I think the reason was that when Visual Studio 6 was used to build VTK
it used to run out of stack space necessitating the large value that
needed to be set.
I am pretty sure that this was the case because I think I was one of
the people who noted this problem in the "old days"!
Andrew
On Mon,
Sorry, forgot to forward this to the list as well.
2009/4/13 Daniel Stonier :
> Just a last piece of information, the guys at ccache have just added a
> patch which fixes ccache...well, its more of a hack. Just tells ccache
> to avoid cacheing anything with a -fprofile flag.
>
> On my gentoo machi
On Sun, Apr 12, 2009 at 10:06 PM, Daniel Stonier wrote:
> Ian, are you using ccache?
No I just really didn't know where to start. Its not obvious from the
wiki that just setting the correct gcc flags triggers a coverage test.
:) I'll try out what Alex says tomorrow and edit the wiki accordingly.
Ian, are you using ccache?
Just found that buggers the process up completely. If you're using a
ccache link to the compiler, then the two-step compile-link process
fails to build the gcno coverage file (or misplaces it). I removed my
ccache links and did it again manually and presto, the coverage
if you using gcc, compile you code with the right options, and you re
good.
http://www.cmake.org/Wiki/CTest:Coverage
Coverage With C++
Currently coverage is only supported on gcc compiler. To perform
coverage test, make sure that your code is built with debug symbols,
without optimization,
So I was able to solve my problem of duplicate python target names
generated from SWIG interface files:
PyTrilinos/NOX/___init__.so
PyTrilinos/NOX/Epetra/___init__.so
by overhauling UseSWIG.cmake to take advantage of the OUTPUT_NAME
property of the SET_TARGET_PROPERTIES() function.
My
Actually, I just discovered an interesting thing.
Cmake generally builds things in two steps, first compile the objects,
then link. Gcc can do it this way, or do it with a one shot command. I
just tried replicating what cmake does manually with gcc on the
command line and nope, no .gcno (gcov) fil
On Sun, Apr 12, 2009 at 8:51 PM, Daniel Stonier wrote:
> I'm having a few troubles understanding how to get CTest setup. The
> wiki is a bit piecemeal and just need a pointer in the right
> direction.
>
> Currently, the ctest results are passing for a simple hello world
> project, but I can't get
I'm having a few troubles understanding how to get CTest setup. The
wiki is a bit piecemeal and just need a pointer in the right
direction.
Currently, the ctest results are passing for a simple hello world
project, but I can't get coverage working. When I run , it will
make all test
Andreas Pakulat wrote:
On 07.04.09 20:29:49, Yevgen Muntyan wrote:
Hi,
I want to convert an autotools project to cmake. But, I don't know how
to solve the problem of the
convenience libraries. The issue is: the project builds a shared library
of bunch of libtool convenience
libraries. Ea
Don't know the reason behind this, but this is how we unset it:
#
# Remove /STACK:1000 set by CMake. This value for stack size
# is very high, limiting the number of threads we can spawn.
# Default value used by Windows is 1MB which is good enough.
#
STRING(REGEX REPLAC
A new idea to extend CMake's scripting capabilities:
JavaScript is much more familiar to C/C++ developers
than the current macro language and others (Lua).
Qt ships with ready to use JavaScript support and is now licensed under LGPL.
But most important, Qt 4.5 comes with a JavaScript debugger!
Th
Hi,
I don't understand the reason behind this. We've a very huge memory
footprint because of this on kde4/windows - kded4.exe has ~30 Threads
which means 300MB of stack!
Is there an easy way to change this?
Thx,
Christian
___
Powered by www.kitware.co
13 matches
Mail list logo
