You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Benjamin Hindman <be...@berkeley.edu> on 2011/12/15 02:01:48 UTC
Re: Review Request: Port mesos build system to use automake, libtool
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/2981/#review3611
-----------------------------------------------------------
Makefile.am
<https://reviews.apache.org/r/2981/#comment8049>
I know you didn't add these spaces, but feel free to remove them here and everywhere else in this patch.
Makefile.am
<https://reviews.apache.org/r/2981/#comment8048>
s/make they're/make sure they're
Makefile.am
<https://reviews.apache.org/r/2981/#comment8063>
Where does this variable come from?
configure.ac
<https://reviews.apache.org/r/2981/#comment8051>
Let's just kill these lines.
configure.ac
<https://reviews.apache.org/r/2981/#comment8052>
Kill.
configure.ac
<https://reviews.apache.org/r/2981/#comment8808>
Can we (do we) test for python before we just invoke it?
configure.ac
<https://reviews.apache.org/r/2981/#comment8809>
Regarding my comment, does autotools give us most of what we want here?
configure.ac
<https://reviews.apache.org/r/2981/#comment8810>
Kill any extra spaces, here and everywhere else.
mesos-build-env.sh.in
<https://reviews.apache.org/r/2981/#comment8811>
I'd like to revisit how bad it would be to just use relative paths and have the build procedure copy artifacts to their necessary spot so that they work given the relative directories after building. I'm concerned about the added mental overhead of getting people used to using these extra scripts as well as exporting environmental variables (can cause frustrating issues) as well as getting this all to work with a debugger. It would be great to enumerate the issues getting the existing mesos-master, mesos-slave, etc to run to determine if we can come up with any (clever) solutions.
mesos-slave-dev.in
<https://reviews.apache.org/r/2981/#comment8812>
s/master/slave
mesos-slave-dev.in
<https://reviews.apache.org/r/2981/#comment8813>
s/master/slave
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8814>
What is this for?
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8815>
Can we do this?
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8816>
Newline.
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8817>
The rest of the file indents.
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8818>
Why the extra variable? Why not just libmesos_la_LIBADD +=? Line length?
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8819>
Wow, I didn't know MKDIR_P existed!!! Or $(@D) ... what does that mean?
src/Makefile.am
<https://reviews.apache.org/r/2981/#comment8820>
Please wrap these lines for the 80 character freaks like me.
src/common/process_utils.hpp
<https://reviews.apache.org/r/2981/#comment8821>
s/hardness/harness
src/common/webui_utils.cpp
<https://reviews.apache.org/r/2981/#comment8822>
Brace on newline.
src/java/src/org/apache/mesos/MesosSchedulerDriver.java
<https://reviews.apache.org/r/2981/#comment8824>
Put throw on new line inside braces.
src/java/src/org/apache/mesos/MesosSchedulerDriver.java
<https://reviews.apache.org/r/2981/#comment8823>
?
src/master/webui.cpp
<https://reviews.apache.org/r/2981/#comment8825>
This comment still applies, even though you moved the code.
src/slave/webui.cpp
<https://reviews.apache.org/r/2981/#comment8826>
Let's call spawnWebui start as well, the namespace says it all: utils::webui::start(...
src/tests/base_zookeeper_test.cpp
<https://reviews.apache.org/r/2981/#comment8827>
Yeah, I like the idea of a source dir and a build dir ...
src/tests/external/LxcIsolation/HoldMoreMemThanRequested.sh
<https://reviews.apache.org/r/2981/#comment8829>
Cool, so are we no longer copying these in the Makefile.am?
src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh
<https://reviews.apache.org/r/2981/#comment8830>
I would have expected first running the mesos-build-env.sh script ... so it is not needed? Or is it expected that it will be run before this? If it's not needed, great!
src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh
<https://reviews.apache.org/r/2981/#comment8831>
Put the args on continued newlines.
src/tests/external/SampleFrameworks/CppFramework.sh
<https://reviews.apache.org/r/2981/#comment8832>
It might be nice to stick a test in each of these files that uses one of the environment variables that is expected and print if they don't exist ... and that these files are expected to get run from the the test binary which sets up all these envs.
src/tests/external/SampleFrameworks/JavaFramework.sh
<https://reviews.apache.org/r/2981/#comment8833>
Why is this one in BUILD, but some others are in SOURCE, like the next one is in SOURCE.
src/tests/main.cpp
<https://reviews.apache.org/r/2981/#comment8828>
s/ROOT_DIR/SOURCE_DIR ?? Pretty please?
