danath / libmv Goto Github PK

What steps will reproduce the problem?
1. make (bin-dbg)

What is the expected output? What do you see instead?

-- Build files have been written to: /Users/joelkronander/Code/libmv/bin-dbg
Scanning dependencies of target libmv_test_main
[  1%] Building CXX object testing/CMakeFiles/libmv_test_main.dir/testing_main.o
/Users/joelkronander/Code/libmv/src/testing/testing_main.cc:27: internal 
compiler error: in reference_to_unused, at dwarf2out.c:10603
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
make[3]: *** [testing/CMakeFiles/libmv_test_main.dir/testing_main.o] Error 1
make[2]: *** [testing/CMakeFiles/libmv_test_main.dir/all] Error 2
make[1]: *** [all] Error 2
make: *** [debug] Error 2

What version of the product are you using? On what operating system?
SVN trunk revision 1176.
On my Mac Book Pro 2.4Ghz Intel Core i5
Mac OS X 10.6.6


Anyone had a similar issue?
Is this related to issue 15?

For release build it works just fine (after updating the glog and gflags libs 
as in svn branches/new_gflags_glog_dasy).

Great work guys! Keep it up.

Even though I'm not a lawyer it seems there might be a licensing conflict 
with SBA package. SBA is under GPL, your library is MIT. MIT is much more 
permissive than GPL, particularly it doesn't forbid distribution of binary 
code (i.e. compiled, close source and even for commercial reasons) and I 
applaud to you for choosing this license. However GPL strictly limits that 
you must be distributing your work as open source. (It might also restrict 
commercial usage, but I'm not sure about this particular problem).

One of the solutions you might look into is SSBA package from Christopher 
Zach: http://cs.unc.edu/~cmzach/opensource.html. It's LGPL, so you can use 
that in close source or commercial apps as long as the SSBA itself is 
compiled as .dll (or .so).

Hope this helps.

However seeing that you plan to write your own bundle adjustment procedure 
- that would be great. Keep it up!

Thank you @Keir for invitation to project. 
(http://groups.google.com/group/libmv-devel/t/1f3f9d00396f0fc0)

I'm making progress with it, but this is an error that I don't think I'll 
be able to understand or fix myself:

3>image_sequence_io.cc
2>L:\tmp\libmv-devel\libmv-
1\src\libmv/correspondence/bipartite_graph.h(260) : error C2679: binary '=' 
: no operator found which takes a right-hand operand of type 
'std::_Tree<_Traits>::const_iterator' (or there is no acceptable 
conversion)
2>        with
2>        [
2>            _Traits=std::_Tset_traits<std::pair<libmv::TrackID,const 
libmv::Feature *>,std::less<std::pair<libmv::TrackID,const libmv::Feature 
*>>,std::allocator<std::pair<libmv::TrackID,const libmv::Feature *>>,false>
2>        ]
2>        I:\Program Files\Microsoft Visual Studio 
9.0\VC\include\xtree(498): could be 'std::_Tree<_Traits>::iterator 
&std::_Tree<_Traits>::iterator::operator =(const 
std::_Tree<_Traits>::iterator &)'
2>        with
2>        [
2>            _Traits=std::_Tset_traits<std::pair<libmv::TrackID,const 
libmv::Feature *>,std::less<std::pair<libmv::TrackID,const libmv::Feature 
*>>,std::allocator<std::pair<libmv::TrackID,const libmv::Feature *>>,false>
2>        ]
2>        while trying to match the argument list 
'(std::_Tree<_Traits>::iterator, std::_Tree<_Traits>::const_iterator)'
2>        with
2>        [
2>            _Traits=std::_Tset_traits<std::pair<libmv::TrackID,const 
libmv::Feature *>,std::less<std::pair<libmv::TrackID,const libmv::Feature 
*>>,std::allocator<std::pair<libmv::TrackID,const libmv::Feature *>>,false>
2>        ]
2>        L:\tmp\libmv-devel\libmv-
1\src\libmv/correspondence/bipartite_graph.h(255) : while compiling class 
template member function 
'libmv::BipartiteGraph<LeftNode,Edge,RightNode>::Iterator 
libmv::BipartiteGraph<LeftNode,Edge,RightNode>::ScanEdgesForLeftNode(const 
LeftNode &)'
2>        with
2>        [
2>            LeftNode=libmv::ImageID,
2>            Edge=const libmv::Feature *,
2>            RightNode=libmv::TrackID
2>        ]
2>        L:\tmp\libmv-devel\libmv-
1\src\libmv/correspondence/correspondence.h(55) : see reference to class 
template instantiation 'libmv::BipartiteGraph<LeftNode,Edge,RightNode>' 
being compiled
2>        with
2>        [
2>            LeftNode=libmv::ImageID,
2>            Edge=const libmv::Feature *,
2>            RightNode=libmv::TrackID
2>        ]

