>>
>>>>
>>>
>>> Ok, I'll check that.
>>>
>>> Regards,
>>> Thierry.
>>>
>>>>
>>>> Andrea
>>>>
>>>> On Tue, Mar 23, 2010 at 11:34 AM, Thierry Lafage
>>>> <thi...@in...> wrote:
>>>>
>>>>
>>>>>
>>>>> Hi Andrea,
>>>>>
>>>>> Here is a simpler patch for as that is meant to be commited in
>>>>>
>>>>>
>>>>> https://clibinutils.svn.sourceforge.net/svnroot/clibinutils/mono-based-binutils/trunk
>>>>>
>>>>> Since the svn url says that it is a mono-based version, the test for
>>>>> Pnet
>>>>> ilasm is surely not worth anymore.
>>>>> I also added a space between -o and the ouput file in order not to
>>>>> confuse
>>>>> with potential -opt or -o<something> ilasm arguments. Otherwise,
>>>>> calling
>>>>> as
>>>>> with -opt gives: "ilasm /ouput:pt"
>>>>>
>>>>> Comparing to my former patch, I don't propose anymore the -O case
>>>>> because,
>>>>> since the short bytecode optimizations are done in ld, there is no more
>>>>> need
>>>>> for the moment to activate optimizations in the assembly pass.
>>>>>
>>>>> Tell me if it's ok for you, then I'll commit.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Thierry Lafage.
>>>>>
>>>>>
>>>>>
>>>>> Andrea Carlo Ornstein a écrit :
>>>>>
>>>>>
>>>>>>
>>>>>> At the beginning when I did the binutils there was a bug in the
>>>>>> release of ilasm.
>>>>>> Only the one from Mono trunk was working.
>>>>>> That is the only reason why the binutils could use also the Pnet ilasm
>>>>>> (You could avoid the build of ilasm from Mono trunk).
>>>>>>
>>>>>> If it is not the case anymore we can switch all to Mono tools and
>>>>>> remove the check in the script.
>>>>>>
>>>>>> Andrea
>>>>>>
>>>>>> On Mon, Mar 22, 2010 at 12:11 PM, Thierry Lafage
>>>>>> <thi...@in...> wrote:
>>>>>>
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> I forgot to mention that (Mono) ilasm does not complains (not even a
>>>>>>> warning) if you pass it an unknown option; so this should be ok with
>>>>>>> older version of ilasm.
>>>>>>> Regarding "Pnet ilasm", I don't know, but I've been told that this
>>>>>>> option would be abandonned. If this is really the case, then we
>>>>>>> should
>>>>>>> also remove the "if [ `$ILASM --version | ..." test. Tell me: I can
>>>>>>> do
>>>>>>> it in this commit.
>>>>>>>
>>>>>>> Regards,
>>>>>>>
>>>>>>> Thierry Lafage.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Thierry Lafage a écrit :
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Here is a patch for our binutils/as script which converts -O option
>>>>>>>> into forthcoming '-opt:ldc' ilasm option.
>>>>>>>>
>>>>>>>> Since the default behavior is not changed, I should commit soon if
>>>>>>>> nobody complains.
>>>>>>>>
>>>>>>>> Regards,
>>>>>>>> Thierry Lafage.
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Download Intel&#174; Parallel Studio Eval
>>>>>>>> Try the new software tools for yourself. Speed compiling, find bugs
>>>>>>>> proactively, and fine-tune applications for parallel performance.
>>>>>>>> See why Intel Parallel Studio got high marks during beta.
>>>>>>>> http://p.sf.net/sfu/intel-sw-dev
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Gcc4cli-devel mailing list
>>>>>>>> Gcc...@li...
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/gcc4cli-devel
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Download Intel&#174; Parallel Studio Eval
>>>>>>> Try the new software tools for yourself. Speed compiling, find bugs
>>>>>>> proactively, and fine-tune applications for parallel performance.
>>>>>>> See why Intel Parallel Studio got high marks during beta.
>>>>>>> http://p.sf.net/sfu/intel-sw-dev
>>>>>>> _______________________________________________
>>>>>>> Gcc4cli-devel mailing list
>>>>>>> Gcc...@li...