- Benjamin
On 2011-11-30 19:15:42, Charles Reiss wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/2981/
> -----------------------------------------------------------
>
> (Updated 2011-11-30 19:15:42)
>
>
> Review request for mesos, Benjamin Hindman and Andy Konwinski.
>
>
> Summary
> -------
>
> This very large patch changes the build system to use automake, libtool. This patch attempts to be feature-complete with the existing build system (exception: Solaris support).
>
> Notable changes:
> - Builds from source control now require configure, automake, libtool to be installed; added ./bootstrap to create configure, Makefile.ins, etc. (autoreconf will _not_ work.)
> - Install location are no longer under $(prefix)/mesos; instead, we follow the configure-set directories mostly, so
> * mesos-master, mesos-slave go into $(sbindir) which defaults to $(prefix)/sbin
> * mesos-local, mesos-log, mesos-mesos (framework submission) go into $(bindir) which defaults to $(prefix)/bin
> * mesos-launcher (utility program for LXC isolation), killtree.sh (utility program) go into $(pkglibexecdir) which defaults to $(prefix)/libexec/mesos
> * mesos.jar goes into $(libdir)/java which defaults to $(prefix)/lib/java
> * libmesos.{la,so,dylib} goes into $(libdir) which defaults to $(prefix)/lib
> * webui python scripts go into $(pkgdatadir) which defaults to $(prefix)/share/mesos
> - New configuration options for finding install locations, with defaults compiled in based on configured install location:
> * webui_dir
> * launcher_dir
> - Created wrapper scripts mesos-master-dev and mesos-slave-dev that set configuration environment variables suitable for running Mesos out of build directory
> - Tests find location of external files using MESOS_BUILD_DIR and MESOS_SOURCE_DIR environment variables set by test harness
> - libmesos_exec.a, libmesos_sched.a have been eliminated
> - configure autodetects Java by default (it can still be disabled)
> - Removes support for building without zookeeper?
> - 'make test' changes names to 'make check'.
>
> Notable things not done:
> - 'make dist' is broken, primarily because of third_party.
> - We make third_party libraries part of libmesos.la; to do this, we link (PIC) .a's and non-convenience static .la's into libmesos.la. This prevents us from building a non-shared libmesos and may not be portable to some platforms. This is not a change from what the existing build system does.
> - the root directory and src have separate Makefile.am's; this means that dependencies in src/ won't cause config.status to regenerate files created a configure time from .in files.
>
>
> This addresses bug MESOS-58.
> https://issues.apache.org/jira/browse/MESOS-58
>
>
> Diffs
> -----
>
> Makefile.am PRE-CREATION
> Makefile.in e8dec7a
> bootstrap PRE-CREATION
> config.guess f32079a
> config.sub 6759825
> configure fc6bf00
> configure.ac f6b6a01
> include/mesos/mesos.hpp PRE-CREATION
> install-sh 4fbbae7
> m4/dummy_m4_contents PRE-CREATION
> mesos-build-env.sh.in PRE-CREATION
> mesos-master-dev.in PRE-CREATION
> mesos-slave-dev.in PRE-CREATION
> src/Makefile.am PRE-CREATION
> src/Makefile.in 516f128
> src/common/process_utils.hpp 8e55d76
> src/common/webui_utils.hpp PRE-CREATION
> src/common/webui_utils.cpp PRE-CREATION
> src/config/config.hpp.in 49f666a
> src/detector/detector.cpp ef4aff7
> src/examples/cpp_test_executor.cpp 9dd244e
> src/examples/cpp_test_framework.cpp ab02805
> src/examples/java/Makefile.in 6500ded
> src/examples/java/TestExceptionFramework.java ba39757
> src/examples/java/TestFramework.java e1b6032
> src/examples/java/TestMultipleExecutorsFramework.java a49ecbb
> src/examples/java/test-exception-framework PRE-CREATION
> src/examples/java/test-executor PRE-CREATION
> src/examples/java/test-framework PRE-CREATION
> src/examples/java/test-multiple-executors-framework PRE-CREATION
> src/examples/java/test_exception_framework be78062
> src/examples/java/test_executor d2fbb05
> src/examples/java/test_framework 2713b34
> src/examples/java/test_multiple_executors_framework 9fad319
> src/examples/python/test-executor PRE-CREATION
