Actsfw 6 12 read sim digi adaptions

This is a couple of updates needed for the ML prototype, I also renamed associatedSurface() for parameters to referenceSurface() because that's what they actually are.

Looking for a clarification here: should the DetectorElementBase share ownership of its DigitizationModule with its clients here, which is the intent expressed by returning a shared_ptr? Or is it the owner of the DigitizationModule, and should it only return a raw pointer to it?

Many detector elements can share one DigitizationModule, if they have the same readout segmentation. A DigitizationModule can have quite a lot of surfaces, so it's potentially quite heavy to make a one-to-one relationship.

I will postpone this change when we have agreed on a strategy

Having a default implementation like this makes it easy to forget about adding the method when building a new DetectorElement class. Since we only have two descendants of DetectorElementBase at present, I would propose to remove the default implementation and explicitly add it to all DetectorElements.

Sounds good, let's change it to WIP.

Probably we should discuss the design here a bit - what I would like to have is that the DetectorElements can have a decoration of a DigitizationModule but don't have to do so. How would you advise that we do that, @hgraslan @cgumpert ?

See above.

Unfortunately my attempt to remove the GCC warnings and CLANG warnings has not worked out, it seems :

// external include(s) #pragma GCC diagnostic push #if defined(__USE_GNU) #pragma GCC diagnostic ignored "-Wmisleading-indentation" #endif #include <Eigen/Dense> #pragma GCC diagnostic pop

has now worked :-( , at least I see the warnings appearing in the GCC jenkins build.

@cgumpert or @hgraslan do you have a better idea ?

Found 2 files with missing MPLv2 license statement:
Tests/Digitization/PlanarModuleStepper_test.cpp
Tests/Digitization/CartesianSegmentation_test.cpp

This merge request does not close a JIRA ticket. Please consider creating a JIRA ticket to describe your changes. This is important for generating a complete ChangeLog for the next release.

Jenkins Build FAILURE

Results available at: Jenkins [ACTS-MERGE #53]

Added 1 commit:

• 95b2e5ac - clang-format: enforce code style [ci-skip]

Compare with previous version

Found 2 files with missing MPLv2 license statement:
Tests/Digitization/PlanarModuleStepper_test.cpp
Tests/Digitization/CartesianSegmentation_test.cpp

This merge request does not close a JIRA ticket. Please consider creating a JIRA ticket to describe your changes. This is important for generating a complete ChangeLog for the next release.

Jenkins Build FAILURE

Results available at: Jenkins [ACTS-MERGE #54]

I don't understand this code: why do you put one pragma inside of the #if block, but not the other two?

I mentioned that in the merge request, I am a bit struggling with that issue. @cgumpert introduced that black in order to avoid a long list of warnings coming from indentation when using gcc. This - however - gives warnings with clang which seems to not know anything about this specific -W option. Actually, this was just a try, but it didn't seem to work anyway (it works for clang to not have warnings, but apparently _USE_GNU is not automatically set ?

Compiler disambiguation can be difficult as clang tries very hard to be gcc. But try this.

#if defined(__GNUC__) && !defined(__clang__)
// GCC specific code goes here
#endif

Looking around further, it seems that icc also has this bad habit of impersonating gcc. Adding them to the blacklist :)

#if defined(__GNUC__) && !defined(__clang__) && !defined(__INTEL_COMPILER)
// GCC specific code goes here
#endif

For more information, try these resources:

• Predefined GCC macros: https://gcc.gnu.org/onlinedocs/cpp/Common-Predefined-Macros.html
• Other compiler identification macros: http://nadeausoftware.com/articles/2012/10/c_c_tip_how_detect_compiler_name_and_version_using_compiler_predefined_macros