2011-05-13T00:54:53 *** nbprashanth has quit IRC 2011-05-13T01:09:48 adityab: what version of cmake are you using? "cmake --help" 2011-05-13T01:10:27 sigblips: 2.8.3 2011-05-13T01:12:33 I'm using cmake 2.8.4 and I got two errors while doing a "cmake Algorithms/src" on Mac OS X. 2011-05-13T01:13:42 sigblips: In a directory outside the source tree, just do a cmake /path/to/top-level/directory 2011-05-13T01:14:39 I did that. Here is the error which is the same for line :7 and :8 2011-05-13T01:14:48 CMake Error at CMakeLists.txt:7 (set_target_properties): 2011-05-13T01:15:00 set_target_properties called with incorrect number of arguments. 2011-05-13T01:16:04 line 7 is set_target_properties(setikit PROPERTIES VERSION ${setikit_VERSION}) 2011-05-13T01:16:16 line 8 is set_target_properties(setikit PROPERTIES SOVERSION ${setikit_VERSION_MAJOR}) 2011-05-13T01:16:48 sigblips: https://github.com/setiQuest/Algorithms/blob/master/src/CMakeLists.txt 2011-05-13T01:16:49 Title: src/CMakeLists.txt at master from setiQuest/Algorithms - GitHub (at github.com) 2011-05-13T01:17:00 Lines 7 and 8 are different here 2011-05-13T01:17:33 ok 2011-05-13T01:17:43 sigblips: do a "cmake Algorithms" 2011-05-13T01:17:53 not Algortihms/src 2011-05-13T01:18:34 The top-level CMakeLists.txt file defines several variables which are then passed down to CMake files in subdirectories. 2011-05-13T01:18:36 Real programmers start at line zero!!! Now the question is where does cmake start? 2011-05-13T01:19:31 sigblips: I tried what you did, and cmake says lines 8 and 9 :) 2011-05-13T01:20:24 I get the same errors on line 7 and 8 when I do "cmake setiQuest-Algorithms-756e166" 2011-05-13T01:22:08 sigblips: Strange. Did you clear the previous cmake run's cache? 2011-05-13T01:22:44 I just built cmake 2.8.4 and this is the first time I've used it. 2011-05-13T01:24:27 ok. Make a fresh directory as a sister directory to the Algorithms/ directory. And then, in this new directory, run "cmake ../Algorithms" 2011-05-13T01:24:42 that should work. 2011-05-13T01:28:13 sigblips: Did that work? 2011-05-13T01:30:56 Nope, the cmake cache is smarter than that and now I'm getting a different error. 2011-05-13T01:32:18 what is it? 2011-05-13T01:36:55 OK, I had to clean up the cache directory mess cmake created and now it cmake'd without those errors. 2011-05-13T01:38:25 sigblips: I'm curious; did you do the very first cmake run from inside the Algorithms/ directory? 2011-05-13T01:39:20 Yes I did. I bet I'm not the first person to do that either. I was thinking ./configure 2011-05-13T01:41:13 *** saksham has quit IRC 2011-05-13T01:41:19 CMake is typically used for out-of-source builds. You make a separate build directory, outside the project dir, and then run cmake. All the makefiles and compiled stuff will be in the build dir and the source dir remains untouched. 2011-05-13T01:46:42 Interesting, I initially did the same as sigblips, though it worked for me 2011-05-13T01:50:41 The cmake directory structure is definitely very different that what I'm used to. 2011-05-13T01:55:52 After some mods because I don't have fftw it builds fine. Are you aware of the warnings on line 44 and 49: "sqsamples.c:44: warning: format ‘%lu’ expects type ‘long unsigned int’, but argument 5 has type ‘uint64_t’" 2011-05-13T02:02:42 /usr/includestdint.h:typedef unsigned long long uint64_t; 2011-05-13T02:03:02 I meant /usr/include/stdint.h:typedef unsigned long long uint64_t; 2011-05-13T02:09:36 adityab: the warning will go away if you change it to a "%llu" 2011-05-13T02:41:04 *** saksham has joined #setiquest 2011-05-13T02:49:58 Adiyab: Do you have a basic, power of 2, FFT that can provide backup in case FFTW is not available? Rob and I have typed in the routine from Numerical Recipes in C on occasion, but surely there is an OS version floating around? 