Any ideas? 

(Don't close the ticket afterwards, I'll post all related stuff to Windows 
compilation here).

In "euclidean_resection.cc" line 636 (into "EuclideanResectionEPnP" function):
// Finally, with all three solutions, select the (R, t) with the best RMSE.
  VLOG(2) << "RMSE for solution 0: " << rmse(0);
  VLOG(2) << "RMSE for solution 1: " << rmse(0);
  VLOG(2) << "RMSE for solution 2: " << rmse(0);

As you see, log output is wrong for solution 1 and 2 (always print rmse(0) )

What steps will reproduce the problem?
1. Download the last version of libmv on the svn
2. Compile the code
3. Run test

What is the expected output? 
Test run fine...

What do you see instead?
image_io test failed :
From log files :
"
Wrong JPEG library version: library is 70, caller expects 62
E0629 11:00:03.951437 25703 image_io.cc:126] Error: Couldn't open 
hopefully_unexisting_file fopen returned 0
Wrong JPEG library version: library is 70, caller expects 62
E0629 11:00:03.951493 25703 image_io.cc:126] Error: Couldn't open 
hopefully_unexisting_file fopen returned 0
Wrong JPEG library version: library is 70, caller expects 62
"

Please use labels and text to provide additional information.

What steps will reproduce the problem?
1. make (main directory, method #1)
2. mkdir alternate_build (use cmake, method #2)

What is the expected output? What do you see instead?
Build is accomplished without a hitch ...

What version of the product are you using? On what operating system?
SVN revision #817. Ubuntu 9.10 64-bit

Please provide any additional information below.

The libmv build fails on Ubuntu. Here's the error I receive:
...
[  3%] Building CXX object 
third_party/glog/src/CMakeFiles/glog.dir/signalhandler.o
/media/dataDrive/workspace/repos/repos_libmv/src/third_party/glog/src/signa
lhandler.cc: In function ‘void* google::<unnamed>::GetPC(void*)’:
/media/dataDrive/workspace/repos/repos_libmv/src/third_party/glog/src/signa
lhandler.cc:72: error: ‘REG_EIP’ was not declared in this scope
...

I had a gflags error but fixed it by installing google gflags from their 
official svn. I am not exactly sure what is going wrong here. I have 
attached a log file. I get the same output for make as opposed to cmake.

Thanks,
bidwej

P.S. I really appreciate libmv. It's fantastic to see opensource work in 
this direction. I suspect I will have some code to contribute in the 
future. 

I fetch the latest libmv source from the 
https://github.com/Matthias-Fauconneau/libmv.git and build it on Fedora 14. It 
got below error info:
-------------------------------------------------------------------------------
../../lib/libcorrespondence_d.so: undefined reference to 
`libmv::Homography2DFromCorrespondences4PointRobust(Eigen::Matrix<double, -1, 
-1, 0, -1, -1> const&, Eigen::Matrix<double, -1, -1, 0, -1, -1> const&, double, 
Eigen::Matrix<double, 3, 3, 0, 3, 3>*, libmv::vector<int, 
Eigen::aligned_allocator<int> >*, double)'
../../lib/libcorrespondence_d.so: undefined reference to `libmv::ReadImage(char 
const*, libmv::Array3D<unsigned char>*)'
../../lib/libcorrespondence_d.so: undefined reference to 
`libmv::BoxFilter(libmv::Array3D<float> const&, int, libmv::Array3D<float>*)'
../../lib/libcorrespondence_d.so: undefined reference to 
`libmv::FundamentalFromCorrespondences7PointRobust(Eigen::Matrix<double, -1, 
-1, 0, -1, -1> const&, Eigen::Matrix<double, -1, -1, 0, -1, -1> const&, double, 
Eigen::Matrix<double, 3, 3, 0, 3, 3>*, libmv::vector<int, 
Eigen::aligned_allocator<int> >*, double)'
collect2: ld returned 1 exit status
make[3]: *** [bin/tests/Array_Matcher_test] Error 1

