Merge lp:~maddevelopers/mg5amcnlo/1.5.10 into lp:~madteam/mg5amcnlo/trunk

[Olivier Mattelaer]

[Olivier Mattelaer (olivier-mattelaer)]

I've made an auto-review and this looks ok (nothing really deep in fact but the order of the helas call).
Valentin could you at least run the tests?

Thanks,

Olivier

review: Approve

[lp:~maddevelopers/mg5amcnlo/1.5.10]

256. By Olivier Mattelaer on 2013-05-12

patch from Luran: Avoid crash when channel contributes less than 1e-100.

[Valentin Hirschi (valentin-hirschi)]

Hi Olivier,

I'll do my best for this review. Since there is nothing major, I went pretty fast on it but still noted the following:

a)

Depending on the c++ compiler, the include
"#include <iostream>"
is mandatory at the beginning of a C++ code in order to use the std:: namespace.
All cpp template have it, could you add it to the new txxxxx template routine too.

b)

This fermion flow flipping business is getting even worse with your necessary modifications. At some point, one will have to rethink that. One possibility would be for example to give up on the dynamical on-the-flight recycling optimisation and perform it only at the very end, after all helas_diagrams have been generated.

Anyway, for the present version the fix seems fine but I don't see any test related to this. Given that we know this part to be a very sensitive point of MG5, it would be worth to have the order of the wfs tested. Something like going through the list of wfs for various processes featuring majoranas and making sure that each wf finds its mothers in the wavefunction list placed before and having smaller nb_wf.

For example, when doing line 1060 to 1070 of helas_objects.py, you are not certain that the mothers of new_wf are before the index i in diagram_wavefunctions list or in previous diagrams in the wavefunctions list. I agree this should be true by construction, but one can never be too careful.

I looked superficially at the rest and it looks fine.

Cheers,

review: Needs Fixing

[Olivier Mattelaer (olivier-mattelaer)]

Hi Valentin,

Thanks so much

> a)
>
> Depending on the c++ compiler, the include
> "#include <iostream>"
> is mandatory at the beginning of a C++ code in order to use the std:: namespace.
> All cpp template have it, could you add it to the new txxxxx template routine too.

Sure will do.

> Anyway, for the present version the fix seems fine but I don't see any test related to this.

In fact I just kept the previous test, written by Johan who were checking those points and edit the
the tests according to the new ordering. So I think that this is fine.

Cheers,

Olivier

On May 14, 2013, at 6:23 AM, Valentin Hirschi <email address hidden> wrote:

> Review: Needs Fixing
>
> Hi Olivier,
>
> I'll do my best for this review. Since there is nothing major, I went pretty fast on it but still noted the following:
>
> a)
>
> Depending on the c++ compiler, the include
> "#include <iostream>"
> is mandatory at the beginning of a C++ code in order to use the std:: namespace.
> All cpp template have it, could you add it to the new txxxxx template routine too.
>
> b)
>
> This fermion flow flipping business is getting even worse with your necessary modifications. At some point, one will have to rethink that. One possibility would be for example to give up on the dynamical on-the-flight recycling optimisation and perform it only at the very end, after all helas_diagrams have been generated.
>
> Anyway, for the present version the fix seems fine but I don't see any test related to this. Given that we know this part to be a very sensitive point of MG5, it would be worth to have the order of the wfs tested. Something like going through the list of wfs for various processes featuring majoranas and making sure that each wf finds its mothers in the wavefunction list placed before and having smaller nb_wf.
>
> For example, when doing line 1060 to 1070 of helas_objects.py, you are not certain that the mothers of new_wf are before the index i in diagram_wavefunctions list or in previous diagrams in the wavefunctions list. I agree this should be true by construction, but one can never be too careful.
>
> I looked superficially at the rest and it looks fine.
>
> Cheers,
>
>
> --
> https://code.launchpad.net/~maddevelopers/madgraph5/1.5.10/+merge/163092
> You proposed lp:~maddevelopers/madgraph5/1.5.10 for merging.

[Valentin Hirschi (valentin-hirschi)]

Hi Olivier,

> Hi Valentin,
>
> Thanks so much
>
> > a)
> >
> > Depending on the c++ compiler, the include
> > "#include <iostream>"
> > is mandatory at the beginning of a C++ code in order to use the std::
> namespace.
> > All cpp template have it, could you add it to the new txxxxx template
> routine too.
>
> Sure will do.
>
> > Anyway, for the present version the fix seems fine but I don't see any test
> related to this.
>
>
> In fact I just kept the previous test, written by Johan who were checking
> those points and edit the
> the tests according to the new ordering. So I think that this is fine.
>

Sorry, I had overlooked the modification of the export_v4_test.
Yes, I'm looking forward to seeing such kind of tests ported to the IOTest framework as it will make their updates so much more handy.
The kind of tests I was referring to where more generic, e.g. not comparing the output vs an hardcoded one but really scanning through the helas wavefunction lists and checking that the order is consistent (all mothers precede their daughters).
But, indeed, I guess the existing 'hardcoded' test is good enough for now.

Cheers,

> Cheers,
>
> Olivier
>
> On May 14, 2013, at 6:23 AM, Valentin Hirschi <email address hidden>
> wrote:
>
> > Review: Needs Fixing
> >
> > Hi Olivier,
> >
> > I'll do my best for this review. Since there is nothing major, I went pretty
> fast on it but still noted the following:
> >
> > a)
> >
> > Depending on the c++ compiler, the include
> > "#include <iostream>"
> > is mandatory at the beginning of a C++ code in order to use the std::
> namespace.
> > All cpp template have it, could you add it to the new txxxxx template
> routine too.
> >
> > b)
> >
> > This fermion flow flipping business is getting even worse with your
> necessary modifications. At some point, one will have to rethink that. One
> possibility would be for example to give up on the dynamical on-the-flight
> recycling optimisation and perform it only at the very end, after all
> helas_diagrams have been generated.
> >
> > Anyway, for the present version the fix seems fine but I don't see any test
> related to this. Given that we know this part to be a very sensitive point of
> MG5, it would be worth to have the order of the wfs tested. Something like
> going through the list of wfs for various processes featuring majoranas and
> making sure that each wf finds its mothers in the wavefunction list placed
> before and having smaller nb_wf.
> >
> > For example, when doing line 1060 to 1070 of helas_objects.py, you are not
> certain that the mothers of new_wf are before the index i in
> diagram_wavefunctions list or in previous diagrams in the wavefunctions list.
> I agree this should be true by construction, but one can never be too careful.
> >
> > I looked superficially at the rest and it looks fine.
> >
> > Cheers,
> >
> >
> > --
> > https://code.launchpad.net/~maddevelopers/madgraph5/1.5.10/+merge/163092
> > You proposed lp:~maddevelopers/madgraph5/1.5.10 for merging.

