From frank.loeffler at aei.mpg.de Tue Aug 8 04:28:16 2006 From: frank.loeffler at aei.mpg.de (Frank Loeffler) Date: Tue, 08 Aug 2006 11:28:16 +0200 Subject: [Patches] extend ICN to (swapped/unswapped) theta-ICN, default does not change Message-ID: <44D85930.3090504@aei.mpg.de> Hi, I still have here this patch, which extends ICN by request to theta-ICN, which can also be swapped, as published recently. The default does not change and the code is simple. This is Cactus Problem Report: 2043 thanks, Frank -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: patch_ICN_MoL Url: http://www.cactuscode.org/pipermail/patches/attachments/20060808/78daa13f/attachment.pl From frank.loeffler at aei.mpg.de Tue Aug 8 05:01:19 2006 From: frank.loeffler at aei.mpg.de (Frank Loeffler) Date: Tue, 08 Aug 2006 12:01:19 +0200 Subject: [Patches] CactusBase/MoL: typo in interface.ccl which prevents compiling Message-ID: <44D860EF.6080907@aei.mpg.de> Hi, There is a binary character instead of an 'r' inside interface.ccl. Strangly I could not track that down to a recent patch. The attached patch should remove this and make it compilable again. While updating MoL before, I got some 'checksum' errors after applying the patches, so it got the whole files. I got this independently on two machines. Is there something going wrong on the repository side? This is Cactus Problem Report: 2044 Frank From frank.loeffler at aei.mpg.de Tue Aug 8 05:02:10 2006 From: frank.loeffler at aei.mpg.de (Frank Loeffler) Date: Tue, 08 Aug 2006 12:02:10 +0200 Subject: [Patches] CactusBase/MoL: typo in interface.ccl which prevents compiling Message-ID: <44D86122.9040904@aei.mpg.de> Hi, There is a binary character instead of an 'r' inside interface.ccl. Strangly I could not track that down to a recent patch. The attached patch should remove this and make it compilable again. While updating MoL before, I got some 'checksum' errors after applying the patches, so it got the whole files. I got this independently on two machines. Is there something going wrong on the repository side? This is Cactus Problem Report: 2044 This email actually contains the patch... Frank -------------- next part -------------- A non-text attachment was scrubbed... Name: patch_MoL_bug Type: application/octet-stream Size: 777 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20060808/65faf80e/attachment.obj From frank.loeffler at aei.mpg.de Tue Aug 8 05:26:53 2006 From: frank.loeffler at aei.mpg.de (Frank Loeffler) Date: Tue, 08 Aug 2006 12:26:53 +0200 Subject: [Patches] CactusBase/MoL: typos(?) all over the files Message-ID: <44D866ED.7080007@aei.mpg.de> There are several typos(?) all over the place. I have the suspicion that these are not real typos, but some error in the cvs repository. The attached patch fixes them, but it might worth looking into the cause of them. This is Cactus Problem Report: 2045 and probably related to 2044 (my previous mail). Frank -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: patch_MoL_bug2 Url: http://www.cactuscode.org/pipermail/patches/attachments/20060808/53be8282/attachment.pl From baiotti at aei.mpg.de Mon Aug 7 05:37:14 2006 From: baiotti at aei.mpg.de (Luca Baiotti) Date: Mon, 07 Aug 2006 10:37:14 -0000 Subject: [Patches] [Developers] Add parameter to choose prolongation type for ADMBase In-Reply-To: References: Message-ID: <44D717D4.6020600@aei.mpg.de> Hi Erik, has this patch already been applied? It seems it hasn't. Is there a specific reason why not? Ciao, Luca. Erik Schnetter wrote: > Introduce a parameter ADMBase::prolongation type, which chooses how the > ADMBase variable are prolongated. This parameter is references in the > ADMBase group tag tables. The default is as is, namely "Lagrange". > Another useful value is "none", disabling prolongation of these > variables, which allows people to activate fewer time levels to save > memory. > > -erik From jthorn at aei.mpg.de Tue Aug 29 08:27:02 2006 From: jthorn at aei.mpg.de (Jonathan Thornburg) Date: Tue, 29 Aug 2006 15:27:02 +0200 (CEST) Subject: [Patches] [patch] make CCTK_VWarn() flush stdout/stderr before printing error msg Message-ID: The current-CVS version of CCTK_VWarn() does not flush stdout or stderr before printing the warning message. Therefore, if Cactus's stdout and stderr are being combined into a single output stream, as with bash% cactus_foo myrun.par 2>&1 | tee logfile csh% cactus_foo myrun.par |& tee logfile then CCTK_VWarn() messages usually appear out of sequence with respect to output from both CCTK_VInfo() and printf(). IMHO