Re: quick fix for brew build on Sierra /XCode8 (was: [darktable-dev] Warning: git master now uses rawspeed as submodule !)

[parafin]

Changes to cmake files is probably the culprit - rawspeed now doesn't inherit 
CFLAGS from darktable and so is missing include path for libjpeg.

P.S.
About switching from stable to master branch and back - there're also issues 
with cherry-picking commits for some reason. We need to come up with correct 
and clean way of branch checkout… Maybe git clean is needed…

> On 4 Jan 2017, at 05:05, Wolfgang Goetz  wrote:
> 
> Roman Lebedev wrote:
> 
>> On Tue, Jan 3, 2017 at 1:46 AM, Wolfgang Goetz
>>  wrote:
>>> MacOS/brew broken:
>>> ...
>>> [  2%] Building CXX object
>>> src/external/rawspeed/RawSpeed/CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o
>>> In file included
>>> from 
>>> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/ArwDecoder.cpp:1:
>>>  
>>> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/StdAfx.h:97:10:
>>> fatal error: 'jpeglib.h' file not found #include "jpeglib.h" ^
>>> 1 error generated.  
>> Hmm, you do have jpeg or jpeg-turbo formula installed?
> 
> 
> 
>> Because travis-ci osx builds do not fail, e.g.
>> https://travis-ci.org/darktable-org/rawspeed/jobs/188323509
> 
> 
> 
> to make it short:  
> http://stackoverflow.com/questions/14525898/cmake-doesnt-set-xcodes-user-include-path
> 
> not sure if it's the right place, looks more like sledgehammer:
> 
> 
> diff --git a/CMakeLists.txt b/CMakeLists.txt
> index 6b7a1530f..d8e57008b 100644
> --- a/CMakeLists.txt
> +++ b/CMakeLists.txt
> @@ -32,7 +32,7 @@ endif()
> IF(DEFINED DT_FORCE_CXX_COMPILER)
>   set(CMAKE_CXX_COMPILER ${DT_FORCE_CXX_COMPILER})
> endif()
> -
> +include_directories(/usr/local/include)
> include(CheckCCompilerFlag)
> include(TestBigEndian)
> 
> 
> 
> build OK, dt 2.3.0+142 running :-)
> 
> 
> 
> kind regards
> Wolfgang
> 
> 
> 
> 
> 
> 
> the long way:
> 
> 
> travis is cool. 
> attention: default is OSX 10.11/XCode 7.3.1 
> 
> i'm on real hardware more like osx_image: xcode8.2  
> exact:macOS Sierra Version 10.12.2, XCode Version 8.2.1 (not avail in travis)
> 
> 
> 
> brew is cool. (in darktable: 'brew bundle' for dependencies, atm just 
> updating to gcc-6.3.0, this will take some time!)
> 
> tags/release-2.2.1 build is ok.
> 
> /usr/local/include/jpeglib.h exists.
> 
> playing around with CC and CXX environments doesnt help. (fun to see gcc-6 as 
> CC and clang as CXX)
> 
> sandboxing: same error
> export PATH=.:$PATH
> brew bundle exec build.sh
> 
> 
> 
> 
> 
> 
> 
> VERBOSE=1 
> 
> 
> with submodule:
> 
> [  3%] Building CXX object 
> src/external/rawspeed/RawSpeed/CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o
> cd /Users/wolfganggoetz/repos/darktable/build/src/external/rawspeed/RawSpeed 
> && 
> /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
>-DDEBUG -DGDK_DISABLE_DEPRECATED -DGTK_DISABLE_DEPRECATED 
> -DGTK_DISABLE_SINGLE_INCLUDES -DOS_OBJECT_USE_OBJC=0 -D_XOPEN_SOURCE=700 
> -D__GDK_KEYSYMS_COMPAT_H__ -isystem 
> /usr/local/Cellar/pugixml/1.8.1/include/pugixml-1.8  -D_DARWIN_C_SOURCE -Wall 
> -fno-strict-aliasing -Wformat -Wformat-security -Wshadow -Wtype-limits -Wvla 
> -Wno-error=varargs -Wframe-larger-than=32768 -Wlarger-than=524288 -std=c++11 
> -fno-strict-aliasing -Wall -fno-strict-aliasing -Wformat -Wformat-security 
> -Wshadow -Wtype-limits -Wvla -Wno-error=varargs -Wframe-larger-than=32768 
> -Wlarger-than=524288 -O2 -g -DNDEBUG -O2 -isysroot 
> /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.12.sdk
>  -fPIC   -march=native -g3 -ggdb3 -o CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o 
> -c 
> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/ArwDecoder.cpp
> In file included from 
> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/AriDecoder.cpp:1:
> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/StdAfx.h:97:10:
> fatal error: 'jpeglib.h' file not found
> 
> 
> 
> 
> good output from release-2.2.1:
> 
> [  2%] Building CXX object 
> src/external/rawspeed/CMakeFiles/rawspeed.dir/RawSpeed/ArwDecoder.cpp.o
> cd /Users/wolfganggoetz/repos/darktable_old/build/src/external/rawspeed && 
> /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
>-DGDK_DISABLE_DEPRECATED -DGDK_VERSION_MIN_REQUIRED=GDK_VERSION_3_14 
> -DGLIB_VERSION_MAX_ALLOWED=GLIB_VERSION_MIN_REQUIRED 
> -DGLIB_VERSION_MIN_REQUIRED=GLIB_VERSION_2_40 -DGTK_DISABLE_DEPRECATED 
> -DGTK_DISABLE_SINGLE_INCLUDES -DHAVE_CONFIG_H -DHAVE_GPHOTO2 
> -DHAVE_GPHOTO_25_OR_NEWER -DHAVE_GRAPHICSMAGICK -DHAVE_HTTP_SERVER 
> -DHAVE_KWALLET -DHAVE_LENSFUN -DHAVE_LIBSECRET -DHAVE_OPENCL -DHAVE_OPENEXR 
> -DHAVE_OPENJPEG -DMAC_INTEGRATION -DOS_OBJECT_USE_OBJC=0 -DUSE_LUA 
> -D_XOPEN_SOURCE=700 -D__GDK_KEYSYMS_COMPAT_H__ 
> -I/Users/wolfganggoetz/repos/darktable_old/src 
> -I/Users/wolfganggoetz/repos/darktable_old/src/external -isystem 
> /usr/local/include/glib-2.0 -isystem 
> 

