compilation error
Hi Ma5 Team,
Upon launching bin/ma5 I get:
Platform: Linux 3.2.0-4-486 [Linux mode]
Reading user settings ...
Checking mandatory packages:
- python [OK]
- python library: numpy [OK]
- g++ [OK]
- GNU Make [OK]
- Root [OK]
- PyRoot libraries [OK]
Checking optional packages:
- pdflatex [DISABLED]
** WARNING: pdflatex disabled. Reports under the pdf format will not be compiled.
- latex [DISABLED]
** WARNING: latex disabled. Reports under the dvi format will not be compiled.
- dvipdf [OK]
- zlib [OK]
- FastJet [OK]
- Delphes [DISABLED]
** WARNING: Header file called '/modules/
** WARNING: Delphes ROOT format will be disabled.
** WARNING: To enable this format, please type 'install delphes'.
- Delphes-MA5tune [DISABLED]
** WARNING: DelphesMA5tune folder not found.
** WARNING: Delphes-MA5tune ROOT format will be disabled.
** WARNING: To enable this format, please type 'install delphesMA5tune'.
- Recasting tools [DISABLED]
** WARNING: Recasting tools are not found and will be disabled.
** WARNING: To enable this functionnality, please type 'install RecastingTools'.
Checking the MadAnalysis library:
=> First use of MadAnalysis (or the library is missing).
****
****
How many cores for the compiling? default = max = 1
Answer: 1
=> Number of cores used for the compilation = 1
Writing the setup files ...
Writing all the Makefiles ...
****
Component 1/8 - library: SampleAnalyzer commons
- Cleaning the project before building the library ...
- Compiling the source files ...
** ERROR: impossible to compile the project. For more details, see the log file:
** ERROR: /home/jonathan/
** ERROR: The library building aborted.
The log file states:
make: *** empty string invalid as file name. Stop.
I'm using:
Ma5 1.1.11
root 5.34
gcc 4.7.2
All the neccessary perequisites are fullfilled so the problem seems to be within the ma5 code.
Please tell me how to fix this issue in order for ma5 to compile correctly.
thanks,
jonathan
Question information
- Language:
- English Edit question
- Status:
- Solved
- Assignee:
- No assignee Edit question
- Solved by:
- jcohen
- Solved:
- Last query:
- Last reply:
Revision history for this message
|
#1 |
Hi Jonathan
This is weird. Could you please try with v.1.1.12beta and tell me whether the problem persists?
Cheers,
Benjamin
Revision history for this message
|
#2 |
Hi Benjamin,
First of all thank you for the prompt response.
Unfortunately the problem persists in v.1.1.12beta.
What do you suggest?
jonathan
Revision history for this message
|
#3 |
Hi Jonathan,
Is this the only information available within the logfile? Could you please send me the associated Makefile (the one present in the same directory as the log file) as well as the full log file? I have so far no idea about the problem.