this is a serious problem: it means that it's hard to tell from looking at the combined stdout/stderr what the program has done before vs after a warning message was generated. I propose the enclosed patch, which fixes this by changing CCTK_VWarn() to flush stdout and stderr before printing its own warning message. (Note that CCTK_VWarn() may print either only to stderr, or to both stdout and stderr, depending on which processor this is, so to be safe we flush both streams.) ciao, -- -- Jonathan Thornburg Max-Planck-Institut fuer Gravitationsphysik (Albert-Einstein-Institut), Golm, Germany, "Old Europe" http://www.aei.mpg.de/~jthorn/home.html "Washing one's hands of the conflict between the powerful and the powerless means to side with the powerful, not to be neutral." -- quote by Freire / poster by Oxfam -------------- next part -------------- Index: WarnLevel.c =================================================================== RCS file: /cactusdevcvs/Cactus/src/main/WarnLevel.c,v retrieving revision 1.78 diff -u -r1.78 WarnLevel.c --- WarnLevel.c 6 Jul 2006 09:14:46 -0000 1.78 +++ WarnLevel.c 29 Aug 2006 13:24:54 -0000 @@ -518,6 +518,19 @@ char *message = NULL; char hostname[MAXNAMELEN+1]; + /* + * If stdout and stderr are merged into a single output stream, as with + * basn% cactus_foo myrun.par 2>&1 | tee logfile + * csh% cactus_foo myrun.par |& tee logfile + * then we need to flush buffers before switching from one to another, + * in order to ensure that stderr messages appear correctly sequenced + * with respect to stdout output. Note we need to flush *both* stdout + * and stderr since this function may print to just stderr, or to both + * of these, depending on which processor this is. + */ + fflush(stdout); + fflush(stderr); + /* Start generating message only if the warbcallback list is not NULL */ if(warncallbacks) { From steve.white at aei.mpg.de Thu Aug 31 07:26:58 2006 From: steve.white at aei.mpg.de (Steve White) Date: Thu, 31 Aug 2006 14:26:58 +0200 Subject: [Patches] Make ADMConstraints GF public In-Reply-To: <44C728C4.6070008@aei.mpg.de> References: <44C728C4.6070008@aei.mpg.de> Message-ID: <20060831122658.GD25551@xeon01.aei.mpg.de> Hi, I tried running the testsuites before and after applying the patch. I compiled in all the distro thorns I could (-= SDF, PETC, LAPACK etc) Find attached the results. With and without Franks patch 34 tests and 34 fail. Unfortunately the results are slightly different. On the other hand, there are just more error messages before the patch than after. This test seems to fail in somewhat different ways in the two cases test_ADM_4 I'm kind of in a rush, so it's hard for me to judge. What's the call? On 26.07.06, Frank Loeffler wrote: > Hi, > > this patch makes the GFs of CactusEinstein/ADMConstraints public. > > The reason is that this way other thorns can use the values already > computed here for their own purposes by inheritage. One example of such > a thorn is ADMDevelopment/MaskedADMConstraints, which requires this > patch to work. > Another possibility to access these GF would be to get their pointer by > variable name. However, I would consider inheritage cleaner. I also > cannot see a reason why these variables should stay private. > > Frank > Index: interface.ccl > =================================================================== > RCS file: /cactusdevcvs/CactusEinstein/ADMConstraints/interface.ccl,v > retrieving revision 1.18 > diff -u -r1.18 interface.ccl > --- interface.ccl 14 Feb 2005 09:50:26 -0000 1.18 > +++ interface.ccl 26 Jul 2006 08:27:08 -0000 > @@ -20,7 +20,7 @@ > USES FUNCTION Boundary_SelectVarForBC > USES FUNCTION Boundary_SelectGroupForBC > > -private: > +public: > > real hamiltonian type=GF tags='tensortypealias="Scalar"' > { > _______________________________________________ > Patches mailing list > Patches at cactuscode.org > http://www.cactuscode.org/mailman/listinfo/patches -- Steve White : Programmer Max-Planck-Institut f?r Gravitationsphysik Albert-Einstein-Institut Am M?hlenberg 1, D-14476 Golm, Germany +49-331-567-7625 -------------- next part -------------- A non-text attachment was scrubbed... Name: ts.gz Type: application/x-gzip Size: 9354 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20060831/e2536ca7/attachment.gz -------------- next part -------------- A non-text attachment was scrubbed... Name: ts-patched.gz Type: application/x-gzip Size: 8106 bytes Desc: not available Url : http://www.cactuscode.org/pipermail/patches/attachments/20060831/e2536ca7/attachment-0001.gz