review: Approve

[lp:~maddevelopers/mg5amcnlo/1.5.10]

257. By Olivier Mattelaer on 2013-05-15

Fix comment on c++ compilation issue reported by Valentin (thanks for that)

258. By Olivier Mattelaer on 2013-05-16

merge with lp:~maddevelopers/madgraph5/1.5.10_mod_enhancement_technique
which correct problem linked to GIM mechanism cancelation (creating problem
for assigning colorflow)

=== modified file 'Template/README.gridpack'
--- Template/README.gridpack	2011-07-11 18:32:26 +0000
+++ Template/README.gridpack	2013-05-16 12:23:29 +0000
@@ -77,7 +77,7 @@
 ./bin/compile
 is executed from the ./madevent directory. As input to specify the
 decays, files with the names
-input_x.in
+decay_x.in
 should be put in the ./DECAY directory. (Where x is a number between 1
 and 20.) These files are read automatically, and one file should be
 put for each decaying particle. (E.g. for ttbar events 2 files are

=== modified file 'Template/Source/PDF/pdg2pdf.f'
--- Template/Source/PDF/pdg2pdf.f	2013-02-26 03:56:16 +0000
+++ Template/Source/PDF/pdg2pdf.f	2013-05-16 12:23:29 +0000
@@ -62,9 +62,11 @@
       enddo
 
 c     Reuse previous result, if possible
-      if (ireuse.gt.0.and.pdflast(iporg,ireuse).ne.-99d9) then
-         pdg2pdf=pdflast(iporg,ireuse)
-         return 
+      if (ireuse.gt.0) then
+         if (pdflast(iporg,ireuse).ne.-99d9) then
+            pdg2pdf=pdflast(iporg,ireuse)
+            return 
+         endif
       endif
 
 c     Bjorken x and/or facrorization scale and/or PDF set are not

=== modified file 'UpdateNotes.txt'
--- UpdateNotes.txt	2013-04-01 15:47:25 +0000
+++ UpdateNotes.txt	2013-05-16 12:23:29 +0000
@@ -1,5 +1,28 @@
 Update notes for MadGraph 5 (in reverse time order)
 
+1.5.10 (xx/05/13) OM: Fix Bug #1170417: fix crash for conjugate routine in presence of 
+                      massless propagator. (introduce in 1.5.9)
+                  OM: Fix question #226810: checking that patch program exists before
+                      trying to update MG5 code.
+                  OM: Fix Bug #1171049: an error in the order of wavefunctions 
+                      making the code to crash (introduce in 1.5.7)
+		  OM: Allow to use an additional syntax for the set command.
+                      set gauge = Feynman is now valid. (Was not valid before due to the '=')
+                  OM: Fix By Arian Abrahantes. Fix SGE cluster which was not working when
+                      running full simulation (PGS/Delphes).
+                  OM: adding txxxxx.cc (Thanks to Aurelijus Rinkevicius for having 
+                      written the routine) 
+                  OM: Fix Bug #1177442. This crash occurs only for very large model. 
+                      None of the model shipped with MG5 are impacted.
+                  OM: Fix Question #228315. On some filesystem, some of the executable 
+                      loose the permission to be executable. Recover those errors 
+                      automatically.
+                  OM: Modify the diagram enhancement technique. When more diagram have 
+                      the same propagator structure we still combine them but we now include
+                      the interference term in the enhancement technique for those diagrams.
+                      This fix a crash for some multi-jet process in presence of non diagonal
+                      ckm matrices.
+
 1.5.9 (01/04/13)  JA: Fix bug in identification of symmetric diagrams, which could
                       give the wrong propagators included in event files for certain
                       processes (such as p p > z z j, z > l+ l-). Apart from the 

=== modified file 'aloha/create_aloha.py'
--- aloha/create_aloha.py	2013-03-18 23:03:14 +0000
+++ aloha/create_aloha.py	2013-05-16 12:23:29 +0000
@@ -724,12 +724,14 @@
         # reorganize the data (in order to use optimization for a given lorentz
         #structure
         request = {}
+
         for list_l_name, tag, outgoing in data:
             #allow tag to have integer for retro-compatibility
+            all_tag = tag[:]
             conjugate = [i for i in tag if isinstance(i, int)]
-            tag =  [i for i in tag if isinstance(i, str)]
-            tag = tag + ['C%s'%i for i in conjugate] 
-            #
+            tag =  [i for i in tag if isinstance(i, str) and not i.startswith('P')]
+            tag = tag + ['C%s'%i for i in conjugate]             
+            tag = tag + [i for i in all_tag if isinstance(i, str) and  i.startswith('P')] 
             
             conjugate = tuple([int(c[1:]) for c in tag if c.startswith('C')])
             loop = any((t.startswith('L') for t in tag))
@@ -784,8 +786,10 @@
                 continue
             #allow tag to have integer for retrocompatibility
             conjugate = [i for i in tag if isinstance(i, int)]
-            tag =  [i for i in tag if isinstance(i, str)]
+            all_tag = tag[:]
+            tag =  [i for i in tag if isinstance(i, str) and not i.startswith('P')]
             tag = tag + ['C%s'%i for i in conjugate] 
+            tag = tag + [i for i in all_tag if isinstance(i, str) and  i.startswith('P')] 
             
             if not self.explicit_combine:
                 lorentzname = list_l_name[0]

