Re: [cmake-developers] New command 'file(LOCK_DIRECTORY ...)'
On 06-Oct-14 16:54, Brad King wrote: On 10/05/2014 04:59 PM, Ruslan Baratov via cmake-developers wrote: So it can't be resolved without some low-level management. I'm thinking about `file` command sub-option like LOCK_DIRECTORY: `file(LOCK_DIRECTORY /path/to/shared-dir)`. Does it sounds doable? I think it is a reasonable proposal. However, more design is needed here, including at least the following items: * There needs to be a way to define the scope of the lock other than until cmake exits. For example, until current function exits. * There needs to be a way to explicitly unlock early, or to explicitly release management of the lock (e.g. to pass to some other process). How it will looks like in terms of CMake code? * The implementation on Windows needs to be handled carefully. The safe way to delete a file there is to open it with a handle that is marked as delete-on-close, and then close the handle. That way the file is eventually deleted even if other processes currently have it open/locked. I don't know if there is an equivalent for a directory. Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Preparing for CMake 3.1.0-rc1
Hello dear developers! 03.10.2014, 16:28, Brad King brad.k...@kitware.com: Hi Folks, In preparation for the first 3.1 release candidate I will feature- freeze master as of 2014-10-09. As of now there are several open topics in 'next' that we should try to land in master by then, but please refrain from adding non-trivial changes in the meantime. Thanks, -Brad CMake is a large set of tools which touches upon many topics. I think everyone is doing their questions and know little about the problems and successes of others. I noticed that the version for developers (from cmake.org/file/dev does not contain the html documentation (I checked for Windows). Fresh build documentation from 7 October: Sphinx http://ifw.podsvirov.pro/cmake/doc/sphinx Doxygen http://ifw.podsvirov.pro/cmake/doc/doxygen In the first place when the distribution CMake version is created Sphinx and it's a good documentation. But I urge you not to forget about documenting code - it will help new developers to understand and to make a contribution. The version of Doxygen documentation is not in the best form, but I have plans to work over this. And advertising: CMake 3.1 will contain the new CPack generator called IFW, which will help to create a nice installer with a graphical user interface on the desktop. Fresh morning building (with html documentation from Sphinx) online installer with update support. Linux: http://ifw.podsvirov.pro/cmake/cmake-dev-linux-x86_64-online.run Windows: http://ifw.podsvirov.pro/cmake/cmake-dev-windows-x86-online.exe Set one is updated constantly :-) All for a good day! Regards, Konstantin Podsvirov -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
[cmake-developers] [CMake 0015198]: ps2pdf not detected when cross-compiling
The following issue has been SUBMITTED. == http://www.cmake.org/Bug/view.php?id=15198 == Reported By:Helfer Thomas Assigned To: == Project:CMake Issue ID: 15198 Category: CMake Reproducibility:always Severity: minor Priority: low Status: new == Date Submitted: 2014-10-07 08:09 EDT Last Modified: 2014-10-07 08:09 EDT == Summary:ps2pdf not detected when cross-compiling Description: Hi, ps2pdf not detected when cross-compiling (and only when cross-compiling), but latex is. The main difference seems that latex is an executable and ps2pdf14 a script (I am using the standard texlive distribution). Steps to Reproduce: 1) decompress the attached archive and launch cmake in the created directory: cmake . -DCMAKE_TOOLCHAIN_FILE=../ToolChain-i686-w64-mingw32.cmake 2) check the PS2PDF_CONVERTER value in CMakeCache.txt: //Path to a program. PS2PDF_CONVERTER:FILEPATH=PS2PDF_CONVERTER-NOTFOUND Additional Information: ps2pdf is detected when no toolchain file is declared == Issue History Date ModifiedUsername FieldChange == 2014-10-07 08:09 Helfer Thomas New Issue 2014-10-07 08:09 Helfer Thomas File Added: test-cmake.tar.bz2 == -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] New command 'file(LOCK_DIRECTORY ...)'