What steps will reproduce the problem?
1. make release
2.
3.

What is the expected output? What do you see instead?
Should compile, but there is an error "libmv/src/tools/tracker.cc:173:37: 
error: variable or field ‘DisplayMatches’ declared void" under archlinux 
system.

What version of the product are you using? On what operating system?
archlinux update to Jul 30 2010

Please provide any additional information below.
I tracked the error and change some codes. Then it compiles. Please check the 
attachment.

What steps will reproduce the problem?
1. Create a fundamental matrix corresponding to two cameras with crossing
viewing directions.
2. Call focal from fundamental.

An assert fails because it tries to compute the sqrt of a negative number.

The function should return some kind of error, so that the caller can do
something accordingly.

Add real building of pthreads, or remove the package after the dependency of 
gflags on pthtreads is gone.

What steps will reproduce the problem?
1. Check out latest version.
2. make
3. make release
4. cd bin-dbg
5. make test

What is the expected output? What do you see instead?

All test should pass. Several of the tests are failing:

The following tests FAILED:
     19 - image_converter_test (Failed)
     20 - image_transform_linear_test (OTHER_FAULT)
     46 - homography_test (Failed)
     50 - panography_test (Failed)
     75 - similarity_parameterization_test (Failed)
     79 - euclidean_parameterization_test (Failed)
Errors while running CTest
make: *** [test] Error 8

What version of the product are you using? On what operating system?

I downloaded the tarball libmv-libmv-7bef9ca.tar.gz and am using Ubuntu 10.10.

Please provide any additional information below.

I am keen to help resolve this issue but will need some guidance from the 
developers as to where to look for the problem, or what the problem may be..

What steps will reproduce the problem?
1. Use images from issue #1
2. Do all matching and calibration steps in order
3. try to switch to 3d view

What is the expected output? What do you see instead?

Almost always I get following assertion failure, a couple of times the 3d
view has actually worked using these exact same steps.

tvr:
/.../libmv/libmv-read-only/src/third_party/eigen/Eigen/src/Core/MatrixStorage.h:
44:
Eigen::ei_matrix_array<T, Size, MatrixOptions, Align>::ei_matrix_array()
[with T = double, int Size = 16, int MatrixOptions = 2, bool Align = true]:
Assert-makro "(reinterpret_cast<size_t>(array) & 0xf) == 0 && "this
assertion is explained here:
http://eigen.tuxfamily.org/dox/UnalignedArrayAssert.html  **** READ THIS
WEB PAGE !!! ****"" ei pidä paikkaansa. (==assertion failed)


What version of the product are you using? On what operating system?
svn rev577, compiled using gcc version 4.3.3 (Debian 4.3.3-13), debian sid

After Windows testing I would need to uncomment new HStack/VStack 
implementation.

What steps will reproduce the problem?
1. run cmake on the src/ folder
2. specify generator as Xcode project and use Default native Compilers
3. Configure eand Generate

What is the expected output? What do you see instead?
The Xcode project that it generates fails with this error.

"Project LIBMV.xcodeproj cannot be opened because the project file cannot
be parsed."

What version of the product are you using? On what operating system?
Code Revision: 847
Mac OS X 10.6.2 
i386 arch
CMake 2.8

I helps development to be able to run Xcode's debugger on the code. 
I hope someone can help me here.