=== added file 'aloha/template_files/txxxxx.cc'
--- aloha/template_files/txxxxx.cc	1970-01-01 00:00:00 +0000
+++ aloha/template_files/txxxxx.cc	2013-05-16 12:23:29 +0000
@@ -0,0 +1,171 @@
+#include <complex>
+#include <cmath> 
+#include "aloha_aux_functions.h"
+using namespace std;
+#include <iostream>
+#include <cstdlib>
+
+void txxxxx(double p[4], double tmass, int nhel, int nst, complex<double> tc[18])
+{
+	complex<double> ft[6][4], ep[4], em[4], e0[4];
+	double pt, pt2, pp, pzpt, emp, sqh, sqs;
+	int i, j;
+	
+	sqh = pow( 0.5, 0.5 );
+	sqs = pow( 0.5/3, 0.5 );
+	
+	pt2 = p[1]*p[1] + p[2]*p[2];
+	pp = min( p[0], pow( pt2+p[3]*p[3], 0.5 ) );
+	pt = min( pp, pow( pt2, 0.5 ) );
+	
+	ft[4][0] = complex<double>( p[0]*nst, p[3]*nst );
+	ft[5][0] = complex<double>( p[1]*nst, p[2]*nst );
+	
+	// construct eps+
+	if( nhel >= 0 )
+	{
+		if( pp == 0 )
+		{
+			ep[0] = complex<double>( 0, 0 );
+			ep[1] = complex<double>( -sqh, 0 );
+			ep[2] = complex<double>( 0, nst*sqh );
+			ep[3] = complex<double>( 0, 0 );
+		}
+		else
+		{
+			ep[0] = complex<double>( 0, 0 );
+			ep[3] = complex<double>( pt/pp*sqh, 0 );
+			
+			if( pt != 0 )
+			{
+				pzpt = p[3]/(pp*pt)*sqh;
+				ep[1] = complex<double>( -p[1]*pzpt, -nst*p[2]/pt*sqh );
+				ep[2] = complex<double>( -p[2]*pzpt, nst*p[1]/pt*sqh );
+			}
+			else
+			{
+				ep[1] = complex<double>( -sqh, 0 );
+				ep[2] = complex<double>( 0, nst*Sgn(sqh,p[3]) );
+			}
+		}
+		
+	}
+	
+	// construct eps-
+	if( nhel <= 0 )
+	{
+		if( pp == 0 )
+		{
+			em[0] = complex<double>( 0, 0 );
+			em[1] = complex<double>( sqh, 0 );
+			em[2] = complex<double>( 0, nst*sqh );
+			em[3] = complex<double>( 0, 0 );
+		}
+		else
+		{
+			em[0] = complex<double>( 0, 0 );
+			em[3] = complex<double>( -pt/pp*sqh, 0 );
+			
+			if( pt != 0 )
+			{
+				pzpt = -p[3]/(pp*pt)*sqh;
+				em[1] = complex<double>( -p[1]*pzpt, -nst*p[2]/pt*sqh );
+				em[2] = complex<double>( -p[2]*pzpt,  nst*p[1]/pt*sqh );
+			}
+			else
+			{
+				em[1] = complex<double>( sqh, 0 );
+				em[2] = complex<double>( 0, nst*Sgn(sqh,p[3]) );
+			}
+		}
+	}
+	
+	// construct eps0
+	if( fabs(nhel) <= 1 )
+	{
+		if( pp == 0 )
+		{
+			e0[0] = complex<double>( 0, 0 );
+			e0[1] = complex<double>( 0, 0 );
+			e0[2] = complex<double>( 0, 0 );
+			e0[3] = complex<double>( 1, 0 );
+		}
+		else
+		{
+			emp = p[0]/(tmass*pp);
+			e0[0] = complex<double>( pp/tmass, 0 );
+			e0[3] = complex<double>( p[3]*emp, 0 );
+			
+			if( pt != 0 )
+			{
+				e0[1] = complex<double>( p[1]*emp, 0 );
+				e0[2] = complex<double>( p[2]*emp, 0 );
+			}
+			else
+			{
+				e0[1] = complex<double>( 0, 0 );
+				e0[2] = complex<double>( 0, 0 );
+			}
+		}
+	}
+	
+	if( nhel == 2 )
+	{
+		for( j=0; j<4; j++ )
+		{
+			for( i=0; i<4; i++ ) ft[i][j] = ep[i]*ep[j];
+		}
+	}
+	else if( nhel == -2 )
+	{
+		for( j=0; j<4; j++ )
+		{
+			for( i=0; i<4; i++ ) ft[i][j] = em[i]*em[j];
+		}
+	}
+	else if( tmass == 0 )
+	{
+		for( j=0; j<4; j++ )
+		{
+			for( i=0; i<4; i++ ) ft[i][j] = 0;
+		}
+	}
+	else if( tmass != 0 )
+	{
+		if( nhel == 1 )
+		{
+			for( j=0; j<4; j++ )
+			{
+				for( i=0; i<4; i++ ) ft[i][j] = sqh*( ep[i]*e0[j] + e0[i]*ep[j] );
+			}
+		}
+		else if( nhel == 0 )
+		{
+			for( j=0; j<4; j++ )
+			{
+				for( i=0; i<4; i++ ) ft[i][j] = sqs*( ep[i]*em[j] + em[i]*ep[j]
+					+ 2.0*e0[i]*e0[j] );
+			}
+		}
+		else if( nhel == -1 )
+		{
+			for( j=0; j<4; j++ )
+			{
+				for( i=0; i<4; i++ ) ft[i][j] = sqh*( em[i]*e0[j] + e0[i]*em[j] );
+			}
+		}
+		else
+		{
+			std::cerr << "Invalid helicity in txxxxx.\n";
+			std::exit(1);
+		}
+	}
+	
+	tc[0] = ft[4][0];
+	tc[1] = ft[5][0];
+	
+	for( j=0; j<4; j++ )
+	{
+		for( i=0; i<4; i++ ) tc[j*4+i+2] = ft[j][i];
+	}
+}