> src/examples/python/test-framework PRE-CREATION
> src/examples/python/test_executor 8d6cc64
> src/examples/python/test_framework 050ab90
> src/examples/python/test_framework.py e6893cf
> src/examples/test_executor.cpp PRE-CREATION
> src/examples/test_framework.cpp PRE-CREATION
> src/java/mesos.pom e5b674d
> src/java/mesos.pom.in PRE-CREATION
> src/java/src/org/apache/mesos/MesosSchedulerDriver.java 04809a6
> src/jvm/jvm.hpp PRE-CREATION
> src/jvm/jvm.cpp PRE-CREATION
> src/log/log.hpp e50969b
> src/log/network.hpp c88822c
> src/master/slaves_manager.cpp 3f67aa1
> src/master/webui.cpp c5a20d1
> src/python/setup.py.in 7c1f6e5
> src/slave/lxc_isolation_module.cpp ab0843a
> src/slave/webui.cpp bb1c780
> src/tests/base_zookeeper_test.hpp e7bb7ed
> src/tests/base_zookeeper_test.cpp 202b66b
> src/tests/external/LxcIsolation/HoldMoreMemThanRequested.sh 7b87a7b
> src/tests/external/LxcIsolation/ScaleUpAndDown.sh 09333e2
> src/tests/external/LxcIsolation/TwoSeparateTasks.sh a2e2883
> src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh 9ac7cd5
> src/tests/external/SampleFrameworks/CFrameworkCmdlineParsing.sh 791cbf1
> src/tests/external/SampleFrameworks/CFrameworkInvalidCmdline.sh 3197c81
> src/tests/external/SampleFrameworks/CFrameworkInvalidEnv.sh 427b33b
> src/tests/external/SampleFrameworks/CppFramework.sh 968d951
> src/tests/external/SampleFrameworks/JavaExceptionFramework.sh bc4a4e0
> src/tests/external/SampleFrameworks/JavaFramework.sh db88d0b
> src/tests/external/SampleFrameworks/PythonFramework.sh 80a7e99
> src/tests/external_tests.cpp 4b08f86
> src/tests/main.cpp ada489d
> src/tests/sample_frameworks_tests.cpp 877fdf1
> src/tests/utils.hpp 65ce4df
> src/tests/utils.cpp 553ab2d
> src/tests/zookeeper_server.hpp cac9f0d
> src/tests/zookeeper_server.cpp 3927785
> third_party/libprocess/include/process/protobuf.hpp 7636822
>
> Diff: https://reviews.apache.org/r/2981/diff
>
>
> Testing
> -------
>
>
> Thanks,
>
> Charles
>
>
Re: Review Request: Port mesos build system to use automake, libtool
Posted by Charles Reiss <wo...@gmail.com>.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > Makefile.am, line 8
> > <https://reviews.apache.org/r/2981/diff/2/?file=61156#file61156line8>
> >
> > I know you didn't add these spaces, but feel free to remove them here and everywhere else in this patch.
Done (in this file).
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > Makefile.am, line 51
> > <https://reviews.apache.org/r/2981/diff/2/?file=61156#file61156line51>
> >
> > s/make they're/make sure they're
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > configure.ac, lines 44-59
> > <https://reviews.apache.org/r/2981/diff/2/?file=61162#file61162line44>
> >
> > Let's just kill these lines.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > configure.ac, lines 82-85
> > <https://reviews.apache.org/r/2981/diff/2/?file=61162#file61162line82>
> >
> > Kill.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > configure.ac, line 101
> > <https://reviews.apache.org/r/2981/diff/2/?file=61162#file61162line101>
> >
> > Regarding my comment, does autotools give us most of what we want here?
$target_os already comes from AC_CANONICAL_TARGET.
configure won't set _XOPEN_SOURCE or the funky Solaris flags (which are clearly wrong on Sparc anyways), and I don't understand why we're not always using -D_XOPEN_SOURCE=700.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > configure.ac, line 157
> > <https://reviews.apache.org/r/2981/diff/2/?file=61162#file61162line157>
> >
> > Kill any extra spaces, here and everywhere else.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > mesos-slave-dev.in, line 19
> > <https://reviews.apache.org/r/2981/diff/2/?file=61168#file61168line19>
> >
> > s/master/slave
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > mesos-slave-dev.in, line 21
> > <https://reviews.apache.org/r/2981/diff/2/?file=61168#file61168line21>
> >
> > s/master/slave
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 22
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line22>
> >
> > What is this for?