What steps will reproduce the problem?
1. video match
2. metric bundle adjust
3. save blender .py

What is the expected output? What do you see instead?
file created... no file created

What version of the product are you using? On what operating system?
lastest svn Ubuntu 10.10

Please provide any additional information below.

was working before recent updates

What steps will reproduce the problem?
Reference: 
    Path: libmv\src\libmv\multiview\autocalibration.cc
    Lines: 105 to 109 (see below)

  // Eigen values should be possitive,
  Vec temp_values = eigen_solver.eigenvalues();
  if (temp_values.sum() < 0) {
    temp_values = -temp_values;
  }

After execution of these lines, I am still finding negative eigen values in 
temp_values.

What is the expected output? What do you see instead?
If I am not wrong, this should code should replaced with the code given below 
to ensure non-negative eigen values:

    for( int i = 0; i < 4; i++)
    {
        if(temp_values[i] < 0){
            temp_values[i] = -temp_values[i];
        }        
    }


What version of the product are you using? On what operating system?
Version: libmv-0.1-full_Win-x86-msvc2008
Operating System: Windows 7


Please provide any additional information below.

What steps will reproduce the problem?
1. Use the five points algorithm some hundreds of times

What is the expected output? What do you see instead?
The expected output is a matrix E (essential matrix).
Instead, the program is bloqued like in infinit loop, but doesn't crash nor 
freeze.

What version of the product are you using? On what operating system?
Ubuntu 11.04, libmv 0.1

Please provide any additional information below.

What steps will reproduce the problem?
1. Try to compile the code under windows (Visual C++ Express 2008)
2. It do not compile.
3.1 -> The glog issue under windows is the biggest problem.
3.2 -> Some minor issue are interpreted as error for visual Compiler.

What is the expected output? What do you see instead?
A binary revision of libmv (lib, test executable)

What version of the product are you using? On what operating system?
(Windows XP, Visual C++ Express 2008).


Please provide any additional information below.

For the minor issue I give the solution below :
imageio.cc
=> Line 34 ->   char *p = strrchr (c, '.');
Must be  : const char *p = strrchr (c, '.');

homography_test.cc
=> Line 35 and 71 ->   int num_h = 3;
Must be : const int num_h = 3;
Because it is used to perform a static allocation of MAT3 array. (It's 
strange that on your linux compiler you don't have an error or a warning).

For the glog windows issue. I can take a look, but do you have any advice 
to give me. First the _WIN32 preprocessor must be added to the glog 
windows project. I thinks this can be done in the CMakeLists.h.

Also I have a problem with :
Error   126 error C2718: 'libmv::Vec2': actual parameter with 
__declspec(align('16')) won't be aligned    e:\2009Projects\libmv\src
\libmv\base\vector_test.cc

Did you have any fix for this ...

Talk to you ;)

When I want to read a RGB image and write it in another file, the output image 
isn't the same.

The code that doesn't work on my ubuntu 10.4 :
Array3Du imageArrayBytes1;
ReadImage ("in.png", &imageArrayBytes1);
WriteImage(imageArrayBytes1,"out.png");

I expected to see the same image, but it's not.
I think there is a bug in the offset function ?
since even a :
      for (size_t y=0; y<imageArrayBytes1.Height(); ++y) {
        for (size_t x=0; x<imageArrayBytes1.Width(); ++x)
          imageArrayBytes2(y,x) = imageArrayBytes1(y,x); 

doesn't fill in the good way (the imageArrayBytes1(y,x) is wrong).
I see some pieces of 1/3 filling. maybe it's a wrong index due to the depth.

What steps will reproduce the problem?
1. Generate msvc 2010 project files
2. Try to compile
3. Fail due to glog

What is the expected output? What do you see instead?
Libmv must compile!

Please use labels and text to provide additional information.
The problem is that glog is not well configured for windows...
Need to regenerate the config.h or something like that.