=== added file 'aloha/template_files/txxxxx.h'
--- aloha/template_files/txxxxx.h	1970-01-01 00:00:00 +0000
+++ aloha/template_files/txxxxx.h	2013-05-16 12:23:29 +0000
@@ -0,0 +1,5 @@
+#ifndef i_guard
+#define i_guard
+#include <complex>
+void txxxxx(double p[4],double tmass,int nhel,int nst,std::complex<double> fi[18]);
+#endif

=== modified file 'bin/create_release.py'
--- bin/create_release.py	2013-01-15 18:59:13 +0000
+++ bin/create_release.py	2013-05-16 12:23:29 +0000
@@ -193,7 +193,6 @@
             path.join(filepath, 'input','mg5_configuration.txt'))
 
 # 2. Create the automatic documentation in the apidoc directory
-
 try:
     status1 = subprocess.call(['epydoc', '--html', '-o', 'apidoc',
                                'madgraph', 'aloha',
@@ -211,7 +210,8 @@
 # 3. tar the MadGraph5_vVERSION directory.
 
 logging.info("Create the tar file " + filename)
-
+# clean all the pyc
+os.system("cd %s;find . -name '*.pyc' -delete" % filepath)
 status2 = subprocess.call(['tar', 'czf', filename, filepath])
 
 if status2:

=== modified file 'madgraph/VERSION'
--- madgraph/VERSION	2013-04-01 15:47:25 +0000
+++ madgraph/VERSION	2013-05-16 12:23:29 +0000
@@ -1,3 +1,3 @@
-version = 1.5.9
-date = 2013-04-01
+version = 1.5.10
+date = 2013-XX-XX
 

=== modified file 'madgraph/core/helas_objects.py'
--- madgraph/core/helas_objects.py	2013-03-27 22:59:57 +0000
+++ madgraph/core/helas_objects.py	2013-05-16 12:23:29 +0000
@@ -1044,7 +1044,30 @@
                         new_wf.set('number', old_wf.get('number'))
                     diagram_wavefunctions[old_wf_index] = new_wf
                 except ValueError:
-                    diagram_wavefunctions.append(new_wf)
+                    # Make sure that new_wf comes before any wavefunction
+                    # which has it as mother
+                    if len(self['mothers']) == 0:
+                        #insert at the beginning
+                        if diagram_wavefunctions:
+                            wf_nb = diagram_wavefunctions[0].get('number')
+                            for w in diagram_wavefunctions:
+                                w.set('number', w.get('number') + 1)
+                            new_wf.set('number', wf_nb)
+                            diagram_wavefunctions.insert(0, new_wf)
+                        else:
+                            diagram_wavefunctions.insert(0, new_wf)
+                    else:
+                        for i, wf in enumerate(diagram_wavefunctions):
+                            if self in wf.get('mothers'):
+                                # Update wf numbers
+                                new_wf.set('number', wf.get('number'))
+                                for w in diagram_wavefunctions[i:]:
+                                    w.set('number', w.get('number') + 1)
+                                # Insert wavefunction
+                                diagram_wavefunctions.insert(i, new_wf)
+                                break
+                        else:
+                            diagram_wavefunctions.append(new_wf)
 
             # Set new mothers
             new_wf.set('mothers', mothers)
@@ -3697,12 +3720,12 @@
         all lorentz structures used by this HelasMatrixElement."""
 
         out = [(tuple(wa.get('lorentz')),  
-                 tuple(list(wa.get('conjugate_indices')) + \
-                  [] if wa.get('particle').get('propagator') =='' else \
-                  ['P%s' % wa.get('particle').get('propagator')]),
-                 wa.find_outgoing_number()) for wa in \
-                self.get_all_wavefunctions()\
-                if wa.get('interaction_id') != 0]  
+                tuple(([] if wa.get('particle').get('propagator') =='' else \
+                           ['P%s' % wa.get('particle').get('propagator')]) + \
+                           list(wa.get('conjugate_indices'))), \
+                wa.find_outgoing_number()) 
+                for wa in self.get_all_wavefunctions() \
+                                               if wa.get('interaction_id') != 0] 
         out += [(tuple(wa.get('lorentz')),  wa.get('conjugate_indices'),
                  wa.find_outgoing_number()) for wa in  self.get_all_amplitudes() \
                 if wa.get('interaction_id') != 0]

=== modified file 'madgraph/interface/madgraph_interface.py'
--- madgraph/interface/madgraph_interface.py	2013-03-30 04:53:23 +0000
+++ madgraph/interface/madgraph_interface.py	2013-05-16 12:23:29 +0000
@@ -873,6 +873,9 @@
 
         if len(args) == 1 and args[0] == 'complex_mass_scheme':
             args.append('True')
+
+        if len(args) > 2 and '=' == args[1]:
+            args.pop(1)
         
         if len(args) < 2:
             self.help_set()
@@ -3237,6 +3240,13 @@
                 raise self.ConfigurationError(error_text)
             return 
         
+        if not misc.which('patch'):
+            error_text = """Not able to find program \'patch\'. Please reload a clean version
+            or install that program and retry."""
+            if mode == 'userrequest':
+                raise self.ConfigurationError(error_text)
+            return            
+        
         
         # read the data present in .autoupdate
         data = {}

=== modified file 'madgraph/iolibs/export_v4.py'
--- madgraph/iolibs/export_v4.py	2013-03-27 15:33:27 +0000
+++ madgraph/iolibs/export_v4.py	2013-05-16 12:23:29 +0000
@@ -121,7 +121,7 @@
                                  '--web'], cwd=self.dir_path)
             else:
                 try:
-                    subprocess.call([pjoin('bin', 'internal', 'clean_template')], \
+                    misc.call([pjoin('bin', 'internal', 'clean_template')], \
                                                                        cwd=self.dir_path)
                 except Exception, why:
                     raise MadGraph5Error('Failed to clean correctly %s: \n %s' \
@@ -554,10 +554,16 @@
                 line = "AMP2(%(num)d)=AMP2(%(num)d)+" % \
                        {"num": (config_to_diag_dict[config][0] + 1)}
 