I originally tried to use suffix rules for protoc, for which this was necessary. But I switched to pattern rules (which probably means no BSD make, not that we care), because suffixes rules couldn't represent the two output files.
Removed.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 65
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line65>
> >
> > Can we do this?
No idea.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 78
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line78>
> >
> > Newline.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 107
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line107>
> >
> > The rest of the file indents.
Indented.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 175
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line175>
> >
> > Why the extra variable? Why not just libmesos_la_LIBADD +=? Line length?
This is an artifact from when I was considering building the (almost) self-contained python egg in this Makefile.am.
Replaced with libmesos_la_LIBADD
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 248
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line248>
> >
> > Wow, I didn't know MKDIR_P existed!!! Or $(@D) ... what does that mean?
>From the GNU make manual. "The directory part of the file name of the target, with the trailing slash removed". This won't work on non-GNU make, if you care about that.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/common/process_utils.hpp, line 27
> > <https://reviews.apache.org/r/2981/diff/2/?file=61171#file61171line27>
> >
> > s/hardness/harness
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/Makefile.am, line 451
> > <https://reviews.apache.org/r/2981/diff/2/?file=61169#file61169line451>
> >
> > Please wrap these lines for the 80 character freaks like me.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/common/webui_utils.cpp, line 18
> > <https://reviews.apache.org/r/2981/diff/2/?file=61173#file61173line18>
> >
> > Brace on newline.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/java/src/org/apache/mesos/MesosSchedulerDriver.java, line 46
> > <https://reviews.apache.org/r/2981/diff/2/?file=61199#file61199line46>
> >
> > Put throw on new line inside braces.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/java/src/org/apache/mesos/MesosSchedulerDriver.java, lines 47-49
> > <https://reviews.apache.org/r/2981/diff/2/?file=61199#file61199line47>
> >
> > ?
Removed.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/master/webui.cpp, line 71
> > <https://reviews.apache.org/r/2981/diff/2/?file=61205#file61205line71>
> >
> > This comment still applies, even though you moved the code.
Added to src/common/webui_utils.cpp
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/slave/webui.cpp, line 73
> > <https://reviews.apache.org/r/2981/diff/2/?file=61208#file61208line73>
> >
> > Let's call spawnWebui start as well, the namespace says it all: utils::webui::start(...
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh, line 16
> > <https://reviews.apache.org/r/2981/diff/2/?file=61214#file61214line16>
> >
> > I would have expected first running the mesos-build-env.sh script ... so it is not needed? Or is it expected that it will be run before this? If it's not needed, great!
external_tests.cpp should set the appropriate environment variables. It was missing MESOS_WEBUI_DIR and MESOS_LAUNCHER_DIR, which I added. I don't think we have anything which runs the LXC tests frequently.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh, line 50
> > <https://reviews.apache.org/r/2981/diff/2/?file=61214#file61214line50>
> >
> > Put the args on continued newlines.
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/tests/external/LxcIsolation/HoldMoreMemThanRequested.sh, line 18
> > <https://reviews.apache.org/r/2981/diff/2/?file=61211#file61211line18>
> >
> > Cool, so are we no longer copying these in the Makefile.am?
Except for the Java framework related scripts.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/tests/external/SampleFrameworks/JavaFramework.sh, line 7
> > <https://reviews.apache.org/r/2981/diff/2/?file=61220#file61220line7>
> >
> > Why is this one in BUILD, but some others are in SOURCE, like the next one is in SOURCE.
The Python framework doesn't need any compilation, so it's run out of the source directory. (If this generates .pyc files or similar, we should reconsider this.)
I didn't go to the effort of modifying the Java framework scripts to handle keeping the scripts in the source directory but finding the Java compiled code in the build directory.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > src/tests/main.cpp, line 57
> > <https://reviews.apache.org/r/2981/diff/2/?file=61223#file61223line57>
> >
> > s/ROOT_DIR/SOURCE_DIR ?? Pretty please?
Done.
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > Makefile.am, line 53
> > <https://reviews.apache.org/r/2981/diff/2/?file=61156#file61156line53>
> >
> > Where does this variable come from?