>>>>>>> https://lists.sourceforge.net/lists/listinfo/gcc4cli-devel
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>
>>>>> Index: scripts/as.in
>>>>> ===================================================================
>>>>> --- scripts/as.in       (revision 14)
>>>>> +++ scripts/as.in       (working copy)
>>>>> @@ -23,7 +23,7 @@
>>>>>         shift
>>>>>         ;;
>>>>>     -o)
>>>>> -          PARAM="$PARAM $1$2";
>>>>> +          PARAM="$PARAM $1 $2";
>>>>>         shift
>>>>>         shift
>>>>>         ;;
>>>>> @@ -34,13 +34,7 @@
>>>>>  esac
>>>>>  done
>>>>>
>>>>> -if test  "x`$ILASM --version | grep Mono | wc -l`" = "x0"; then
>>>>> -    #needed by Pnet ilasm
>>>>> -    PARAM=`echo "--format dll  $PARAM" | sed 's? -o? -o ?'`;
>>>>> -else
>>>>> -    #needed by Mono ilasm
>>>>> -    PARAM=`echo "-dll -quiet $PARAM" | sed 's? -o? -output:?'`;
>>>>> -fi
>>>>> +PARAM=`echo "-dll -quiet $PARAM" | sed 's? -o ? -output:?'`;
>>>>>
>>>>>  CMD="$ILASM $PARAM"
>>>>>  if test -n "$VERBOSE"; then
>>>>>
>>>>>
>>>>>
>>>>>
>
> Index: tools/gcc4cli/utils/CObject.cs
> ===================================================================
> --- tools/gcc4cli/utils/CObject.cs      (revision 14)
> +++ tools/gcc4cli/utils/CObject.cs      (working copy)
> @@ -15,8 +15,6 @@
>         {
>             MainType = Module.Types["<Module>"];
>             FillTables ();
> -            if (ref_types.Contains ("<ModuleExtern>"))
> -                fix_DOTGNU_Errors ();
>         }
>
>         internal CObject (string name, AssemblyResolver asm_resolver)
> @@ -30,30 +28,6 @@
>             Module.Types.Add (MainType);
>         }
>
> -        void fix_DOTGNU_Errors ()
> -        {
> -            foreach (MethodDefinition method in MainType.Methods) {
> -                foreach (Instruction instr in method.Body.Instructions) {
> -                    if (   instr.OpCode.OperandType ==
> OperandType.InlineMethod
> -                        || (instr.OpCode.OperandType ==
> OperandType.InlineTok && instr.Operand is MethodReference)) {
> -                        MethodReference mref = instr.Operand as
> MethodReference;
> -                        if (mref.DeclaringType.Name == "<ModuleExtern>") {
> -                            string m_name = mref.Name;
> -                            if (def_symbols.Contains (m_name)) {
> -                                mref.DeclaringType =
> ((MethodReference)def_symbols[m_name]).DeclaringType;
> -                            }
> -                            else if (private_symbols.Contains (m_name)) {
> -                                mref.DeclaringType =
> ((MethodReference)private_symbols[m_name]).DeclaringType;
> -                            }
> -                            else
> -                                throw new Exception ("Error in hack for
> DOTGNU");
> -                        }
> -                    }
> -                }
> -            }
> -            ref_types.Remove ("<ModuleExtern>");
> -        }
> -
>         static int renamer_id = 0;
>
>         bool MethodIsVarArg(MethodReference method)
> Index: scripts/as.in
> ===================================================================
> --- scripts/as.in       (revision 14)
> +++ scripts/as.in       (working copy)
> @@ -23,7 +23,7 @@
>           shift
>           ;;
>       -o)
> -          PARAM="$PARAM $1$2";
> +          PARAM="$PARAM -output:$2";
>           shift
>           shift
>           ;;
> @@ -34,13 +34,7 @@
>   esac
>  done
>
> -if test  "x`$ILASM --version | grep Mono | wc -l`" = "x0"; then
> -    #needed by Pnet ilasm
> -    PARAM=`echo "--format dll  $PARAM" | sed 's? -o? -o ?'`;
> -else
> -    #needed by Mono ilasm
> -    PARAM=`echo "-dll -quiet $PARAM" | sed 's? -o? -output:?'`;
> -fi
> +PARAM="-dll -quiet $PARAM"
>
>  CMD="$ILASM $PARAM"
>  if test -n "$VERBOSE"; then
>
>