#1955: Multipole: Add possibility to have out_dir distinct from regular IO:out_dir, but default to the old behavior ------------------------------------+--------------------------------------- Reporter: knarf | Owner: Type: enhancement | Status: reviewed_ok Priority: minor | Milestone: Component: EinsteinToolkit thorn | Version: development version Resolution: | Keywords: ------------------------------------+---------------------------------------
Comment (by rhaas):
Even when adding "Multipole:" the merge request, git log still is lacking the required information: {{{ $ git log commit 0ef3e71f999c3f7cf2f925989595b86bf7d53e20 Merge: cd658d7 c25f761 Author: Roland Haas rhaas@aei.mpg.de Date: Tue Aug 30 14:46:04 2016 -0500
Multipole: Merge branch own_io_dir
commit c25f7615292e7bacdb652001e1ad826ba5984277 Author: Frank Löffler knarf@cct.lsu.edu Date: Tue Aug 30 11:47:34 2016 -0500
Multipole: create out_dir if it does not exist
commit 953e5aac90e58f61d8d66c72fa727ea73410f088 Author: Frank Löffler knarf@cct.lsu.edu Date: Mon Aug 29 16:56:57 2016 -0500
initialize my_out_dir with correct value
commit d61c3ad38f49a74a3c3832b0417c0acb9619d9a0 Author: Frank Löffler knarf@cct.lsu.edu Date: Mon Aug 29 13:01:34 2016 -0500
Add possibility to have out_dir distinct from regular IO:out_dir, but default to the old behavior
commit cd658d7a1cfee286c611cb4e171e99aefbdb07ef Author: Erik Schnetter schnetter@gmail.com Date: Thu Jul 21 16:52:47 2016 -0400
EHFinder: Correct error in division
}}} is what it looks like (see https://bitbucket.org/rhaas80/einsteinanalysis).
There is {{{CCTK_VError}}} which is quite old (at least in the current released version) so should be safe to use (compared to {{{CCTK_VERROR}}}) and still lets the compiler know that the routine won't return.
The branch you created was a "private branch" I thought, not really intended for use by end-users. Is that correct? In that case you could as well (if eg you had not write permission to einsteinanalysis) have forked the repo (into one of your own), done the modification there and then created a pull request? So since the branch (seems to) have served no other purpose than to allow for the pull request, I would treat its history as something that does not need to be preserved (contrary to eg master's or the release branches' history which must never change).
Note: these are both optional (since they only affect developers), returning the error code to the user would be more important to me to tell the truth.