-                line += "+".join(["AMP(%(num)d)*dconjg(AMP(%(num)d))" % \
-                                  {"num": a.get('number')} for a in \
+                amp = "+".join(["AMP(%(num)d)" % {"num": a.get('number')} for a in \
                                   sum([diagrams[idiag].get('amplitudes') for \
                                        idiag in config_to_diag_dict[config]], [])])
+                
+                # Not using \sum |M|^2 anymore since this creates troubles
+                # when ckm is not diagonal due to the JIM mechanism.
+                if '+' in amp:
+                    line += "(%s)*dconjg(%s)" % (amp, amp)
+                else:
+                    line += "%s*dconjg(%s)" % (amp, amp)
                 ret_lines.append(line)
         else:
             for idiag, diag in enumerate(matrix_element.get('diagrams')):
@@ -3160,7 +3166,13 @@
                                             self.p_to_f.parse(param.expr)))
 
         fsock.write_comments("\nDefinition of the EW coupling used in the write out of aqed\n")
-        fsock.writelines(""" gal(1) = 1d0
+        if ('aEWM1',) in self.model['parameters']:
+            fsock.writelines(""" gal(1) = 3.5449077018110318 / DSQRT(aEWM1)
+                                 gal(2) = 1d0
+                         """)            
+        else:
+            logger.warning('$RED aEWM1 not define in MODEL. AQED will not be written correcty in LHE FILE')
+            fsock.writelines(""" gal(1) = 1d0
                              gal(2) = 1d0
                          """)
 

=== modified file 'madgraph/iolibs/template_files/cpp_hel_amps_cc.inc'
--- madgraph/iolibs/template_files/cpp_hel_amps_cc.inc	2011-03-27 17:13:28 +0000
+++ madgraph/iolibs/template_files/cpp_hel_amps_cc.inc	2013-05-16 12:23:29 +0000
@@ -4,6 +4,11 @@
 //==========================================================================
 
 #include "HelAmps_%(model_name)s.h"
+#include <complex>
+#include <cmath>
+#include <iostream>
+#include <cstdlib>
+using namespace std;
 
 namespace %(namespace)s_%(model_name)s {
 

=== modified file 'madgraph/various/cluster.py'
--- madgraph/various/cluster.py	2013-03-30 19:28:32 +0000
+++ madgraph/various/cluster.py	2013-05-16 12:23:29 +0000
@@ -559,6 +559,9 @@
             stderr = '/dev/null'
         elif stderr == -2: # -2 is subprocess.STDOUT
             stderr = stdout
+        else:
+            stderr = self.def_get_path(stderr)
+            
         if log is None:
             log = '/dev/null'
         else:

=== modified file 'madgraph/various/sum_html.py'
--- madgraph/various/sum_html.py	2012-10-03 15:25:35 +0000
+++ madgraph/various/sum_html.py	2013-05-16 12:23:29 +0000
@@ -60,8 +60,9 @@
                     try:
                         return float(d)
                     except ValueError:
-                        if re.search(r'''[+-]?[\d.]*-\d*''', d):
-                            return 0.0
+                        m=re.search(r'''([+-]?[\d.]*)([+-]\d*)''', d)
+                        if m:
+                            return float(m.group(1))*10**(float(m.group(2)))
                         return 
                     
                 data = [secure_float(d) for d in line.split()]

=== modified file 'models/import_ufo.py'
--- models/import_ufo.py	2013-03-18 23:03:14 +0000
+++ models/import_ufo.py	2013-05-16 12:23:29 +0000
@@ -1221,39 +1221,53 @@
         if simplify:
             # check if the parameters is still usefull:
             re_str = '|'.join(special_parameters)
-            re_pat = re.compile(r'''\b(%s)\b''' % re_str)
-            used = set()
-            # check in coupling
-            for name, coupling_list in self['couplings'].items():
-                for coupling in coupling_list:
-                    for use in  re_pat.findall(coupling.expr):
-                        used.add(use)
+            if len(re_str) > 25000: # size limit on mac
+                split = len(special_parameters) // 2
+                re_str = ['|'.join(special_parameters[:split]),
+                          '|'.join(special_parameters[split:])]
+            else:
+                re_str = [ re_str ]
+            for expr in re_str:
+                re_pat = re.compile(r'''\b(%s)\b''' % expr)
+                used = set()
+                # check in coupling
+                for name, coupling_list in self['couplings'].items():
+                    for coupling in coupling_list:
+                        for use in  re_pat.findall(coupling.expr):
+                            used.add(use)
         else:
             used = set([i for i in special_parameters if i])
         
         
         # simplify the regular expression
-        re_str = '|'.join([param for param in special_parameters 
-                                                      if param not in used])
-        re_pat = re.compile(r'''\b(%s)\b''' % re_str)            
-        param_info = {}
-        # check in parameters
-        for dep, param_list in self['parameters'].items():
-            for tag, parameter in enumerate(param_list):
-                # update information concerning zero/one parameters
-                if parameter.name in special_parameters:
-                    param_info[parameter.name]= {'dep': dep, 'tag': tag, 
-                                                           'obj': parameter}
-                    continue
-                                    
-                # Bypass all external parameter
-                if isinstance(parameter, base_objects.ParamCardVariable):
-                    continue
-
-                # check the presence of zero/one parameter
-                if simplify:
-                    for use in  re_pat.findall(parameter.expr):
-                        used.add(use)
+        re_str = '|'.join([param for param in special_parameters if param not in used])
+        if len(re_str) > 25000: # size limit on mac
+            split = len(special_parameters) // 2
+            re_str = ['|'.join(special_parameters[:split]),
+                          '|'.join(special_parameters[split:])]
+        else:
+            re_str = [ re_str ]
+        for expr in re_str:                                                      
+            re_pat = re.compile(r'''\b(%s)\b''' % expr)
+               
+            param_info = {}
+            # check in parameters
+            for dep, param_list in self['parameters'].items():
+                for tag, parameter in enumerate(param_list):
+                    # update information concerning zero/one parameters
+                    if parameter.name in special_parameters:
+                        param_info[parameter.name]= {'dep': dep, 'tag': tag, 
+                                                               'obj': parameter}
+                        continue
+                                        
+                    # Bypass all external parameter
+                    if isinstance(parameter, base_objects.ParamCardVariable):
+                        continue
+    
+                    # check the presence of zero/one parameter
+                    if simplify:
+                        for use in  re_pat.findall(parameter.expr):
+                            used.add(use)
                         
