remove piraha cache files in make clean

Comments (18)

1. 

   Steven R. Brandt

   • changed status to open
   • assigned issue to
   • removed comment

   • 2017-06-05T08:19:07+00:00
2. 

   Erik Schnetter

   • changed title to remove piraha cache files in make clean
   • removed comment

   • 2017-06-13T09:56:35+00:00
3. 

   Roland Haas reporter

   • changed milestone to ET_2017_05
   • removed comment

   • 2017-06-28T07:46:21+00:00
4. 

   Roland Haas reporter

   • removed comment

   Any progress on this? I think this really should be in before the release since otherwise there is no way of getting rid of the cache files other than rm -rf configs/sim.

   • 2017-06-28T07:47:10+00:00
5. 

   Steven R. Brandt

   • removed comment

   This week is really busy for me.

   • 2017-06-28T08:32:50+00:00
6. 

   Roland Haas reporter

   • removed comment

   Would you like to re-assign this then? Currently the release plan https://docs.einsteintoolkit.org/et-docs/Release_Details calls for a release by 2017-06-29 (which is impossible, given that it is tomorrow).

   Frank: given that those dates are not possible, is this just a typo and was supposed to read July in stead of June?

   • 2017-06-28T08:44:06+00:00
7. 

   Steven R. Brandt

   • removed comment

   Actually, I think this suffices:

   diff --git a/Makefile b/Makefile
   index b4ed2e9..31ac107 100644
   --- a/Makefile
   +++ b/Makefile
   @@ -505,6 +505,7 @@ ifneq ($strip($(CONFIGURATIONS)),)
    $(addsuffix -clean,$(CONFIGURATIONS)):
           $(NOTIFY_DIVIDER)
           @echo Cleaning configuration $(@:%-clean=%)
   +       @rm -f $(CONFIGS_DIR)/$(@:%-clean=%)/piraha
           cd $(CONFIGS_DIR)/$(@:%-clean=%)
           $(MAKE) -f $(CCTK_HOME)/lib/make/make.configuration TOP=$(CONFIGS_DIR)/$(@:%-clean=%) CCTK_HOME=$(CCTK_HOME) clean
           $(NOTIFY_DIVIDER)
   

   • 2017-06-28T08:56:24+00:00
8. 

   Erik Schnetter

   • removed comment

   Doesn't this require an -r flag as well?

   • 2017-06-28T09:07:33+00:00
9. 

   Roland Haas reporter

   • removed comment

   Shouldn't it be part of -realclean as well? Possible also cleandeps and cleanobjs though those are less clear to me.

   • 2017-06-28T10:38:23+00:00
10. 

    Steven R. Brandt

    • removed comment

    OK, so updating...

    diff --git a/Makefile b/Makefile
    index b4ed2e9..d0e4539 100644
    --- a/Makefile
    +++ b/Makefile
    @@ -505,6 +505,7 @@ ifneq ($strip($(CONFIGURATIONS)),)
     $(addsuffix -clean,$(CONFIGURATIONS)):
            $(NOTIFY_DIVIDER)
            @echo Cleaning configuration $(@:%-clean=%)
    +       @rm -fr $(CONFIGS_DIR)/$(@:%-clean=%)/piraha
            cd $(CONFIGS_DIR)/$(@:%-clean=%)
            $(MAKE) -f $(CCTK_HOME)/lib/make/make.configuration TOP=$(CONFIGS_DIR)/$(@:%-clean=%) CCTK_HOME=$(CCTK_HOME) clean
            $(NOTIFY_DIVIDER)
    @@ -591,6 +592,7 @@ ifneq ($strip($(CONFIGURATIONS)),)
     $(addsuffix -realclean,$(CONFIGURATIONS)):
            $(NOTIFY_DIVIDER)
            @echo Cleaning configuration $(@:%-realclean=%)
    +       @rm -fr $(CONFIGS_DIR)/$(@:%-clean=%)/piraha
            cd $(CONFIGS_DIR)/$(@:%-realclean=%)
            $(MAKE) -f $(CCTK_HOME)/lib/make/make.configuration TOP=$(CONFIGS_DIR)/$(@:%-realclean=%) CCTK_HOME=$(CCTK_HOME) realclean
            $(NOTIFY_DIVIDER)
    

    • 2017-06-28T12:12:13+00:00
11. 

    anonymous

    • removed comment

    The -realclean target should use @rm -fr $(CONFIGS_DIR)/$(@:%-realclean=%)/piraha I think.

    • 2017-06-28T13:25:47+00:00
12. 

    Roland Haas reporter

    • removed comment

    I would also suggest to add this to make.configuration whichsi where the actual work happens, not the wrapper Makefile in the Cactus root. Finally it may be better to make piraha a subdir of build (or scratch) which would automatically clean it in realclean (and clean).

    • 2017-07-03T09:26:03+00:00
13. 

    Roland Haas reporter

    • removed comment

    One more (slightly unexpected, but actually perfectly correct) observed behaviour: I also found that once I remove the piraha directory and call make sim again, it does not trigger CST ie the piraha cache is not re-generated which may be ok (since the C code generated from the ccl files did not change). I guess that one would need a rebuild to force CST to re-run (after all the piraha files are just caches so need only be re-genreated when the ccl files change).

    • 2017-07-03T09:29:37+00:00
14. 

    Roland Haas reporter

    • changed status to open
    • removed comment

    I have a fix for this (and an issue on BW) here:

    https://bitbucket.org/cactuscode/cactus/pull-requests/40/rhaas-piraha-fixes/diff

    • 2017-07-03T09:33:09+00:00
15. 

    Roland Haas reporter

    • removed comment

    Just to be clear, since I am proposing the fix, Steve, would you mind to review them, please?

    • 2017-07-03T10:26:13+00:00
16. 

    Steven R. Brandt

    • changed status to open
    • removed comment

    • 2017-07-03T10:42:01+00:00
17. 

    Roland Haas reporter

    • changed status to resolved
    • removed comment

    Thank you.

    Applied as git hash 8fb34581 of the flesh.

    • 2017-07-03T10:57:52+00:00
18. 

    Roland Haas reporter

    • edited description
    • changed status to closed

    • 2019-02-21T19:49:33+00:00
19. Log in to comment