#2908: Cactus: Fixed arg list too long error.
Reporter: Max Morris
Status: submitted
Milestone:
Version:
Type: bug
Priority: major
Component:
[https://bitbucket.org/cactuscode/cactus/pull-requests/175](https://bitbucke…
There is a check in `lib/make/make.thornlib` to ensure that `sh` is not passed an argument string which is too long. However, this check is not robust enough because it counts words instead of characters/bytes and compares against arbitrary magic numbers rather than checking the max argument string length of the system. I am working on a thorn which fails to build as a direct result of the faulty check.
This PR replaces the existing check with a more robust one based on `$ getconf ARG_MAX`, with reasonable fallback behavior if the system shell somehow doesn’t support this command.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2908/cactus-fixed-arg-…
#2906: Cactus: New cGH field cctk_patch_is_cartesian
Reporter: Roland Haas
Status: new
Milestone:
Version:
Type: enhancement
Priority: major
Component: Cactus
Comment (by Roland Haas):
The change itself looks ok, though I’d keep the two empty lines I comment on the pull request.
This will require matching changes in Carpet and PUGH.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2906/cactus-new-cgh-fi…
#2898: CarpetX: Interpolator caches
Reporter: Lucas Timotheo Sanches
Status: new
Milestone:
Version:
Type: enhancement
Priority: major
Component:
Comment (by Roland Haas):
@{5f2f18c8e8c45600229698f5} udpated the pull request to split the functionality into a “setup” and an “interpolation” step.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2898/carpetx-interpola…
#2907: Arith uses very generic names for its (global) include files
Reporter: Roland Haas
Status: new
Milestone:
Version:
Type: bug
Priority: minor
Component: CarpetX
Thorn `Arith` in `CarpetX` provides a number of helper functions for arithmetic operations. However it uses a large number of include files in its `interface.ccl` `INCLUDES HEADER` section. Since these file names are globally visible \(not even restriction to other thorns `REQUIRE`ing `Arith` or having a `USES HEADER` in their `interface.ccl`, this pollutes \(a\) global name space \(namely the `configs/sim/bindings/include/` which _all_ thorns use to search for include files using a `-I` compiler option\). Right now the files are:
```
INCLUDES HEADER: arr.hxx IN arr.hxx
INCLUDES HEADER: cons.hxx IN cons.hxx
INCLUDES HEADER: cplx.hxx IN cplx.hxx
INCLUDES HEADER: defs.hxx IN defs.hxx
INCLUDES HEADER: div.hxx IN div.hxx
INCLUDES HEADER: dual.hxx IN dual.hxx
INCLUDES HEADER: mat.hxx IN mat.hxx
INCLUDES HEADER: rational.hxx IN rational.hxx
INCLUDES HEADER: rten.hxx IN rten.hxx
INCLUDES HEADER: simd.hxx IN simd.hxx
INCLUDES HEADER: smallvector.hxx IN smallvector.hxx
INCLUDES HEADER: spvect.hxx IN spvect.hxx
INCLUDES HEADER: sum.hxx IN sum.hxx
INCLUDES HEADER: ten3.hxx IN ten3.hxx
INCLUDES HEADER: tuple.hxx IN tuple.hxx
INCLUDES HEADER: vec.hxx IN vec.hxx
INCLUDES HEADER: vect.hxx IN vect.hxx
```
of which `vec.hxx` and `vect.hxx` seem to be the most problematic.
It may be a good idea to consolidate these into fewer files and / or prefixing with `arith_` or so to create a namespace.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2907/arith-uses-very-g…