quick fix for brew build on Sierra /XCode8 (was: [darktable-dev] Warning: git master now uses rawspeed as submodule !)

[Wolfgang Goetz]

Roman Lebedev wrote:

> On Tue, Jan 3, 2017 at 1:46 AM, Wolfgang Goetz
>  wrote:
> > MacOS/brew broken:
> > ...
> > [  2%] Building CXX object
> > src/external/rawspeed/RawSpeed/CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o
> > In file included
> > from 
> > /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/ArwDecoder.cpp:1:
> >  
> > /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/StdAfx.h:97:10:
> > fatal error: 'jpeglib.h' file not found #include "jpeglib.h" ^
> > 1 error generated.  
> Hmm, you do have jpeg or jpeg-turbo formula installed?



> Because travis-ci osx builds do not fail, e.g.
> https://travis-ci.org/darktable-org/rawspeed/jobs/188323509



to make it short:  
http://stackoverflow.com/questions/14525898/cmake-doesnt-set-xcodes-user-include-path

not sure if it's the right place, looks more like sledgehammer:


diff --git a/CMakeLists.txt b/CMakeLists.txt
index 6b7a1530f..d8e57008b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -32,7 +32,7 @@ endif()
 IF(DEFINED DT_FORCE_CXX_COMPILER)
   set(CMAKE_CXX_COMPILER ${DT_FORCE_CXX_COMPILER})
 endif()
-
+include_directories(/usr/local/include)
 include(CheckCCompilerFlag)
 include(TestBigEndian)



build OK, dt 2.3.0+142 running :-)



kind regards
Wolfgang






the long way:


travis is cool. 
attention: default is OSX 10.11/XCode 7.3.1 

i'm on real hardware more like osx_image: xcode8.2  
exact:macOS Sierra Version 10.12.2, XCode Version 8.2.1 (not avail in travis)



brew is cool. (in darktable: 'brew bundle' for dependencies, atm just updating 
to gcc-6.3.0, this will take some time!)

tags/release-2.2.1 build is ok.