On 10/07/2014 02:49 AM, Ruslan Baratov wrote: How it will looks like in terms of CMake code? That's what I'm asking you to think through and propose ;) Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] fix-OSX-bundle-rpaths-and-Qt5 topic
On 10/07/2014 12:34 AM, clin...@elemtech.com wrote: I don't have a 10.5 machine Oops, I must have mixed up my browser tabs on that one. but I've put in a commit which I hope solves the problem. 36c509b9 OSX: Only enable @rpath support on OS X 10.6 or greater. Thanks! -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/07/2014 03:22 AM, Amine Khaldi wrote: Please note that from the http://www.cmake.org/Bug/view.php?id=14972 fixes on, we can no longer compile ReactOS. Thanks for trying the development version and reporting this regression before it was released. We have many DEPENDS on files that exist in the source (they are simply input files, they're not generated or so) like spec files, inf files, idl files...etc and they end up in the Unknown Build Time Dependencies which should not happen, and did not happen with CMake 2.8.x Would you please construct a CMake code example demonstrating this in a small test case? We'd rather not have to try building all of ReactOS to reproduce this. Also, what build-time failure do you actually get? Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/7/2014 8:47 AM, Brad King wrote: Please note that from thehttp://www.cmake.org/Bug/view.php?id=14972 fixes on, we can no longer compile ReactOS. Also, if you really care about ReactOS, you could run a nightly cmake dashboard. Or it is almost certain that it will get broken again. If we are not testing it, it is broken... http://www.cmake.org/testing/ Thanks. -Bill -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/07/2014 04:09 PM, Bill Hoffman wrote: On 10/7/2014 8:47 AM, Brad King wrote: Please note that from thehttp://www.cmake.org/Bug/view.php?id=14972 fixes on, we can no longer compile ReactOS. Also, if you really care about ReactOS, you could run a nightly cmake dashboard. Or it is almost certain that it will get broken again. If we are not testing it, it is broken... http://www.cmake.org/testing/ To clarify as far as I understand it this particular issue is with building ReactOS on regular Windows rather than using CMake under ReactOS. Which of course doesn't mean that there shouldn't be a dashboard for it nonetheless :) Nils -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
[cmake-developers] FindThreads_overhaul topic
Eike, This fails on a few dashboard machines: http://open.cdash.org/testDetails.php?test=286158250build=3519076 CMake Error at /.../Modules/FindThreads.cmake:207 (add_library): add_library cannot create imported target CMake::Threads because another target with the same name already exists. Instead of using a variable like HAVE_CMAKE_THREADS_TARGET, just test for the target. Look at FindOpenGL for an example: if(OPENGL_FOUND) if(NOT TARGET OpenGL::GL) add_library(OpenGL::GL UNKNOWN IMPORTED) + add_library(CMake::Threads INTERFACE IMPORTED GLOBAL) The imported target should not be GLOBAL. Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
To clarify as far as I understand it this particular issue is with building ReactOS on regular Windows rather than using CMake under ReactOS. Which of course doesn't mean that there shouldn't be a dashboard for it nonetheless :) Yup, this is what I understand. It isn't a problem with compiling CMake on ReactOS, but problem of compiling ReactOS with latest CMake master. I'll follow the discussion with Amine and try to sort it out. --Adam -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Preparing for CMake 3.1.0-rc1
On 10/03/2014 08:27 AM, Brad King wrote: In preparation for the first 3.1 release candidate I will feature- freeze master as of 2014-10-09. As of now there are several open topics in 'next' that we should try to land in master by then, but please refrain from adding non-trivial changes in the meantime. In order to get the dashboard cleaned up for final merges to 'master' before the release, please refrain from merging any changes other than dashboard fixes and documentation updates to 'next'. Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
As a result of this, it's no longer possible to use the new CMake to compile ReactOS, which makes us see this as a severe regression, considering that CMake is the build system we use to compile the project. Adam, Brad and co, can you please help ? If I understand correctly the change introduced by a33cf6d08853ea4c79324bdd36c04f311a23f20a (Ninja: Consider only custom commands deps as side-effects (#14972)) caused that regression? Can you please point to me any test case I can run in order to see it failing on CMake 3.x and running fine on CMake 2.8. If this is a bug I am willing to fix it ASAP. However please note that if these files are produced by some custom commands that not specify them as an output and then these are inputs (dependencies) to some regular build command, then it is expected behavior. Indeed it working was working previously in 2.8 just by coincidence, because CMake was generating suboptimal build graph simply adding all custom commands regardless of anything as implicit dependencies, even it