Thanks in advance,
Benjamin
Revision history for this message
|
#4 |
Hi Benjamin,
Yes this is the only information in the logfile. Here the Makefile as you requested:
#######
# MAKEFILE DEVOTED TO SAMPLEANALYZER COMMONS #
#######
# Compilers
CXX = g++
# C++ Compilation options
CXXFLAGS = -Wall -O3 -fPIC -I$(MA5_
CXXFLAGS += $(shell root-config --cflags)
CXXFLAGS += -DROOT_USE
# Linking options
LIBFLAGS =
LIBFLAGS += $(shell root-config --libs) -lEG
# Files
SRCS = $(wildcard */*.cpp)
HDRS = $(wildcard */*.h)
OBJS = $(SRCS:.cpp=.o)
# Name of the library
LIBRARY = libcommons_
# Defining colours
GREEN = "\\033[1;32m"
RED = "\\033[1;31m"
PINK = "\\033[1;35m"
BLUE = "\\033[1;34m"
YELLOW = "\\033[1;33m"
CYAN = "\\033[1;36m"
NORMAL = "\\033[0;39m"
# All target
all: header compile_header compile link_header link
# Header target
header:
@echo -e $(YELLOW)
@echo -e " Building SampleAnalyzer commons "
@echo -e "------
# Compile_header target
compile_header:
@echo -e $(YELLOW)
@echo -e " Compilation "
@echo -e "------
# Link_header target
link_header:
@echo -e $(YELLOW)
@echo -e " Linking "
@echo -e "------
# clean_header target
clean_header:
@echo -e $(YELLOW)
@echo -e " Removing intermediate files from building "
@echo -e "------
# mrproper_header target
mrproper_header:
@echo -e $(YELLOW)
@echo -e " Cleaning all the project "
@echo -e "------
# Precompile target
precompile:
# Compile target
compile: precompile $(OBJS)
# Compile each file
%.o: %.cpp $(HDRS)
$(CXX) $(CXXFLAGS) -o $@ -c $<
# Link target
link: $(OBJS)
$(CXX) -shared -o ../Lib/$(LIBRARY) $(OBJS) $(LIBFLAGS)
# Phony target
.PHONY: do_clean header compile_header link_header
# Clean target
clean: clean_header do_clean
# Do clean target
do_clean:
@rm -f $(OBJS)
# Mr Proper target
mrproper: mrproper_header do_mrproper
# Do Mr Proper target
do_mrproper: do_clean
@rm -f ../Lib/$(LIBRARY)
@rm -f *~ */*~
@rm -f compilation.log linking.log cleanup.log mrproper.log
thanks,
jonathan
Revision history for this message
|
#5 |
Hi Jonathan,
Could you please remove the line "LIBFLAGS = " from the Makefile and tell me whether this solves the problem?
Cheers,
Benjamin
On 19 Jan 2015, at 09:16 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
>
> Hi Benjamin,
>
> Yes this is the only information in the logfile. Here the Makefile as
> you requested:
>
> #######
> # MAKEFILE DEVOTED TO SAMPLEANALYZER COMMONS #
> #######
>
> # Compilers
> CXX = g++
>
> # C++ Compilation options
> CXXFLAGS = -Wall -O3 -fPIC -I$(MA5_
> CXXFLAGS += $(shell root-config --cflags)
> CXXFLAGS += -DROOT_USE
>
> # Linking options
> LIBFLAGS =
> LIBFLAGS += $(shell root-config --libs) -lEG
>
> # Files
> SRCS = $(wildcard */*.cpp)
> HDRS = $(wildcard */*.h)
> OBJS = $(SRCS:.cpp=.o)
>
> # Name of the library
> LIBRARY = libcommons_
>
> # Defining colours
> GREEN = "\\033[1;32m"
> RED = "\\033[1;31m"
> PINK = "\\033[1;35m"
> BLUE = "\\033[1;34m"
> YELLOW = "\\033[1;33m"
> CYAN = "\\033[1;36m"
> NORMAL = "\\033[0;39m"
>
> # All target
> all: header compile_header compile link_header link
>
> # Header target
> header:
> @echo -e $(YELLOW)
> @echo -e " Building SampleAnalyzer commons "
> @echo -e "------
>
> # Compile_header target
> compile_header:
> @echo -e $(YELLOW)
> @echo -e " Compilation "
> @echo -e "------
>
> # Link_header target
> link_header:
> @echo -e $(YELLOW)
> @echo -e " Linking "
> @echo -e "------
>
> # clean_header target
> clean_header:
> @echo -e $(YELLOW)
> @echo -e " Removing intermediate files from building "
> @echo -e "------
>
> # mrproper_header target
> mrproper_header:
> @echo -e $(YELLOW)
> @echo -e " Cleaning all the project "
> @echo -e "------
>
> # Precompile target
> precompile:
>
> # Compile target
> compile: precompile $(OBJS)
>
> # Compile each file
> %.o: %.cpp $(HDRS)
> $(CXX) $(CXXFLAGS) -o $@ -c $<
>
> # Link target
> link: $(OBJS)
> $(CXX) -shared -o ../Lib/$(LIBRARY) $(OBJS) $(LIBFLAGS)
>
> # Phony target
> .PHONY: do_clean header compile_header link_header
>
> # Clean target
> clean: clean_header do_clean
>
> # Do clean target
> do_clean:
> @rm -f $(OBJS)
>
> # Mr Proper target
> mrproper: mrproper_header do_mrproper
>
> # Do Mr Proper target
> do_mrproper: do_clean
> @rm -f ../Lib/$(LIBRARY)
> @rm -f *~ */*~
> @rm -f compilation.log linking.log cleanup.log mrproper.log
>
>
> thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#6 |
For some reason it worked for me if I told the program to use 2 cores for compiling but not when I told it to use 1 (?).