         # modify the object for those which are still used
         for param in used:

=== modified file 'tests/unit_tests/iolibs/test_export_v4.py'
--- tests/unit_tests/iolibs/test_export_v4.py	2013-03-19 08:18:53 +0000
+++ tests/unit_tests/iolibs/test_export_v4.py	2013-05-16 12:23:29 +0000
@@ -5277,7 +5277,7 @@
 
         self.assertEqual(helas_call_writers.FortranHelasCallWriter(mybasemodel).\
                                    get_matrix_element_calls(matrix_element),
-                         """CALL VXXXXX(P(0,1),ZERO,NHEL(1),-1*IC(1),W(1,1))
+                                   """CALL VXXXXX(P(0,1),ZERO,NHEL(1),-1*IC(1),W(1,1))
 CALL IXXXXX(P(0,2),BMASS,NHEL(2),+1*IC(2),W(1,2))
 CALL OXXXXX(P(0,3),TMASS,NHEL(3),+1*IC(3),W(1,3))
 CALL OXXXXX(P(0,4),TMASS,NHEL(4),+1*IC(4),W(1,4))
@@ -5290,26 +5290,26 @@
 CALL FSIXXX(W(1,5),W(1,2),GT1GOM,MGO,WGO,W(1,7))
 # Amplitude(s) for diagram number 1
 CALL IOSXXX(W(1,7),W(1,4),W(1,9),GT1GOP,AMP(1))
-CALL IXXXXX(P(0,7),MX1,NHEL(7),-1*IC(7),W(1,7))
-CALL HIOCXX(W(1,7),W(1,4),GB1X1P,MB1,WB1,W(1,10))
-CALL FSIXXX(W(1,8),W(1,2),GT1X1M,MX1,WX1,W(1,11))
-CALL OXXXXX(P(0,5),TMASS,NHEL(5),+1*IC(5),W(1,12))
+CALL OXXXXX(P(0,5),TMASS,NHEL(5),+1*IC(5),W(1,7))
+CALL IXXXXX(P(0,7),MX1,NHEL(7),-1*IC(7),W(1,10))
+CALL HIOCXX(W(1,10),W(1,4),GB1X1P,MB1,WB1,W(1,11))
+CALL FSIXXX(W(1,8),W(1,2),GT1X1M,MX1,WX1,W(1,12))
 # Amplitude(s) for diagram number 2
-CALL IOSCXX(W(1,11),W(1,12),W(1,10),GB1X1M,AMP(2))
-CALL OXXXXX(P(0,2),BMASS,NHEL(2),-1*IC(2),W(1,11))
-CALL FVOCXX(W(1,11),W(1,1),GG,BMASS,ZERO,W(1,13))
-CALL FSOCXX(W(1,13),W(1,10),GB1GOM,MGO,WGO,W(1,11))
+CALL IOSCXX(W(1,12),W(1,7),W(1,11),GB1X1M,AMP(2))
+CALL OXXXXX(P(0,2),BMASS,NHEL(2),-1*IC(2),W(1,12))
+CALL FVOCXX(W(1,12),W(1,1),GG,BMASS,ZERO,W(1,13))
+CALL FSOCXX(W(1,13),W(1,11),GB1GOM,MGO,WGO,W(1,12))
 # Amplitude(s) for diagram number 3
-CALL IOSXXX(W(1,5),W(1,11),W(1,2),GT1GOM,AMP(3))
-CALL HIOCXX(W(1,7),W(1,3),GB1X1P,MB1,WB1,W(1,11))
-CALL HIOXXX(W(1,6),W(1,4),GT1N1P,MT1,WT1,W(1,7))
-CALL FSOCXX(W(1,13),W(1,11),GB1GOM,MGO,WGO,W(1,6))
+CALL IOSXXX(W(1,5),W(1,12),W(1,2),GT1GOM,AMP(3))
+CALL HIOCXX(W(1,10),W(1,3),GB1X1P,MB1,WB1,W(1,12))
+CALL HIOXXX(W(1,6),W(1,4),GT1N1P,MT1,WT1,W(1,10))
+CALL FSOCXX(W(1,13),W(1,12),GB1GOM,MGO,WGO,W(1,6))
 # Amplitude(s) for diagram number 4
-CALL IOSXXX(W(1,5),W(1,6),W(1,7),GT1GOM,AMP(4))
-CALL FSIXXX(W(1,8),W(1,7),GT1X1M,MX1,WX1,W(1,6))
+CALL IOSXXX(W(1,5),W(1,6),W(1,10),GT1GOM,AMP(4))
+CALL FSIXXX(W(1,8),W(1,10),GT1X1M,MX1,WX1,W(1,6))
 # Amplitude(s) for diagram number 5
-CALL IOSCXX(W(1,6),W(1,12),W(1,11),GB1X1M,AMP(5))
-CALL FSIXXX(W(1,5),W(1,7),GT1GOM,MGO,WGO,W(1,6))
+CALL IOSCXX(W(1,6),W(1,7),W(1,12),GB1X1M,AMP(5))
+CALL FSIXXX(W(1,5),W(1,10),GT1GOM,MGO,WGO,W(1,6))
 # Amplitude(s) for diagram number 6
 CALL IOSXXX(W(1,6),W(1,3),W(1,9),GT1GOP,AMP(6))""".split('\n'))
         