was not necessary or expressed anywhere, causing severe clutter in Ninja build graph. So, (1) Are these .spec, .inf or .idl generated by some custom commands? (2) If yes, are these files specified as an output of these subcommand commands? If no, why? --Adam -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/07/2014 04:56 PM, Adam Strzelecki wrote: (1) Are these .spec, .inf or .idl generated by some custom commands? (2) If yes, are these files specified as an output of these subcommand commands? If no, why? From what I remember from the IRC discussion ... They are regular (not generated) source files that are listed as dependencies (but not outputs) of custom commands. Nils -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/7/2014 10:42 AM, Adam Strzelecki wrote: Which of course doesn't mean that there shouldn't be a dashboard for it nonetheless:) Yup, this is what I understand. It isn't a problem with compiling CMake on ReactOS, but problem of compiling ReactOS with latest CMake master. I'll follow the discussion with Amine and try to sort it out. I see. In that case a contract test might be a good thing... https://github.com/Kitware/CMake/tree/master/Tests/Contracts The idea is to add a build of a project as a test to CMake. See CMAKE_CONTRACT_PROJECTS in Tests/CMakeLists.txt for an idea of how this works. -Bill -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] FindThreads_overhaul topic
Am Dienstag, 7. Oktober 2014, 10:35:24 schrieb Brad King: Eike, This fails on a few dashboard machines: http://open.cdash.org/testDetails.php?test=286158250build=3519076 CMake Error at /.../Modules/FindThreads.cmake:207 (add_library): add_library cannot create imported target CMake::Threads because another target with the same name already exists. Instead of using a variable like HAVE_CMAKE_THREADS_TARGET, just test for the target. Look at FindOpenGL for an example: if(OPENGL_FOUND) if(NOT TARGET OpenGL::GL) add_library(OpenGL::GL UNKNOWN IMPORTED) + add_library(CMake::Threads INTERFACE IMPORTED GLOBAL) The imported target should not be GLOBAL. Both issues should be fixed now. Eike signature.asc Description: This is a digitally signed message part. -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Improving the version selection behavior of EXACT
On 10/06/2014 04:04 PM, Rolf Eike Beer wrote: Topic FPHSA_exact_version pushed to next. Thanks. The impl looks okay to me. Please look at extending the FindPackageTest with a case to cover this. It already has at least one case for FPHSA. Alternatively you could add cases to RunCMake.find_package if that proves easier. Thanks, -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
From what I remember from the IRC discussion ... They are regular (not generated) source files that are listed as dependencies (but not outputs) of custom commands. Okay now I get it. So actually 7243c951 (Ninja: Don't limit custom cmd side-effects to build folder (#14972)) is to blame not a33cf6d0 (Ninja: Consider only custom commands deps as side-effects (#14972)). Previously in 2.8 only files in build folder might have been be side-effects, but as we have discussed with Brad this is pretty inaccurate assumption. So we have decided that ANY file that is dependency of custom command may be a side effect, therefore ReactOS files landing in Unknown Build Time Dependencies. that serve only one purpose, to not bail out Ninja on very beginning if file does not exist as file may appear later on during build. But it does NOT stop Ninja to properly relaunch custom command if the file gets modified. So I guess we may point to the wrong direction. Simple example: -- test.txt 1 2 3 test -- CMakeLists.txt cmake_minimum_required(VERSION 2.6) project(DependTest C) add_custom_command( OUTPUT test-out.txt COMMAND cp test.txt test-out.txt DEPENDS test.txt) add_custom_target(depend-test ALL cat test-out.txt DEPENDS test-out.txt) (1) first run $ ninja -v [1/2] cd /tmp/depend cp test.txt test-out.txt [2/2] cd /tmp/depend cat test-out.txt 1 2 3 test (2) 2nd run $ ninja -v [1/1] cd /tmp/depend cat test-out.txt 1 2 3 test (3) touching test.txt $ touch test.txt $ ninja -v [1/2] cd /tmp/depend cp test.txt test-out.txt [2/2] cd /tmp/depend cat test-out.txt 1 2 3 test So with proper test case and detailed reports, something more than vague latest CMake makes our project not to compile anymore I am unable to help, but I am willing to help. So please either provide test case, or some info how to run this problematic build (I have access to OSX, Linux or Windows system), or find me on #cmake IRC channel under OnO. --Adam -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