/usr/local/include/jpeglib.h exists.

playing around with CC and CXX environments doesnt help. (fun to see gcc-6 as 
CC and clang as CXX)

sandboxing: same error
export PATH=.:$PATH
brew bundle exec build.sh







VERBOSE=1 


with submodule:

[  3%] Building CXX object 
src/external/rawspeed/RawSpeed/CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o
cd /Users/wolfganggoetz/repos/darktable/build/src/external/rawspeed/RawSpeed && 
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
   -DDEBUG -DGDK_DISABLE_DEPRECATED -DGTK_DISABLE_DEPRECATED 
-DGTK_DISABLE_SINGLE_INCLUDES -DOS_OBJECT_USE_OBJC=0 -D_XOPEN_SOURCE=700 
-D__GDK_KEYSYMS_COMPAT_H__ -isystem 
/usr/local/Cellar/pugixml/1.8.1/include/pugixml-1.8  -D_DARWIN_C_SOURCE -Wall 
-fno-strict-aliasing -Wformat -Wformat-security -Wshadow -Wtype-limits -Wvla 
-Wno-error=varargs -Wframe-larger-than=32768 -Wlarger-than=524288 -std=c++11 
-fno-strict-aliasing -Wall -fno-strict-aliasing -Wformat -Wformat-security 
-Wshadow -Wtype-limits -Wvla -Wno-error=varargs -Wframe-larger-than=32768 
-Wlarger-than=524288 -O2 -g -DNDEBUG -O2 -isysroot 
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.12.sdk
 -fPIC   -march=native -g3 -ggdb3 -o CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o 
-c 
/Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/ArwDecoder.cpp
In file included from 
/Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/AriDecoder.cpp:1:
/Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/StdAfx.h:97:10:
fatal error: 'jpeglib.h' file not found




good output from release-2.2.1:

[  2%] Building CXX object 
src/external/rawspeed/CMakeFiles/rawspeed.dir/RawSpeed/ArwDecoder.cpp.o
cd /Users/wolfganggoetz/repos/darktable_old/build/src/external/rawspeed && 
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
   -DGDK_DISABLE_DEPRECATED -DGDK_VERSION_MIN_REQUIRED=GDK_VERSION_3_14 
-DGLIB_VERSION_MAX_ALLOWED=GLIB_VERSION_MIN_REQUIRED 
-DGLIB_VERSION_MIN_REQUIRED=GLIB_VERSION_2_40 -DGTK_DISABLE_DEPRECATED 
-DGTK_DISABLE_SINGLE_INCLUDES -DHAVE_CONFIG_H -DHAVE_GPHOTO2 
-DHAVE_GPHOTO_25_OR_NEWER -DHAVE_GRAPHICSMAGICK -DHAVE_HTTP_SERVER 
-DHAVE_KWALLET -DHAVE_LENSFUN -DHAVE_LIBSECRET -DHAVE_OPENCL -DHAVE_OPENEXR 
-DHAVE_OPENJPEG -DMAC_INTEGRATION -DOS_OBJECT_USE_OBJC=0 -DUSE_LUA 
-D_XOPEN_SOURCE=700 -D__GDK_KEYSYMS_COMPAT_H__ 
-I/Users/wolfganggoetz/repos/darktable_old/src 
-I/Users/wolfganggoetz/repos/darktable_old/src/external -isystem 
/usr/local/include/glib-2.0 -isystem /usr/local/include/../lib/glib-2.0/include 
-isystem /usr/local/Cellar/gtk+3/3.22.5/include/gtk-3.0 -isystem 
/usr/local/Cellar/glib/2.50.2/include/gio-unix-2.0 -isystem 
/usr/local/Cellar/cairo/1.14.8/include/cairo -isystem 
/usr/local/Cellar/libepoxy/1.3.1/include -isystem 
/usr/local/Cellar/pango/1.40.3/include/pango-1.0 -isystem 
/usr/local/Cellar/harfbuzz/1.3.4_1/include/harfbuzz -isystem 
/usr/local/Cellar/atk/2.22.0/include/atk-1.0 -isystem 
/usr/local/Cellar/pixman/0.34.0/include/pixman-1 -isystem 
/usr/local/Cellar/fontconfig/2.12.1_2/include -isystem 
/usr/local/opt/freetype/include/freetype2 -isystem 
/usr/local/Cellar/libpng/1.6.27/include/libpng16 -isystem 