Revision history for this message
|
#7 |
Hi Benjamin,
The problem still persists after erasing the line that you mentioned, I get the same error.
Strangely after the run ./ma5 the line in makefile that I previously erased now reappears!
I forgot to mention that I'm working on a virtual machine that has debian wheezy 32 bits.
Which linux distribution you are working with?
As for Jory's comment, this may be the reason the max # of cores by default in my case is:
How many cores for the compiling? default = max = 1
I appreciate your help,
jonathan
Revision history for this message
|
#8 |
Hi Jonathan,
Sorry for not having clearer. Please change the Makefile, go into the tools/SampleAna
I will check further this max core stuff later this week or next week.
Cheers,
Benjamin
On 20 Jan 2015, at 16:46 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Hi Benjamin,
>
> The problem still persists after erasing the line that you mentioned, I get the same error.
> Strangely after the run ./ma5 the line in makefile that I previously erased now reappears!
>
> I forgot to mention that I'm working on a virtual machine that has debian wheezy 32 bits.
> Which linux distribution you are working with?
>
> As for Jory's comment, this may be the reason the max # of cores by
> default in my case is:
>
> How many cores for the compiling? default = max = 1
>
> I appreciate your help,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#9 |
Hi Benjamin,
I'm sorry but I followed your instructions and still encounter the same error when doing ./bin/ma5
Interestingly after typing make in the appropriate directory there is no error but when I try to run ma5 I get the same error as before.
After running ma5 it appears to change the makefile to the original one (the line LIBFLAGS reappears).
thank you for your cooperation,
jonathan
Revision history for this message
|
#10 |
Dear Jonathan, Jory,
The issue is now fixed. You can either update via bazaar your ma5 copy, or modify the file
madanalysis/
Go to line 317 and replace
commands = ['make'
by
if strcores=='':
commands = ['make'
else:
commands = ['make'
(be careful with the spacing).
Cheers,
Benjamin
On 21 Jan 2015, at 13:01 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Hi Benjamin,
>
> I'm sorry but I followed your instructions and still encounter the same
> error when doing ./bin/ma5
>
> Interestingly after typing make in the appropriate directory there is no
> error but when I try to run ma5 I get the same error as before.
>
> After running ma5 it appears to change the makefile to the original one
> (the line LIBFLAGS reappears).
>
> thank you for your cooperation,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#11 |
Hi Benjamin,
I downloaded the new version and it stopped in phase 5/8 with a fatal error.
In tools/SampleAna
CDFJetCluPlugin.hh and CDFMidPointPlug
g++ -Wall -O3 -fPIC -I/home/
fastjet/
compilation terminated.
make: *** [fastjet/
So a friend of mine downloaded the two libraries and also changed line 31 in both ClusterAlgoCDFJ
from
#include <fastjet/
to
#include "SampleAnalyzer
and the same for ClusterAlgoCDFM
After that step 5 worked OK but now I gota new error:
Component 6/8 - test program: interface to Fastjet
- Cleaning the project before building the test program ...