On 10/07/2014 02:21 PM, Adam Strzelecki wrote: So please either provide test case, or some info [snip] On 10/07/2014 03:20 PM, Adam Strzelecki wrote: I had a talk with Amine, and the problem was that due 7243c951 their build.ninja was getting numerous extra phony entries that were leading Ninja to crash! So yes we could point our fingers on Ninja and close the case, but I propose we do something better: For reference, that commit was: Ninja: Don't limit custom cmd side-effects to build folder (#14972) http://cmake.org/gitweb?p=cmake.git;a=commitdiff;h=7243c951 So now a large number of phony rules for files in the source tree appear that did not before. The case in question uses custom command dependencies so is not helped by the parent change: Ninja: Consider only custom commands deps as side-effects (#14972) http://www.cmake.org/gitweb?p=cmake.git;a=commitdiff;h=a33cf6d0 I kindly ask to make possible side-effects phony rules as deprecated by some new policy. Basically long story short all these side-effects tricks were to make some old projects don't explicitly specifying custom command outputs work with Ninja that stat files once at start, so if there is a side-effect file that isn't explicitly specified as command output it will be NOT restat, that will lead to some other command having that file as dependency to fail. That situation is discussed thoroughly here: Add explicit specification of custom command side effect outputs http://www.cmake.org/Bug/view.php?id=14963 and here: https://github.com/martine/ninja/issues/760#issuecomment-46540858 Ideally both CMake and Ninja will fix their pieces of the problem. Fixing the Ninja side as I propose in the above-linked comment would solve the problem outright with no further changes. The fix on the CMake side will also require projects to be updated to use the new feature, but will help generate more efficient Ninja build rules. Since Make does restat dependencies on each rule this won't happen with Make. So we put side-effects in phony ensuring Ninja to pickup the rules even the file didn't exist when Ninja was started. No. For build systems besides Ninja the side effects cannot be listed as outputs of custom commands. Their timestamps are not always updated when the custom command runs, so the rules may be left in an always-rerun state because the side-effect outputs will never be newer than inputs. Since CMake was designed for such build systems before Ninja existed, there is no interface to specify the side-effects explicitly. My original intention was actually to limit these phony rules and it was done by: (1) a33cf6d0 (Ninja: Consider only custom commands deps as side-effects (#14972)) However somewhere during discussion with Brad I raised question why such side-effect are to be in build folder in fact they can be anywhere and it we want to ensure Ninja build won't fail when Make build goes well we need introduced: (2) 7243c951 (Ninja: Don't limit custom cmd side-effects to build folder (#14972)) This unfortunately lead Ninja to crash with many extra side-effect phony rules in ReactOS case. So again I propose we make side-effect as policy enabled only for projects 3.1. That is not possible because CMake currently does not have enough information to produce valid Ninja files without the phony rules. The proper fix to address issue 14963 and provide projects with an interface to specify custom command side-effects explicitly. Then dependencies can be hooked up correctly for Ninja. I invite you to propose an interface to achieve this. This will not be done before the freeze for 3.1 on Thursday. Reverting 7243c951 will resolve the problem for ReactOS in out-of-source builds. So, we either revert that or hope Ninja can be fixed to deal with the large dependency lists w/out crashing. -Brad -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Severe regression caused by #14972 fixes
This will not be done before the freeze for 3.1 on Thursday. Reverting 7243c951 will resolve the problem for ReactOS in out-of-source builds. So, we either revert that or hope Ninja can be fixed to deal with the large dependency lists w/out crashing. In meantime I've pushed stage/cmp0055-disable-ninja-side-effects that make CMake warn about this compatibility layer when such phony rules are about to be emitted. Setting policy to NEW or making all deps to be explicit outputs disables warning, setting to OLD removed warning keeping phony rules. IHMO this is win-win for everybody, since it will emit warnings once someone installs 3.1 and possibly relies on side-effect. One thing I am not sure about is if the CMP0055.rst provided is clear enough for the reader. --Adam -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
[cmake-developers] Contract testing of CMake
Hi Bill: In a recent thread on list you brought up the topic of contract testing https://github.com/Kitware/CMake/tree/master/Tests/Contracts where apparently the idea is CMake is tested by building some git version of CMake than building some fixed version of another project against that version to make sure no regression in CMake behaviour has crept in. I would like to help CMake (and protect PLplot against potential future CMake regressions) by doing informal contract PLplot build tests by hand for some fixed version of PLplot to to start with. Which CMake branch would you recommend for this general purpose? Should I be following the tip of maint, master, next, or release with such tests? Of course, once I got such a contract test to work by hand, I would probably want to move to a formal automated procedure so your advice on that would be appreciated as well. Alan __ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __ Linux-powered Science __ -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
[cmake-developers] Any ideas for accessing the Dart source code?