2011-05-13T03:02:11 *** Gamegoofs2 has left #setiquest 2011-05-13T03:08:15 *** nave1990 has quit IRC 2011-05-13T03:09:23 FFT has to take complex-valued input 2011-05-13T03:13:29 *** gerryharp1 has quit IRC 2011-05-13T03:25:59 *** jrseti_ has quit IRC 2011-05-13T04:27:35 *** Adityab has quit IRC 2011-05-13T04:29:09 *** Adityab has joined #setiquest 2011-05-13T06:22:39 *** saksham has quit IRC 2011-05-13T06:57:54 *** michaelMo has joined #setiquest 2011-05-13T07:03:14 *** nbprashanth has joined #setiquest 2011-05-13T07:07:06 *** nbprashanth has quit IRC 2011-05-13T07:20:39 *** nbprashanth has joined #setiquest 2011-05-13T07:25:38 *** Pranav_rcmas has joined #setiquest 2011-05-13T07:31:59 *** Adityab has quit IRC 2011-05-13T07:50:18 *** sigblips has quit IRC 2011-05-13T08:08:05 *** nbprashanth has quit IRC 2011-05-13T08:24:29 *** michaelMo has quit IRC 2011-05-13T08:26:38 *** devon_hillard has quit IRC 2011-05-13T08:27:07 *** devon_hillard has joined #setiquest 2011-05-13T08:37:05 *** Pranav_rcmas has quit IRC 2011-05-13T09:35:47 *** Pranav_rcmas has joined #setiquest 2011-05-13T09:48:27 *** nbprashanth has joined #setiquest 2011-05-13T10:43:29 *** Adityab_ has joined #setiquest 2011-05-13T11:27:18 *** nbprashanth has quit IRC 2011-05-13T13:41:13 *** Pranav_rcmas has quit IRC 2011-05-13T14:47:04 *** Pranav_rcmas has joined #setiquest 2011-05-13T15:32:20 *** Libertiny has joined #setiquest 2011-05-13T15:37:21 *** devon_hillard has joined #setiquest 2011-05-13T15:37:21 *** devon_hillard has joined #setiquest 2011-05-13T15:51:24 *** nbprashanth has joined #setiquest 2011-05-13T17:23:09 *** nbprashanth has quit IRC 2011-05-13T17:28:59 *** Adityab_ has quit IRC 2011-05-13T18:25:15 *** sigblips has joined #setiquest 2011-05-13T18:38:02 *** Adityab_ has joined #setiquest 2011-05-13T18:46:01 *** nbprashanth has joined #setiquest 2011-05-13T19:09:30 *** nbprashanth has quit IRC 2011-05-13T19:11:11 *** nbprashanth has joined #setiquest 2011-05-13T19:17:51 Adityab_: did you see my printf "%llu" fixes the uint64_t warning comment? 2011-05-13T19:21:42 sigblips: Yes. I dozed off at the keyboard yesterday. Saw it this morning. 2011-05-13T19:22:18 Although I never got that warning with my compiler. 2011-05-13T19:27:37 My compiler is gcc 4.2.1 and it is building 64-bit executables. I wonder if the warning has to do with the compiler version, stdint.h system defines, or stdio.h (standard lib)? 2011-05-13T19:29:28 Mine is 4.5.2, 64-bit. 2011-05-13T19:36:05 I'm thinking the warning more likely is system or standard lib related. Mac OS X is BSD based and does some things slightly different than the GNU way. 2011-05-13T19:37:56 I just tested building sqsamples.c on my AWS 64-bit micro-instance which has gcc 4.1.2 and it did not give the warning. So this doesn't look like a gcc version issue. 2011-05-13T19:38:38 The AWS micro-instance is running 64-bit Linux. 2011-05-13T19:49:25 ok 2011-05-13T19:50:11 yes, probably some OS X/BSD issue. 2011-05-13T19:50:55 which reminds me, I should install a BSD in a VM and test once the code is big enough. 2011-05-13T19:54:40 That probably would be a good idea. I've found having baudline build on different OS platforms has really helped in finding and reducing bugs. Same thing holds for building with mixed 32/64-bit and big/little endian architectures. 2011-05-13T20:43:19 *** Pranav_rcmas has quit IRC 2011-05-13T21:39:37 *** jrseti_ has joined #setiquest 2011-05-13T23:20:57 *** nbprashanth has quit IRC 2011-05-13T23:31:52 Is anybody here running a 32-bit Linux kernel with >= 4 GB of RAM? 2011-05-13T23:32:37 *** devon_hillard has quit IRC 2011-05-13T23:37:21 sigblips: Is that normally possible? 2011-05-13T23:39:26 I think it is possible with a Linux extension. This probably goes for BSD too but my Mac is happy with 4 GB of RAM while running a 32-bit kernel.