I cannot compile libmv on linux
What steps will reproduce the problem?
1. get a linux (ubuntu 32b)
2. compile libmv
3. Get this error
[  1%] Built target gflags
[  2%] Built target libmv_test_main
Scanning dependencies of target glog
[  2%] Building CXX object third_party/glog/src/CMakeFiles/glog.dir/demangle.o  


[  2%] Building CXX object third_party/glog/src/CMakeFiles/glog.dir/logging.o   


[  3%] Building CXX object 
third_party/glog/src/CMakeFiles/glog.dir/raw_logging.o                          

[  3%] Building CXX object 
third_party/glog/src/CMakeFiles/glog.dir/signalhandler.o                        

src/third_party/glog/src/signalhandler.cc: In function 'void* 
google::<unnamed>::GetPC(void*)':                                
src/third_party/glog/src/signalhandler.cc:75:28: error: 'REG_RIP' was not 
declared in this scope

it seems to be a known solved issue of glog:
http://code.google.com/p/google-glog/issues/detail?id=46

I think our config_linux.h is wrong.

What steps will reproduce the problem?
1. make

What is the expected output? What do you see instead?

- Expected: cmake calling libraries and dependencies. 
- Actual:   I receive QT and gflag errors (see attached file).

What version of the product are you using? On what operating system?

- Revision 756. MacOSX 10.6 Snow Leopard.

Please provide any additional information below.

Build fails on snow leopard. I searched online. The error is QT related but
I am not sure how to resolve this issue. 

CMAKE_MODULE_PATH = /repos/libmv-read-only/src/CMakeModules
-- Looking for Q_WS_X11
-- Looking for Q_WS_X11 - not found.
-- Looking for Q_WS_WIN
-- Looking for Q_WS_WIN - not found.
-- Looking for Q_WS_QWS
-- Looking for Q_WS_QWS - not found.
-- Looking for Q_WS_MAC
-- Looking for Q_WS_MAC - not found.
-- Found Qt-Version 4.5.3
-- Found OpenSSL: /usr/lib/libssl.dylib
-- Configuring done
-- Generating done
-- Build files have been written to: /repos/libmv-read-only/bin-dbg

Any suggestions would be appreciated. I am running OSX 10.6 and just
installed the latest QT SDK disk image "qt-sdk-mac-opensource-2009.04.dmg".
I also build and installed gflags from source and posted an issue on their
site. My make log file is attached for reference.

Thanks! 

P.S I am interested in contributing to the project and am working out the
details of a geometric min-hash algorithm to evaluate its application on
mobile phones.

What steps will reproduce the problem?
1. Modify the homography H_real in line 104 of autocalibration_test.cc so
that it is different from the identity.
2. Run the autocalibration_test.

What is the expected output? What do you see instead?
The test fails because of a bug in the autocalibration code.  The computed
homography does not seem to be a correct metric update.

What steps will reproduce the problem?
1. building on Ubuntu 11.04 64 bit fails on surf
2.
3.

What is the expected output? What do you see instead?
should build was building before.....

What version of the product are you using? On what operating system?
lastest svn checkout