I thought it would be interesting to install my own local dart server to learn how to use CTest as a dart client. However, the dart server software development effort seems to have a largely broken web presence. For example, http://www.cmake.org/dart/HTML/Install.shtml points to a Dart.pdf documentation file that is a broken link. The Dart Wiki at http://na-mic.org/Wiki/index.php/Dart2Summary which is recommended by the above cmake.org site is still up and running, but it also refers to that same broken link for Dart.pdf. From the Wiki the latest Dart release seems to be 1.0.9 (from 7 (!) years ago). I discovered that release does include the Dart.pdf file which is good, but only includes jar files and not source code which is bad. For example, I am concerned those 7-year old *.class files in the jar files might not run properly with modern java. Therefore, I would like to build those jar files from source. According to the above wiki you can get access to the source code using svn co http://svn.na-mic.org/svn/Dart/trunk but that command immediately returns svn: OPTIONS of 'http://svn.na-mic.org/svn/Dart/trunk': 200 OK In contrast, Dart.pdf states that source code is available using svn checkout http://svn.na-mic.org:8000/svn/Dart However, that command times out with cannot connect to server. Does anyone here know how to access the Dart source or is that gone forever? Alan __ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __ Linux-powered Science __ -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Any ideas for accessing the Dart source code?
Consider using CDash instead. It was Dart's drop-in replacement... and is actively maintained today. www.cdash.org On Tue, Oct 7, 2014 at 6:46 PM, Alan W. Irwin ir...@beluga.phys.uvic.ca wrote: I thought it would be interesting to install my own local dart server to learn how to use CTest as a dart client. However, the dart server software development effort seems to have a largely broken web presence. For example, http://www.cmake.org/dart/HTML/Install.shtml points to a Dart.pdf documentation file that is a broken link. The Dart Wiki at http://na-mic.org/Wiki/index.php/Dart2Summary which is recommended by the above cmake.org site is still up and running, but it also refers to that same broken link for Dart.pdf. From the Wiki the latest Dart release seems to be 1.0.9 (from 7 (!) years ago). I discovered that release does include the Dart.pdf file which is good, but only includes jar files and not source code which is bad. For example, I am concerned those 7-year old *.class files in the jar files might not run properly with modern java. Therefore, I would like to build those jar files from source. According to the above wiki you can get access to the source code using svn co http://svn.na-mic.org/svn/Dart/trunk but that command immediately returns svn: OPTIONS of 'http://svn.na-mic.org/svn/Dart/trunk': 200 OK In contrast, Dart.pdf states that source code is available using svn checkout http://svn.na-mic.org:8000/svn/Dart However, that command times out with cannot connect to server. Does anyone here know how to access the Dart source or is that gone forever? Alan __ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __ Linux-powered Science __ -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Any ideas for accessing the Dart source code?