- Compiling the source files ...
- Linking the test program ...
** ERROR: impossible to link the project. For more details, see the log file:
** ERROR: /home/jonathan/
** ERROR: The test program building aborted.
Finally the new linking_fasjet.log states:
g++ Fastjet/Test.o -L/home/
/usr/lib/
/usr/lib/
/home/jonathan/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/home/jonathan/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
/usr/lib/
collect2: error: ld returned 1 exit status
make: *** [link] Error 1
Thanks,
jonathan
Revision history for this message
|
#12 |
Hi Jonathan,
Can you try without modifying the code, v1.1.12beta, and by swtiching off fastjet (modifying the file madanalysis/
Other question: do you have several copies of fastjet on your laptop? If yes, maybe impose madanalysis to use one of these by modifying te same file as above.
Cheers,
Benjamin
On 27 Jan 2015, at 11:56 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Hi Benjamin,
>
> I downloaded the new version and it stopped in phase 5/8 with a fatal error.
> In tools/SampleAna
> CDFJetCluPlugin.hh and CDFMidPointPlug
>
> g++ -Wall -O3 -fPIC -I/home/
> fastjet/
> compilation terminated.
> make: *** [fastjet/
>
> So a friend of mine downloaded the two libraries and also changed line
> 31 in both ClusterAlgoCDFJ
>
> from
> #include <fastjet/
> to
> #include "SampleAnalyzer
>
> and the same for ClusterAlgoCDFM
>
> After that step 5 worked OK but now I gota new error:
>
> Component 6/8 - test program: interface to Fastjet
> - Cleaning the project before building the test program ...
> - Compiling the source files ...
> - Linking the test program ...
> ** ERROR: impossible to link the project. For more details, see the log file:
> ** ERROR: /home/jonathan/
> ** ERROR: The test program building aborted.
>
>
> Finally the new linking_fasjet.log states:
>
>
> g++ Fastjet/Test.o -L/home/
> /usr/lib/
> /usr/lib/
> /home/jonathan/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /home/jonathan/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> /usr/lib/
> collect2: error: ld returned 1 exit status
> make: *** [link] Error 1
>
>
> Thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#13 |
Hi Benjamin,
Unfortunatley after switching off the fastjet the original compile (at step 5/8) error still persists.
when I run fastjet-config --version I get that I have the 3.0.2
I use Debian and I installed fastjet from the dependencies.
What system are you working on?
thanks,
jonathan
Revision history for this message
|
#14 |
Hi Jonathan
I am using both SL5, SL6 and MacOSX. Could you paste me the log file associated with the current crash? Thanks.
Cheers,
Benjamin
Revision history for this message
|
#15 |
Sorry about that, here is the logfile:
g++ -Wall -O3 -fPIC -I/home/
fastjet/
compilation terminated.
make: *** [fastjet/
thanks,
jonathan
Revision history for this message
|
#16 |
Hi Jonathan.
Could you please switch off totally fastjet in the configuration file and tell me whether it compiles?
Cheers,
Benjamin
On 03 Feb 2015, at 10:46 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Sorry about that, here is the logfile:
>
> g++ -Wall -O3 -fPIC -I/home/
> fastjet/
> compilation terminated.
> make: *** [fastjet/
>
> thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#17 |
Hi Benjamin,
As I said before, the compilation problem appears after I switched off fastjet by setting
fastjet_veto = 1 in installation_
so ma5 appears to be ignoring this veto.
Is that what you meant by "totally switching off" or is there some other way of doing it?
thanks,
jonathan
Revision history for this message
|
#18 |
This cannot be. Are you sure that you have removed the '#' in front of the line?
Cheers
Benjamin
On 03 Feb 2015, at 12:11 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Hi Benjamin,
>
> As I said before, the compilation problem appears after I switched off fastjet by setting
> fastjet_veto = 1 in installation_
> so ma5 appears to be ignoring this veto.