Please provide any additional information below.
[ 40%] Building CXX object 
libmv/descriptor/CMakeFiles/descriptor.dir/simpliest_descriptor.o
[ 40%] Building CXX object 
libmv/descriptor/CMakeFiles/descriptor.dir/surf_descriptor.o
In file included from 
/home/pilgrims/libmv/libmv/src/libmv/descriptor/surf_descriptor.cc:26:0:
/home/pilgrims/libmv/libmv/src/./libmv/image/integral_image.h: In function 
‘typename TImage::Scalar libmv::BoxIntegral(const TImage&, int, int, int, 
int) [with TImage = Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001>, typename TImage::Scalar = unsigned int]’:
/home/pilgrims/libmv/libmv/src/libmv/descriptor/surf_descriptor.cc:39:70:   
instantiated from ‘float libmv::descriptor::HarrX(const TImage&, int, int, 
int) [with TImage = Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001>]’
/home/pilgrims/libmv/libmv/src/libmv/descriptor/surf_descriptor.cc:137:78:   
instantiated from ‘void libmv::descriptor::MSURFDescriptor(const TImage&, 
const TPointFeature&, libmv::Vecf*) [with int blocks = 4, int samples_per_block 
= 9, TImage = Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001>, TPointFeature = libmv::PointFeature, libmv::Vecf = 
Eigen::Matrix<float, -0x00000000000000001, 1>]’
/home/pilgrims/libmv/libmv/src/libmv/descriptor/surf_descriptor.cc:174:74:   
instantiated from here
/home/pilgrims/libmv/libmv/src/./libmv/image/integral_image.h:84:60: error: no 
matching function for call to ‘min(int&, 
Eigen::PlainObjectBase<Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001, 0, -0x00000000000000001, -0x00000000000000001> 
>::Index)’
/home/pilgrims/libmv/libmv/src/./libmv/image/integral_image.h:85:60: error: no 
matching function for call to ‘min(int&, 
Eigen::PlainObjectBase<Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001, 0, -0x00000000000000001, -0x00000000000000001> 
>::Index)’
/home/pilgrims/libmv/libmv/src/./libmv/image/integral_image.h:86:60: error: no 
matching function for call to ‘min(int, 
Eigen::PlainObjectBase<Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001, 0, -0x00000000000000001, -0x00000000000000001> 
>::Index)’
/home/pilgrims/libmv/libmv/src/./libmv/image/integral_image.h:87:60: error: no 
matching function for call to ‘min(int, 
Eigen::PlainObjectBase<Eigen::Matrix<unsigned int, -0x00000000000000001, 
-0x00000000000000001, 0, -0x00000000000000001, -0x00000000000000001> 
>::Index)’
make[3]: *** [libmv/descriptor/CMakeFiles/descriptor.dir/surf_descriptor.o] 
Error 1
make[3]: Leaving directory `/home/pilgrims/libmv/libmv/bin-dbg'
make[2]: *** [libmv/descriptor/CMakeFiles/descriptor.dir/all] Error 2
make[2]: Leaving directory `/home/pilgrims/libmv/libmv/bin-dbg'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/home/pilgrims/libmv/libmv/bin-dbg'
make: *** [debug] Error 2

Hi, I have an infinite loop in EuclideanResection.

What steps will reproduce the problem?
call libmv::resection::EuclideanResection(x,X,&R,&t) with
x = 
114 584 160 99 112
33 41 50 54 55
X = 
-144.7 89.4 -58.89 -139.7 -162.1
-200.3 -63.25 -103 -160.9 -200.4
443.2 146.4 246.1 392.2 490.7

there is an infinite loop during the SVD decomposition in function 
AbsoluteOrientation (euclidean_resection.cc) at line 64.
The N matrix has nan values...

What is the expected output? What do you see instead?
infinite loop ! 
by the way, shouldn't Eigen check if the matrix has nan values ?

What steps will reproduce the problem?
1. Launch tvr in debug mode.
2. Load the two attached images.
3. Click the Compute features option in the menu.

The program stops because of an assert in UnsafeBoxFilter. The assert
checks that the box integral of a positive image is positive, which is not
the case.

What steps will reproduce the problem?
1. Download images from Issue 1
2. Use TVR to compute SURF points
3. Note that the SURF points are sometimes in homogeneous areas, and that
there are near-duplicates.

What is the expected output? What do you see instead?

SURF should do a better job of finding distinct areas, and should not find
keypoints in homogeneous areas. Also, multiple close keypoints should be
pruned.

Please use labels and text to provide additional information.

What steps will reproduce the problem?
1. make clean
2. make

Linking of bundle_test will fail, arguing undefined references to lapack
functions.

Simply re-running make will solve the problem.

The tangential distortion code seems wrong to me but I don't want to commit a 
fix without testing.

It defines "double radius_squared = radius * radius" when radius is actually 
already squared, so I think we need to remove this and rename radius to 
radius_squared.

Since we only input the first 2 tangential coefficient, maybe we could also 
remove the loop which is supposed to compute higher orders.

I think using tangential_distortion_[0] and tangential_distortion_[1] would be 
clearer than tangential_distortion_.x() and tangential_distortion_.y().

