Integration crashes for e-p collisions

[sschmidt]

WHIZARDs integration crashes for hadron-electron and hadron-photon collisions:

---INTEGRATE---
| Initializating integration for process test1:
Runtime Error: strfun.f90, line 721: Subscript 1 of SFCHAIN%STRFUN (value 0) is out of range (1:1)
Program terminated by fatal error
strfun.f90, line 721: Error occurred in STRFUN:STRFUN_CHAIN_MAKE_EVALUATORS
processes.f90, line 1920: Called by PROCESSES:PROCESS_CONNECT_STRFUN
integrations.f90, line 277: Called by INTEGRATIONS:INTEGRATION_SETUP_BEAMS
integrations.f90, line 595: Called by INTEGRATIONS:INTEGRATION_INIT0
integrations.f90, line 729: Called by INTEGRATIONS:INTEGRATE_PROCESS0
integrations.f90, line 749: Called by INTEGRATIONS:INTEGRATE_PROCESS1
commands.f90, line 2880: Called by COMMANDS:CMD_INTEGRATE_EXECUTE
commands.f90, line 930: Called by COMMANDS:COMMAND_EXECUTE
commands.f90, line 5544: Called by COMMANDS:COMMAND_LIST_EXECUTE
whizard.f90, line 278: Called by WHIZARD:WHIZARD_PROCESS_STREAM
whizard.f90, line 256: Called by WHIZARD:WHIZARD_PROCESS_FILE
main.f90, line 301: Called by MAIN
Aborted

See attached file for an example input file.

[sschmidt]

example input file

[sschmidt]

Sorry, the failure for gamma-p was because of keeping the none statement instead of lhapdf for the photon.

[sschmidt]

corrected version for ep-collision

[dwiesler]

confirmed, gdb with gfortran 4.5 compiled version reports this:

---INTEGRATE---
| Initializating integration for process test2:
 -----EXITING INIT------------

Program received signal SIGSEGV, Segmentation fault.
__interactions_MOD_external_link_get_ptr () at interactions.f90:203
203     interactions.f90: No such file or directory.
        in interactions.f90
Current language:  auto; currently fortran
(gdb) bt full
#0  __interactions_MOD_external_link_get_ptr () at interactions.f90:203
No locals.
#1  0xb78a7d96 in __interactions_MOD_interaction_find_link () at interactions.f90:1091
No locals.
#2  0xb78f25e5 in __strfun_MOD_strfun_chain_make_evaluators () at strfun.f90:724
No locals.
#3  0xb796e3ba in __processes_MOD_process_connect_strfun () at processes.f90:1920
No locals.
#4  0x080ef5cc in ?? ()
No symbol table info available.
#5  0xbf80b72c in ?? ()
No symbol table info available.
#6  0x00000001 in ?? ()
No symbol table info available.
#7  0x00000000 in ?? ()
No symbol table info available.
(gdb)

[Juergen Reuter]

It seems that all in all a single structure function is not supported. It either triggers an error

 Initializating integration for process ilc_zh:
 -----EXITING INIT------------
******************************************************************************
******************************************************************************
*** WHIZARD BUG: Structure functions: colliding particles can't be determined
******************************************************************************
******************************************************************************
WHIZARD run aborted.

(which it should not!) or seems to raise the aforementioned seg fault. This is really bad, as WK promised to exactly solve this problem!!!!

[Juergen Reuter]

The guy who finally solves this ticket should add an example for e-p collisions and also tell Hans to include this in the HUDSON setup. Or -- we simply add an additional test for e-p collisions in the test suite.

[kilian]

Currently looking into this. For a proper e-p example, I'll make the option available to set one or both beam energies separately. (Currently, a c.m. momentum can be set, but this is not very convenient.) Implementing also the crossing angle.

[kilian]

r2779 eliminates the segfault and adds the missing beam options (asymmetric beams, crossing angle etc.).