@@ -5439,7 +5439,7 @@
         # account)
         self.assertEqual(helas_call_writers.FortranHelasCallWriter(mybasemodel).\
                                    get_matrix_element_calls(matrix_element),
-                         """CALL OXXXXX(P(0,1),me,NHEL(1),-1*IC(1),W(1,1))
+        """CALL OXXXXX(P(0,1),me,NHEL(1),-1*IC(1),W(1,1))
 CALL IXXXXX(P(0,2),me,NHEL(2),+1*IC(2),W(1,2))
 CALL SXXXXX(P(0,3),+1*IC(3),W(1,3))
 CALL SXXXXX(P(0,4),+1*IC(4),W(1,4))
@@ -5450,25 +5450,25 @@
 CALL HIOXXX(W(1,5),W(1,7),MGVX494,Msl2,Wsl2,W(1,9))
 # Amplitude(s) for diagram number 1
 CALL IOSXXX(W(1,8),W(1,6),W(1,9),MGVX350,AMP(1))
-CALL IXXXXX(P(0,1),me,NHEL(1),+1*IC(1),W(1,9))
-CALL FSICXX(W(1,9),W(1,3),MGVX350,Mneu1,Wneu1,W(1,10))
-CALL HIOXXX(W(1,10),W(1,6),MGVX350,Msl2,Wsl2,W(1,9))
-CALL OXXXXX(P(0,2),me,NHEL(2),-1*IC(2),W(1,10))
-CALL FSOCXX(W(1,10),W(1,4),MGVX494,Mneu1,Wneu1,W(1,11))
+CALL OXXXXX(P(0,2),me,NHEL(2),-1*IC(2),W(1,9))
+CALL IXXXXX(P(0,1),me,NHEL(1),+1*IC(1),W(1,10))
+CALL FSICXX(W(1,10),W(1,3),MGVX350,Mneu1,Wneu1,W(1,11))
+CALL HIOXXX(W(1,11),W(1,6),MGVX350,Msl2,Wsl2,W(1,10))
+CALL FSOCXX(W(1,9),W(1,4),MGVX494,Mneu1,Wneu1,W(1,11))
 # Amplitude(s) for diagram number 2
-CALL IOSXXX(W(1,5),W(1,11),W(1,9),MGVX494,AMP(2))
-CALL FSIXXX(W(1,5),W(1,4),MGVX494,Mneu1,Wneu1,W(1,10))
+CALL IOSXXX(W(1,5),W(1,11),W(1,10),MGVX494,AMP(2))
+CALL FSIXXX(W(1,5),W(1,4),MGVX494,Mneu1,Wneu1,W(1,9))
 CALL HIOXXX(W(1,2),W(1,7),MGVX494,Msl2,Wsl2,W(1,12))
 # Amplitude(s) for diagram number 3
-CALL IOSXXX(W(1,10),W(1,6),W(1,12),MGVX350,AMP(3))
+CALL IOSXXX(W(1,9),W(1,6),W(1,12),MGVX350,AMP(3))
 CALL OXXXXX(P(0,5),me,NHEL(5),+1*IC(5),W(1,12))
 CALL FSOCXX(W(1,12),W(1,4),MGVX494,Mneu1,Wneu1,W(1,7))
 # Amplitude(s) for diagram number 4
-CALL IOSXXX(W(1,2),W(1,7),W(1,9),MGVX494,AMP(4))
-CALL FSOXXX(W(1,6),W(1,3),MGVX350,Mneu1,Wneu1,W(1,9))
+CALL IOSXXX(W(1,2),W(1,7),W(1,10),MGVX494,AMP(4))
+CALL FSOXXX(W(1,6),W(1,3),MGVX350,Mneu1,Wneu1,W(1,10))
 CALL HIOXXX(W(1,8),W(1,1),MGVX350,Msl2,Wsl2,W(1,6))
 # Amplitude(s) for diagram number 5
-CALL IOSXXX(W(1,5),W(1,9),W(1,6),MGVX494,AMP(5))
+CALL IOSXXX(W(1,5),W(1,10),W(1,6),MGVX494,AMP(5))
 CALL IXXXXX(P(0,6),me,NHEL(6),-1*IC(6),W(1,6))
 CALL FSICXX(W(1,6),W(1,3),MGVX350,Mneu1,Wneu1,W(1,8))
 CALL HIOXXX(W(1,8),W(1,1),MGVX350,Msl2,Wsl2,W(1,6))
@@ -5476,9 +5476,9 @@
 CALL IOSXXX(W(1,5),W(1,11),W(1,6),MGVX494,AMP(6))
 # Amplitude(s) for diagram number 7
 CALL IOSXXX(W(1,2),W(1,7),W(1,6),MGVX494,AMP(7))
-CALL HIOXXX(W(1,10),W(1,1),MGVX350,Msl2,Wsl2,W(1,6))
+CALL HIOXXX(W(1,9),W(1,1),MGVX350,Msl2,Wsl2,W(1,6))
 # Amplitude(s) for diagram number 8
-CALL IOSXXX(W(1,2),W(1,9),W(1,6),MGVX494,AMP(8))""".split('\n'))
+CALL IOSXXX(W(1,2),W(1,10),W(1,6),MGVX494,AMP(8))""".split('\n'))
 
         # Test find_outgoing_number
         goal_numbers = [1, 2, 3, 2, 3, 1, 2, 3, 1, 1, 3, 2, 3, 3]
@@ -8562,13 +8562,13 @@
 CALL HIOXXX(W(1,1),W(1,5),MGVX494,Msl2,Wsl2,W(1,3))
 # Amplitude(s) for diagram number 1
 CALL IOSXXX(W(1,7),W(1,2),W(1,3),MGVX350,AMP(1))
+CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,3))
+CALL FSICXX(W(1,3),W(1,4),MGVX350,Mneu1,Wneu1,W(1,7))
 CALL OXXXXX(P(0,5),zero,NHEL(5),+1*IC(5),W(1,3))
-CALL FSOXXX(W(1,3),W(1,6),MGVX350,Mneu1,Wneu1,W(1,7))
-CALL HIOXXX(W(1,1),W(1,7),MGVX494,Msl2,Wsl2,W(1,3))
-CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,7))
-CALL FSICXX(W(1,7),W(1,4),MGVX350,Mneu1,Wneu1,W(1,1))
+CALL FSOXXX(W(1,3),W(1,6),MGVX350,Mneu1,Wneu1,W(1,4))
+CALL HIOXXX(W(1,1),W(1,4),MGVX494,Msl2,Wsl2,W(1,3))
 # Amplitude(s) for diagram number 2
-CALL IOSXXX(W(1,1),W(1,2),W(1,3),MGVX350,AMP(2))""".split('\n'))
+CALL IOSXXX(W(1,7),W(1,2),W(1,3),MGVX350,AMP(2))""".split('\n'))
 
         exporter = export_v4.ProcessExporterFortranME()
 