Besides, the radial_distortion code is quite unclear, Since this code shouldn't 
be performance critial, I propose using:
for (int i = 0; i < radial_distortion_.size(); ++i)
 coef_radial += radial_distortion_[i] * pow(radius,i+1);

What steps will reproduce the problem?
1. svn checkout ....
2. cmake -G "Unix Makefiles"
3. make

What do you see instead?

Linking CXX executable image_test
/usr/bin/ld: cannot open output file image_test: >> Is a directory <<

The problem line is ./src/libmv/image/CMakeLists.txt:
IMAGE_TEST(image)

What steps will reproduce the problem?
1. Download libmv-0.1-dev.tar.bz2 from the website.
2. unzip and mkdir build
3. run cmake (successful) then make

What is the expected output? What do you see instead?
I would expect the build to be successful, but I see the following error log:

------------------
[  7%] Building CXX object 
libmv/base/CMakeFiles/scoped_ptr_test.dir/scoped_ptr_test.o
In file included from 
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/ei
gen/Eigen/Core:96:0,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/eigen/Eigen/Array:4,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./libmv/numeric/numeric.h:28,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./testing/testing.h:24,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/scoped_ptr_test.cc:22:
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/ei
gen/Eigen/src/Core/util/Memory.h:280:13: error: ‘ptrdiff_t’ does not name a 
type
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:33:1: warning: 
‘libmv::{anonymous}::ScopedPtr_NullDoesNothing_Test::test_info_’ defined 
but not used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:38:1: warning: 
‘libmv::{anonymous}::ScopedPtr_FreesWhenOutOfScope_Test::test_info_’ 
defined but not used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:47:1: warning: 
‘libmv::{anonymous}::ScopedPtr_Operators_Test::test_info_’ defined but not 
used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:54:1: warning: 
‘libmv::{anonymous}::ScopedPtr_Reset_Test::test_info_’ defined but not used 
[-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:62:1: warning: 
‘libmv::{anonymous}::ScopedPtr_ReleaseAndGet_Test::test_info_’ defined but 
not used [-Wunused-variable]
make[3]: *** [libmv/base/CMakeFiles/scoped_ptr_test.dir/scoped_ptr_test.o] 
Error 1
make[3]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make[2]: *** [libmv/base/CMakeFiles/scoped_ptr_test.dir/all] Error 2
make[2]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make[1]: *** [all] Error 2
make[1]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make: *** [debug] Error 2


then I edited the Memory.h file with #include <stddef.h>, and error vanishes, 
but reoccurs in other files, WHICH ALSO INCLUDES SYSTEM FILES, like below:
--------------------
[  7%] Building CXX object 
libmv/base/CMakeFiles/scoped_ptr_test.dir/scoped_ptr_test.o
In file included from 
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/gt
est/include/gtest/internal/gtest-type-util.h:51:0,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/gtest/include/gtest/internal/gtest-internal.h:57,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./third_party/gtest/include/gtest/gtest.h:65,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./testing/testing.h:25,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/scoped_ptr_test.cc:22:
/usr/include/c++/4.6/cxxabi.h:366:5: error: ‘ptrdiff_t’ does not name a type
/usr/include/c++/4.6/cxxabi.h:442:23: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:454:18: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:464:26: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:490:18: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:496:26: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:535:18: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:541:26: error: ‘ptrdiff_t’ has not been 
declared
/usr/include/c++/4.6/cxxabi.h:561:4: error: ‘ptrdiff_t’ has not been 
declared
In file included from 
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/gt
est/include/gtest/gtest-param-test.h:157:0,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./third_party/gtest/include/gtest/gtest.h:69,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/./testing/testing.h:25,
                 from /home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/scoped_ptr_test.cc:22:
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/third_party/gt
est/include/gtest/internal/gtest-param-util.h:122:11: error: ‘ptrdiff_t’ 
does not name a type
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:33:1: warning: 
‘libmv::{anonymous}::ScopedPtr_NullDoesNothing_Test::test_info_’ defined 
but not used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:38:1: warning: 
‘libmv::{anonymous}::ScopedPtr_FreesWhenOutOfScope_Test::test_info_’ 
defined but not used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:47:1: warning: 
‘libmv::{anonymous}::ScopedPtr_Operators_Test::test_info_’ defined but not 
used [-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:54:1: warning: 
‘libmv::{anonymous}::ScopedPtr_Reset_Test::test_info_’ defined but not used 
[-Wunused-variable]
/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/src/libmv/base/sco
ped_ptr_test.cc:62:1: warning: 
‘libmv::{anonymous}::ScopedPtr_ReleaseAndGet_Test::test_info_’ defined but 
not used [-Wunused-variable]
make[3]: *** [libmv/base/CMakeFiles/scoped_ptr_test.dir/scoped_ptr_test.o] 
Error 1
make[3]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make[2]: *** [libmv/base/CMakeFiles/scoped_ptr_test.dir/all] Error 2
make[2]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make[1]: *** [all] Error 2
make[1]: Leaving directory 
`/home/deepakd/Development/thirdparty/libmv/libmv-0.1.1104-dev/bin-dbg'
make: *** [debug] Error 2