Re: [darktable-dev] An effort-less way to provide DT to advanced Windows users

[Joel Brunetti]

Martin,

I'm not a Darktable developer but I do subscribe to the mailing list
and see how often people bring up the idea of a Windows port on this
list.
I'm sure the developers are tired of dealing with this so let me share
with you my understanding if this and their frustration.

Surely in your search for a solution you would have read this blog
post: http://www.darktable.org/2015/07/why-dont-you-provide-a-windows-build/
If you read it carefully you will see that the challenge of compiling
Darktable on Windows is not what is preventing an official port. It is
the ongoing testing and support the developers will feel forced to
provide that prevents this. None of the core developers have a
personal need for this port and the Darktable project, as a
non-commercial piece of software, is driven by the personal needs of
its developers.

If you are serious about wanting a Windows Darktable port you need to
demonstrate to the core developers that you are reliable and committed
to the Darktable project. I see two ways of doing so:
1. Become a Darktable developer. This means not simply for the Windows
port but develop features and fix existing bugs. Once you are a
reliable contributor that is trusted then the core developers may be
open to an official port for which you will be responsible for testing
and maintaining.
2. Create a fork as you mentioned with its own webpage and issue
tracker. Show commitment to not just providing a fork or build
instructions but to providing ongoing maintenance and support to a
fork that tracks the official Darktable repository. Given time the
core developers may feel fit to merge the two projects should you
still be open to that.

Good luck!

On 3 January 2017 at 14:10, Martin Pecka  wrote:
> I wasn't sure if it was irony or not, but now I am... And I'm sad, because
> I've brought a completely new viewpoint in the Windows-Darktable problem.
>
> Why not take this in a constructive way? Should we (Windows users) create a
> fork of DT with its own webpage and issue tracker?
>
>
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>



-- 
Joel Brunetti
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] An effort-less way to provide DT to advanced Windows users

[Martin Pecka]

I wasn't sure if it was irony or not, but now I am... And I'm sad, 
because I've brought a completely new viewpoint in the Windows-Darktable 
problem.


Why not take this in a constructive way? Should we (Windows users) 
create a fork of DT with its own webpage and issue tracker?


___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] An effort-less way to provide DT to advanced Windows users

[Roman Lebedev]

btw, from now on, each time such mail appears in our maillsts, i'll be
pushing the moment i start looking at that pr by one month.

On Tue, Jan 3, 2017 at 11:43 PM, Roman Lebedev  wrote:
> On Tue, Jan 3, 2017 at 11:32 PM, Martin Pecka  wrote:
>> Hello DT developers.
>>
>> I know there is some discussion about Windows ports of DT. And I even know
>> there is a PR that looks promising and is adding Windows support the
>> "proper" and user-friendly way. However, it still seems to me you're
>> undecided if the Windows port is a wanted "feature" or not.
>> But today I've tested a way that allows to compile DT on Windows 10 with no
>> code modifications at all (well, with one character being edited in
>> src/CMakeLists.txt). Which means no additional burden for the regular
>> Linux/Mac developers. My suggestion doesn't result in an installer or
>> binary, but just in a compilation guide, where it can be clearly written
>> that bugreports for this DT "distribution" will be ignored. And since every
>> user that compiles DT on Windows would have to read this guide, he should
>> also read and remember not to post bugreports.
>>
>> My approach uses the Windows Subsystem for Linux added in Win10 (
>> https://en.wikipedia.org/wiki/Windows_Subsystem_for_Linux ), which is "a
>> compatibility layer for running Linux binary executables (in ELF format)
>> natively on Windows 10". This practically means you can run bash in your
>> Windows console, and that bash runs in a fictional Ubuntu 14.04. But this
>> "fictional Ubuntu" can do most of the things real Ubuntu can do, most
>> notably installing software via apt-get and compiling using gcc. I was
>> really impressed by the subset of things that are working seamlessly.
>>
>> Together with Xming (X server port for Windows), this allows for quite easy
>> compilation and running of Darktable on Windows.
>>
>> I've recorded the needed steps in the following Gist:
>> https://gist.github.com/peci1/222a6bdfa028aa043f9c509c78bf9bc4 . It's mostly
>> just about adding newer versions of libraries to the fictional Trusty
>> system, and then using the standard Ubuntu compilation guide from DT wiki.
>> (I think it's generally also a guide on how to build DT on real Trusty
>> system).
>>
>> Recent news (
>> http://www.omgubuntu.co.uk/2016/10/windows-10-linux-subsystem-ubuntu-16-04 )
>> mentioned it will even be possible to get Xenial instead of Trusty on
>> Windows. That would even mean darktable could be compiled on Windows with
>> exactly the steps from Xenial compilation guide. However, I'm not able to
>> test this ATM (it's not yet released to public, and I run some other
>> software that requires Trusty).
>>
>> My question/suggestion is: should I add this guide to the official DT wiki?
> Forcing something down our throat sure will help. Go on.
>
> Roman.
>> ___
>> darktable developer mailing list
>> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] An effort-less way to provide DT to advanced Windows users