@@ -8615,13 +8615,13 @@
 CALL HIOXXX(W(1,1),W(1,5),MGVX494,Msl2,Wsl2,W(1,3))
 # Amplitude(s) for diagram number 1
 CALL IOSXXX(W(1,7),W(1,2),W(1,3),MGVX350,AMP(1))
+CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,3))
+CALL FSICXX(W(1,3),W(1,4),MGVX350,Mneu1,Wneu1,W(1,7))
 CALL OXXXXX(P(0,5),zero,NHEL(5),+1*IC(5),W(1,3))
-CALL FSOCXX(W(1,3),W(1,6),MGVX494,Mneu1,Wneu1,W(1,7))
-CALL HIOXXX(W(1,1),W(1,7),MGVX494,Msl2,Wsl2,W(1,3))
-CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,7))
-CALL FSICXX(W(1,7),W(1,4),MGVX350,Mneu1,Wneu1,W(1,1))
+CALL FSOCXX(W(1,3),W(1,6),MGVX494,Mneu1,Wneu1,W(1,4))
+CALL HIOXXX(W(1,1),W(1,4),MGVX494,Msl2,Wsl2,W(1,3))
 # Amplitude(s) for diagram number 2
-CALL IOSXXX(W(1,1),W(1,2),W(1,3),MGVX350,AMP(2))""".split('\n'))
+CALL IOSXXX(W(1,7),W(1,2),W(1,3),MGVX350,AMP(2))""".split('\n'))
 
         self.assertEqual(exporter.get_JAMP_lines(me)[0],
                          "JAMP(1)=+AMP(1)-AMP(2)")
@@ -8662,7 +8662,7 @@
 
         # This has been checked against v4
         self.assertEqual(myfortranmodel.get_matrix_element_calls(me),
-                         """CALL IXXXXX(P(0,1),zero,NHEL(1),+1*IC(1),W(1,1))
+        """CALL IXXXXX(P(0,1),zero,NHEL(1),+1*IC(1),W(1,1))
 CALL OXXXXX(P(0,2),zero,NHEL(2),-1*IC(2),W(1,2))
 CALL OXXXXX(P(0,3),zero,NHEL(3),+1*IC(3),W(1,3))
 CALL SXXXXX(P(0,4),+1*IC(4),W(1,4))
@@ -8688,24 +8688,24 @@
 CALL IOSXXX(W(1,11),W(1,2),W(1,3),MGVX350,AMP(3))
 # Amplitude(s) for diagram number 4
 CALL IOSXXX(W(1,12),W(1,2),W(1,3),MGVX350,AMP(4))
-CALL OXXXXX(P(0,6),zero,NHEL(6),+1*IC(6),W(1,3))
-CALL FVOXXX(W(1,3),W(1,9),MGVX12,zero,zero,W(1,12))
-CALL FSOXXX(W(1,12),W(1,8),MGVX350,Mneu1,Wneu1,W(1,9))
+CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,3))
+CALL FVICXX(W(1,3),W(1,5),MGVX12,zero,zero,W(1,12))
+CALL FSICXX(W(1,12),W(1,4),MGVX350,Mneu1,Wneu1,W(1,5))
+CALL OXXXXX(P(0,6),zero,NHEL(6),+1*IC(6),W(1,12))
+CALL FVOXXX(W(1,12),W(1,9),MGVX12,zero,zero,W(1,4))
+CALL FSOXXX(W(1,4),W(1,8),MGVX350,Mneu1,Wneu1,W(1,9))
+CALL HIOXXX(W(1,1),W(1,9),MGVX494,Msl2,Wsl2,W(1,4))
+# Amplitude(s) for diagram number 5
+CALL IOSXXX(W(1,5),W(1,2),W(1,4),MGVX350,AMP(5))
+CALL FSOXXX(W(1,12),W(1,7),MGVX350,Mneu1,Wneu1,W(1,9))
 CALL HIOXXX(W(1,1),W(1,9),MGVX494,Msl2,Wsl2,W(1,12))
-CALL IXXXXX(P(0,3),zero,NHEL(3),-1*IC(3),W(1,9))
-CALL FVICXX(W(1,9),W(1,5),MGVX12,zero,zero,W(1,8))
-CALL FSICXX(W(1,8),W(1,4),MGVX350,Mneu1,Wneu1,W(1,5))
-# Amplitude(s) for diagram number 5
-CALL IOSXXX(W(1,5),W(1,2),W(1,12),MGVX350,AMP(5))
-CALL FSOXXX(W(1,3),W(1,7),MGVX350,Mneu1,Wneu1,W(1,8))
-CALL HIOXXX(W(1,1),W(1,8),MGVX494,Msl2,Wsl2,W(1,3))
 # Amplitude(s) for diagram number 6
-CALL IOSXXX(W(1,5),W(1,2),W(1,3),MGVX350,AMP(6))
-CALL FSICXX(W(1,9),W(1,10),MGVX350,Mneu1,Wneu1,W(1,5))
+CALL IOSXXX(W(1,5),W(1,2),W(1,12),MGVX350,AMP(6))
+CALL FSICXX(W(1,3),W(1,10),MGVX350,Mneu1,Wneu1,W(1,5))
 # Amplitude(s) for diagram number 7
-CALL IOSXXX(W(1,5),W(1,2),W(1,12),MGVX350,AMP(7))
+CALL IOSXXX(W(1,5),W(1,2),W(1,4),MGVX350,AMP(7))
 # Amplitude(s) for diagram number 8
-CALL IOSXXX(W(1,5),W(1,2),W(1,3),MGVX350,AMP(8))""".split('\n'))
+CALL IOSXXX(W(1,5),W(1,2),W(1,12),MGVX350,AMP(8))""".split('\n'))
 
         # Test amp2 lines        
         amp2_lines = \