What version of the product are you using? On what operating system?
libmv-0.1-dev.tar.bz2 on Arch linux
libmv-0.1-dev.tar.bz2 on Ubuntu

Please provide any additional information below.

What steps will reproduce the problem?
1. compile in release mode

System:
svn revision 532. (and before)
Ubuntu x86_64
gcc (Ubuntu 4.3.2-1ubuntu12) 4.3.2

Output:
cc1plus: warnings being treated as errors
./libmv/src/third_party/eigen2/Eigen/src/Core/SolveTriangular.h: In member
function ‘bool Eigen::LU<MatrixType>::solve(const
Eigen::MatrixBase<OtherDerived>&, ResultType*) const [with OtherDerived =
Eigen::CwiseUnaryOp<Eigen::ei_scalar_opposite_op<double>,
Eigen::Matrix<double, 3, 1, 0, 3, 1> >, ResultType = libmv::Vec3,
MatrixType = Eigen::Matrix<double, 3, 3, 0, 3, 3>]’:
./libmv/src/third_party/eigen2/Eigen/src/Core/SolveTriangular.h:189: error:
array subscript is above array bounds
compilation terminated due to -Wfatal-errors.

What steps will reproduce the problem?
1. Obtained code from https://github.com/wg-perception/libmv
2. Do the following as per the README 
 upto 
$ make test


What is the expected output? What do you see instead?

The

$ make test

step says:


The following tests FAILED:
     19 - image_converter_test (Failed)
     20 - image_transform_linear_test (OTHER_FAULT)
     46 - homography_test (Failed)
     50 - panography_test (Failed)
     75 - similarity_parameterization_test (Failed)
     79 - euclidean_parameterization_test (Failed)
Errors while running CTest
make: *** [test] Error 8


What version of the product are you using? On what operating system?
Source checked out from  https://github.com/wg-perception/libmv
(upto date as of now)
on Ubuntu 11.10

Please provide any additional information below.

Output from running

 $ make
 $ make release
 $ cd bin-dbg
 $ make test

as per README are attached.

Thanks in advance..

Srimal.

What steps will reproduce the problem?
1. Start ui/tvr
2. Load image
3. Press "compute features"

What is the expected output? What do you see instead?

tvr: /root/libmv-read-
only1/src/third_party/eigen/Eigen/src/Core/MatrixStorage.h:44: 
Eigen::ei_matrix_array<T, Size, MatrixOptions, Align>::ei_matrix_array() 
[with T = float, int Size = 64, int MatrixOptions = 2, bool Align = true]: 
Assertion `(reinterpret_cast<size_t>(array) & 0xf) == 0 && "this assertion 
is explained here: http://eigen.tuxfamily.org/dox/UnalignedArrayAssert.html  
**** READ THIS WEB PAGE !!! ****"' failed.


What version of the product are you using? On what operating system?


Please provide any additional information below.

What steps will reproduce the problem?
Run tests from https://github.com/Matthias-Fauconneau/libmv/.

What is the expected output? What do you see instead?
Tests should pass. CameraIntrinsics fails.

The issue might be related to inverting the 1D function with Levenberg 
optimization.