[Roman Lebedev]

On Tue, Jan 3, 2017 at 11:32 PM, Martin Pecka  wrote:
> Hello DT developers.
>
> I know there is some discussion about Windows ports of DT. And I even know
> there is a PR that looks promising and is adding Windows support the
> "proper" and user-friendly way. However, it still seems to me you're
> undecided if the Windows port is a wanted "feature" or not.
> But today I've tested a way that allows to compile DT on Windows 10 with no
> code modifications at all (well, with one character being edited in
> src/CMakeLists.txt). Which means no additional burden for the regular
> Linux/Mac developers. My suggestion doesn't result in an installer or
> binary, but just in a compilation guide, where it can be clearly written
> that bugreports for this DT "distribution" will be ignored. And since every
> user that compiles DT on Windows would have to read this guide, he should
> also read and remember not to post bugreports.
>
> My approach uses the Windows Subsystem for Linux added in Win10 (
> https://en.wikipedia.org/wiki/Windows_Subsystem_for_Linux ), which is "a
> compatibility layer for running Linux binary executables (in ELF format)
> natively on Windows 10". This practically means you can run bash in your
> Windows console, and that bash runs in a fictional Ubuntu 14.04. But this
> "fictional Ubuntu" can do most of the things real Ubuntu can do, most
> notably installing software via apt-get and compiling using gcc. I was
> really impressed by the subset of things that are working seamlessly.
>
> Together with Xming (X server port for Windows), this allows for quite easy
> compilation and running of Darktable on Windows.
>
> I've recorded the needed steps in the following Gist:
> https://gist.github.com/peci1/222a6bdfa028aa043f9c509c78bf9bc4 . It's mostly
> just about adding newer versions of libraries to the fictional Trusty
> system, and then using the standard Ubuntu compilation guide from DT wiki.
> (I think it's generally also a guide on how to build DT on real Trusty
> system).
>
> Recent news (
> http://www.omgubuntu.co.uk/2016/10/windows-10-linux-subsystem-ubuntu-16-04 )
> mentioned it will even be possible to get Xenial instead of Trusty on
> Windows. That would even mean darktable could be compiled on Windows with
> exactly the steps from Xenial compilation guide. However, I'm not able to
> test this ATM (it's not yet released to public, and I run some other
> software that requires Trusty).
>
> My question/suggestion is: should I add this guide to the official DT wiki?
Forcing something down our throat sure will help. Go on.

Roman.
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Roman Lebedev]

On Tue, Jan 3, 2017 at 10:11 PM, johannes hanika  wrote:
> yeah that is annoying. is it a good idea to add that file to
> .gitignore? or is this indeed the mechanism to update the submodule?
Looking at 
https://github.com/darktable-org/darktable/commit/3b4c49d16b859758ddeccdf8a261719ac3f0c5f9.patch
i'm not seeing any other changes that would contain the rawspeed repo commit,
so that is indeed the the mechanism to update the submodule.