It's in the generated Makefile; thought it was a documented automake variable, but apparently not. Replaced with a listing of generated non-Makefile files in src/
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > configure.ac, line 97
> > <https://reviews.apache.org/r/2981/diff/2/?file=61162#file61162line97>
> >
> > Can we (do we) test for python before we just invoke it?
Except for --with-webui, we do. (See AM_PATH_PYTHON below.)
> On 2011-12-15 01:01:48, Benjamin Hindman wrote:
> > mesos-build-env.sh.in, line 1
> > <https://reviews.apache.org/r/2981/diff/2/?file=61166#file61166line1>
> >
> > I'd like to revisit how bad it would be to just use relative paths and have the build procedure copy artifacts to their necessary spot so that they work given the relative directories after building. I'm concerned about the added mental overhead of getting people used to using these extra scripts as well as exporting environmental variables (can cause frustrating issues) as well as getting this all to work with a debugger. It would be great to enumerate the issues getting the existing mesos-master, mesos-slave, etc to run to determine if we can come up with any (clever) solutions.
Using relative paths only works if you know the current working directory when the artifacts are run, which you won't get for free.
We could use the executable path to do detect installed versus uninstalled binaries, but this would require different code for essentially every platform (unless we want to trust argv[0] + PATH resolution). http://stackoverflow.com/questions/1023306/finding-current-executables-path-without-proc-self-exe gives ways of an executable path. We'd need to look at carefully about what we want to do when the path involves symbolic links or when the executable is in neither the build directory nor the install location.
Ideally, we could find an elegant way of hooking libtool's wrapper scripts (on platforms where they exist) to set the relevant environment variables.
- Charles
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/2981/#review3611
-----------------------------------------------------------
On 2011-12-15 19:28:39, Charles Reiss wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/2981/
> -----------------------------------------------------------
>
> (Updated 2011-12-15 19:28:39)
>
>
> Review request for mesos, Benjamin Hindman and Andy Konwinski.
>
>
> Summary
> -------
>
> This very large patch changes the build system to use automake, libtool. This patch attempts to be feature-complete with the existing build system (exception: Solaris support).
>
> Notable changes:
> - Builds from source control now require configure, automake, libtool to be installed; added ./bootstrap to create configure, Makefile.ins, etc. (autoreconf will _not_ work.)
> - Install location are no longer under $(prefix)/mesos; instead, we follow the configure-set directories mostly, so
> * mesos-master, mesos-slave go into $(sbindir) which defaults to $(prefix)/sbin
> * mesos-local, mesos-log, mesos-mesos (framework submission) go into $(bindir) which defaults to $(prefix)/bin
> * mesos-launcher (utility program for LXC isolation), killtree.sh (utility program) go into $(pkglibexecdir) which defaults to $(prefix)/libexec/mesos
> * mesos.jar goes into $(libdir)/java which defaults to $(prefix)/lib/java
> * libmesos.{la,so,dylib} goes into $(libdir) which defaults to $(prefix)/lib
> * webui python scripts go into $(pkgdatadir) which defaults to $(prefix)/share/mesos
> - New configuration options for finding install locations, with defaults compiled in based on configured install location:
> * webui_dir
> * launcher_dir
> - Created wrapper scripts mesos-master-dev and mesos-slave-dev that set configuration environment variables suitable for running Mesos out of build directory
> - Tests find location of external files using MESOS_BUILD_DIR and MESOS_SOURCE_DIR environment variables set by test harness
> - libmesos_exec.a, libmesos_sched.a have been eliminated
> - configure autodetects Java by default (it can still be disabled)
> - Removes support for building without zookeeper?
> - 'make test' changes names to 'make check'.
>
> Notable things not done:
> - 'make dist' is broken, primarily because of third_party.
> - We make third_party libraries part of libmesos.la; to do this, we link (PIC) .a's and non-convenience static .la's into libmesos.la. This prevents us from building a non-shared libmesos and may not be portable to some platforms. This is not a change from what the existing build system does.
> - the root directory and src have separate Makefile.am's; this means that dependencies in src/ won't cause config.status to regenerate files created a configure time from .in files.
>
>
> This addresses bug MESOS-58.