On 2014-10-07 19:06-0400 David Cole wrote: Consider using CDash instead. It was Dart's drop-in replacement... and is actively maintained today. www.cdash.org Thanks, Dave, for that information. In reviewing why I wasted my time looking at the moribund dart rather than cdash, dart clients are mentioned several times in the ctest-2.8.12.2 documentation (and I missed the one reference to cdash that also occurs there) so I did a google search, found http://www.cmake.org/dart/HTML/Install.shtml and went galloping off in the wrong direction. :-( I have now looked at the documentation for ctest-3.0.2, and cdash is mentioned a lot more but then so is dart. I think that documentation should be updated to point users exclusively at cdash (possibly with a reference to the dart protocol or dart standard but definitely not the dart software which really is moribund or dart servers unless you are referring to the protocol and not the software). Furthermore, to prevent others being misled like I was by http://www.cmake.org/dart/HTML/Install.shtml, could someone with write access to that web server either drop that page or modify that page so that it at least mentions that dart is no longer maintained and cdash is the suggested alternative? Currently http://www.cdash.org/overview/ does not mention dart at all which is probably the best approach and what the ctest documentation should do as well. Alan __ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __ Linux-powered Science __ -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Any ideas for accessing the Dart source code?
On Tue, Oct 7, 2014 at 9:23 PM, Alan W. Irwin ir...@beluga.phys.uvic.ca wrote: On 2014-10-07 19:06-0400 David Cole wrote: Consider using CDash instead. It was Dart's drop-in replacement... and is actively maintained today. www.cdash.org Thanks, Dave, for that information. In reviewing why I wasted my time looking at the moribund dart rather than cdash, dart clients are mentioned several times in the ctest-2.8.12.2 documentation (and I missed the one reference to cdash that also occurs there) so I did a google search, found http://www.cmake.org/dart/HTML/Install.shtml and went galloping off in the wrong direction. :-( I have now looked at the documentation for ctest-3.0.2, and cdash is mentioned a lot more but then so is dart. I think that documentation should be updated to point users exclusively at cdash (possibly with a reference to the dart protocol or dart standard but definitely not the dart software which really is moribund or dart servers unless you are referring to the protocol and not the software). Furthermore, to prevent others being misled like I was by http://www.cmake.org/dart/HTML/Install.shtml, could someone with write access to that web server either drop that page or modify that page so that it at least mentions that dart is no longer maintained and cdash is the suggested alternative? Currently http://www.cdash.org/overview/ does not mention dart at all which is probably the best approach and what the ctest documentation should do as well. Alan __ Alan W. Irwin Astronomical research affiliation with Department of Physics and Astronomy, University of Victoria (astrowww.phys.uvic.ca). Programming affiliations with the FreeEOS equation-of-state implementation for stellar interiors (freeeos.sf.net); the Time Ephemerides project (timeephem.sf.net); PLplot scientific plotting software package (plplot.sf.net); the libLASi project (unifont.org/lasi); the Loads of Linux Links project (loll.sf.net); and the Linux Brochure Project (lbproject.sf.net). __ Linux-powered Science __ -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers We should probably wipe those references or just update that for 3.0 -- Jacky Alciné - http://jalcine.me -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Contract testing of CMake
On 10/7/2014 4:59 PM, Alan W. Irwin wrote: Of course, once I got such a contract test to work by hand, I would probably want to move to a formal automated procedure so your advice on that would be appreciated as well. The idea is to test next CMake against a version of the contract project that is known to work with the current release of CMake. We don't want development of the contract project to cause the test to fail. So, use a fixed version of PLplot and build it with next CMake. By using next CMake bugs will be noticed as soon as they are introduced. -Bill -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
Re: [cmake-developers] Any ideas for accessing the Dart source code?
On 10/08/2014 12:46 AM, Alan W. Irwin wrote: I thought it would be interesting to install my own local dart server to learn how to use CTest as a dart client. You probably want the newer Kitware maintained replacement CDash: http://www.cdash.org/ Additional instructions for a local installation can be found here: http://public.kitware.com/Wiki/CDash:Installation CMake's own CDash Dashboard can be seen in action here: http://open.cdash.org/index.php?project=CMake Nils -- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake-developers