>
> Is that what you meant by "totally switching off" or is there some other
> way of doing it?
>
> thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#19 |
Oh sorry I forgot about that...
Now FastJet is successfully disabled and when I run ./bin/ma5 it passes all the tests and after that I get:
MadGraph 5 NOT found:
=> Particle labels from input/particles
=> 87 particles successfully exported.
=> Multiparticle labels from madanalysis/
=> Creation of the label 'invisible' (-> missing energy).
=> Creation of the label 'hadronic' (-> jet energy).
=> 8 multiparticles successfully exported.
ma5>
Is this normal?
So now I beleive we isolated the problem to the linking with fastjet...What's the next step then?
thanks,
jonathan
Revision history for this message
|
#20 |
Hi Jonathan,
Everything is fine so far. So, how many instances of fastjet do you have on your computer? Try to enable fastjet in the configuration file and properly provide the path where to find the librairies. If this does not work, maybe could you uninstall all instances of fastjet from your computer and reinstall it from the ma5 shell?
Cheers,
Benjamin
On 03 Feb 2015, at 13:01 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Oh sorry I forgot about that...
>
> Now FastJet is successfully disabled and when I run ./bin/ma5 it passes
> all the tests and after that I get:
>
> MadGraph 5 NOT found:
> => Particle labels from input/particles
> => 87 particles successfully exported.
> => Multiparticle labels from madanalysis/
> => Creation of the label 'invisible' (-> missing energy).
> => Creation of the label 'hadronic' (-> jet energy).
> => 8 multiparticles successfully exported.
> ma5>
>
> Is this normal?
> So now I beleive we isolated the problem to the linking with fastjet...What's the next step then?
>
> thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#21 |
Hi Benjamin,
I've checked the instances of fastjet on my computer and I wish to remove them and reinstall fastjet in the ma5 shell as you suggested, here are the files I get after searching fastjet on my system, some of which are linked to root/c++/fortran so could you please tell me which of these should I remove so that this procedure will run smoothly:
Fast C++ code for te kt and other jet algorythims
Developement files of fastjet
Fortran bindings of fastjet
Fastjet plugins
Tools for fastjet
Tools for fastjet - developement files
Fastjet macro for CERN ROOT
Thanks,
jonathan
Revision history for this message
|
#22 |
Hi Jonathan,
I am afraid I do not understand what you did on your system. This is probably more up to you to fix this.
What you can do, IMO, is to edit the installation options to force madanalysis5 to use a given fastjet instance available on your system (just correctly setup the paths in the configuration file).
Cheers,
Benjamin
On 16 Feb 2015, at 11:01 , jcohen <email address hidden> wrote:
> Question #260550 on MadAnalysis 5 changed:
> https:/
>
> Status: Answered => Open
>
> jcohen is still having a problem:
> Hi Benjamin,
>
> I've checked the instances of fastjet on my computer and I wish to
> remove them and reinstall fastjet in the ma5 shell as you suggested,
> here are the files I get after searching fastjet on my system, some of
> which are linked to root/c++/fortran so could you please tell me which
> of these should I remove so that this procedure will run smoothly:
>
> Fast C++ code for te kt and other jet algorythims
> Developement files of fastjet
> Fortran bindings of fastjet
> Fastjet plugins
> Tools for fastjet
> Tools for fastjet - developement files
> Fastjet macro for CERN ROOT
>
> Thanks,
> jonathan
>
> --
> You received this question notification because you are an answer
> contact for MadAnalysis 5.
Revision history for this message
|
#23 |
Hi Benjamin,
After removing ALL instances of fastjet I typed in the ma5 interface:
install fastjet
After the installation is complete I turn off the veto on fastjet (adding the appropriate # in installation_
The compilation of ma5 + fastjet runs without any error.
I thank you again for your patience in overcoming this ordeal.
jonathan