> https://issues.apache.org/jira/browse/MESOS-58
>
>
> Diffs
> -----
>
> Makefile.am PRE-CREATION
> Makefile.in e8dec7a
> bootstrap PRE-CREATION
> config.guess f32079a
> config.sub 6759825
> configure fc6bf00
> configure.ac f6b6a01
> include/mesos/mesos.hpp PRE-CREATION
> install-sh 4fbbae7
> m4/dummy_m4_contents PRE-CREATION
> mesos-build-env.sh.in PRE-CREATION
> mesos-master-dev.in PRE-CREATION
> mesos-slave-dev.in PRE-CREATION
> src/Makefile.am PRE-CREATION
> src/Makefile.in 516f128
> src/common/process_utils.hpp 8e55d76
> src/common/webui_utils.hpp PRE-CREATION
> src/common/webui_utils.cpp PRE-CREATION
> src/config/config.hpp.in 49f666a
> src/detector/detector.cpp ef4aff7
> src/examples/cpp_test_executor.cpp 9dd244e
> src/examples/cpp_test_framework.cpp ab02805
> src/examples/java/Makefile.in 6500ded
> src/examples/java/TestExceptionFramework.java ba39757
> src/examples/java/TestFramework.java e1b6032
> src/examples/java/TestMultipleExecutorsFramework.java a49ecbb
> src/examples/java/test-exception-framework PRE-CREATION
> src/examples/java/test-executor PRE-CREATION
> src/examples/java/test-framework PRE-CREATION
> src/examples/java/test-multiple-executors-framework PRE-CREATION
> src/examples/java/test_exception_framework be78062
> src/examples/java/test_executor d2fbb05
> src/examples/java/test_framework 2713b34
> src/examples/java/test_multiple_executors_framework 9fad319
> src/examples/python/test-executor PRE-CREATION
> src/examples/python/test-framework PRE-CREATION
> src/examples/python/test_executor 8d6cc64
> src/examples/python/test_framework 050ab90
> src/examples/python/test_framework.py e6893cf
> src/examples/test_executor.cpp PRE-CREATION
> src/examples/test_framework.cpp PRE-CREATION
> src/java/mesos.pom e5b674d
> src/java/mesos.pom.in PRE-CREATION
> src/java/src/org/apache/mesos/MesosSchedulerDriver.java 04809a6
> src/jvm/jvm.hpp PRE-CREATION
> src/jvm/jvm.cpp PRE-CREATION
> src/log/log.hpp e50969b
> src/log/network.hpp c88822c
> src/master/slaves_manager.cpp 3f67aa1
> src/master/webui.cpp c5a20d1
> src/python/setup.py.in 7c1f6e5
> src/slave/lxc_isolation_module.cpp ab0843a
> src/slave/webui.cpp bb1c780
> src/tests/base_zookeeper_test.hpp e7bb7ed
> src/tests/base_zookeeper_test.cpp 202b66b
> src/tests/external/LxcIsolation/HoldMoreMemThanRequested.sh 7b87a7b
> src/tests/external/LxcIsolation/ScaleUpAndDown.sh 09333e2
> src/tests/external/LxcIsolation/TwoSeparateTasks.sh a2e2883
> src/tests/external/LxcIsolation/run_scheduled_memhog_test.sh 9ac7cd5
> src/tests/external/SampleFrameworks/CFrameworkCmdlineParsing.sh 791cbf1
> src/tests/external/SampleFrameworks/CFrameworkInvalidCmdline.sh 3197c81
> src/tests/external/SampleFrameworks/CFrameworkInvalidEnv.sh 427b33b
> src/tests/external/SampleFrameworks/CppFramework.sh 968d951
> src/tests/external/SampleFrameworks/JavaExceptionFramework.sh bc4a4e0
> src/tests/external/SampleFrameworks/JavaFramework.sh db88d0b
> src/tests/external/SampleFrameworks/PythonFramework.sh 80a7e99
> src/tests/external_tests.cpp 4b08f86
> src/tests/main.cpp ada489d
> src/tests/sample_frameworks_tests.cpp 877fdf1
> src/tests/utils.hpp 65ce4df
> src/tests/utils.cpp 553ab2d
> src/tests/zookeeper_server.hpp cac9f0d
> src/tests/zookeeper_server.cpp 3927785
> third_party/libprocess/include/process/protobuf.hpp 7636822
>
> Diff: https://reviews.apache.org/r/2981/diff
>
>
> Testing
> -------
>
>
> Thanks,
>
> Charles
>
>