On 27 Aug 2012, at 21:19, rhaas@tapir.caltech.edu wrote:
User: rhaas Date: 2012/08/27 02:19 PM
Modified: /trunk/src/ GRHydro_UpdateMaskM.F90
Log: GRHydro: clear atmosphere_mask_real in MHD AtmosphereReset
From: Roland Haas roland.haas@physics.gatech.edu
Hi Roland,
Looks like something went wrong (on one of the commits, not necessarily the above). The build and test system is giving a build failure on datura:
/home/ianhin/Cactus/EinsteinToolkitTests/arrangements/EinsteinInitialData/GRHydro_InitData/src/GRHydro_Bondi.c(811): error: "GRHydro_reflevel" has already been de clared in the current scope CCTK_INT GRHydro_reflevel=0; ^
compilation aborted for /home/ianhin/Cactus/EinsteinToolkitTests/configs/EinsteinToolkitTests_4dbbcf02976dfa1bc57dd21dc12b375f60e0db06/build/GRHydro_InitData/GRHy dro_Bondi.c (code 2)
This seems to be declared both in interface.ccl as well as in this (and another) source file.
Hello Ian, Erik, all,
This seems to be declared both in interface.ccl as well as in this (and another) source file.
Yup. I see it. Erik actually beat you to it, Ian. I do not know why this ever compiled. The line in question was last changed on August 8th. It declares (so Erik tells me, I am still compiling) a variable that is never used.
The simple fix is to remove the unused declaration.
I'll try and hunt down why this started to fail now.
Yours, Roland
Hello all,
The simple fix is to remove the unused declaration.
That was the correct thing to do and I have commited it. This change is actually part of the proposed changes for GRHydro_InitData (looks like they are from the same git commit but since I split them up by thorn for subversion, I missed it).
I'll try and hunt down why this started to fail now.
This started to fails since a GRHydro commit made GRHydro_reflevel public (used to be private before) which pushed it into GRHydro_InitData's context. Not sure why it failed even with this since apparently DECLARE_CCTK_PARAMETERS creates a new scope after it, so it should have been at best a shadow variable. Anyhow what happened was actually the "right thing" and I understand why it happened.
Thank you for being so quick in pointing out the error.
Yours, Roland
users@lists.einsteintoolkit.org
-
Ian Hinder -
Roland Haas