As for rawspeed being separate library, it will add one more
cmake&& install
chain, which is also not cool.

> -jo
Roman.

> On Wed, Jan 4, 2017 at 7:57 AM, Ulrich Pegelow
>  wrote:
>> Am 03.01.2017 um 17:29 schrieb Roman Lebedev:
>>>
>>> On Tue, Jan 3, 2017 at 7:22 PM, Ulrich Pegelow

 What now?
>>>
>>> $ git checkout -f darktable-2.2.x
>>>
>>
>> Yep, works. But then:
>>
>> $ git checkout master
>> M   src/external/rawspeed
>> Switched to branch 'master'
>>
>> Which means we generate a pseudo modification in src/external/rawspeed:
>>
>> diff --git a/src/external/rawspeed b/src/external/rawspeed
>> --- a/src/external/rawspeed
>> +++ b/src/external/rawspeed
>> @@ -1 +1 @@
>> -Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6
>> +Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6-dirty
>>
>>
>> If we always need to force a checkout there is quite some risk of
>> accidentally loosing changes. Forcing a checkout will just delete any
>> uncommited modification.
>>
>> Ulrich
>>
>>
>>
>> ___
>> darktable developer mailing list
>> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>>
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

[darktable-dev] Re: compile DT with new exiv2 version

[Matthias Bodenbinder]

Am 03.01.2017 um 17:31 schrieb c...@0n0e.com:
> On 01/03/2017 03:22 AM, Matthias Bodenbinder wrote:
>> Hello,
>>
>> I am running Mint LMDE and can not easily install exiv2 0.25 from testing 
>> because it has too many dependencies.
>> I compiled version 0.25 by myself and installed it in /home/software/exiv2. 
>> In parallel I still have the default exiv2 version 0.24.
>>
>> How can I tell my self compiled darktable to use the new exvi2 0.25? It is 
>> still grabing the old version:
>> libexiv2.so.13 => /usr/lib/x86_64-linux-gnu/libexiv2.so.13
> 
> Does the following work? (Change filepaths and filenames as needed.)
> 
> $ export LD_PRELOAD=/home/software/exiv2/libexiv2.so.13
> $ /opt/darktable/bin/darktable

That is interesting. With this setting it does ot work either but I see a 
different behaviour. The lens identifier in DT is left blank and on the 
commandline I see the following errors during import of the CR2 files:

[exiv2] /home/matthias/Bilder/Fotos/tmp/16-35/IMG_0972.CR2: Image type 
/home/matthias/Bilder/Fotos/tmp/16-35/IMG_0972.CR2 is not supported
[exiv2] /home/matthias/Bilder/Fotos/tmp/16-35/IMG_0972.CR2: Image type 
/home/matthias/Bilder/Fotos/tmp/16-35/IMG_0972.CR2 is not supported
[exiv2] /home/matthias/Bilder/Fotos/tmp/16-35/IMG_0983.CR2: Image type 
/home/matthias/Bilder/Fotos/tmp/16-35/IMG_0983.CR2 is not supported

Weird.
Matthias



___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[johannes hanika]

yeah that is annoying. is it a good idea to add that file to
.gitignore? or is this indeed the mechanism to update the submodule?

-jo

On Wed, Jan 4, 2017 at 7:57 AM, Ulrich Pegelow
 wrote:
> Am 03.01.2017 um 17:29 schrieb Roman Lebedev:
>>
>> On Tue, Jan 3, 2017 at 7:22 PM, Ulrich Pegelow
>>>
>>> What now?
>>
>> $ git checkout -f darktable-2.2.x
>>
>
> Yep, works. But then:
>
> $ git checkout master
> M   src/external/rawspeed
> Switched to branch 'master'
>
> Which means we generate a pseudo modification in src/external/rawspeed:
>
> diff --git a/src/external/rawspeed b/src/external/rawspeed
> --- a/src/external/rawspeed
> +++ b/src/external/rawspeed
> @@ -1 +1 @@
> -Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6
> +Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6-dirty
>
>
> If we always need to force a checkout there is quite some risk of
> accidentally loosing changes. Forcing a checkout will just delete any
> uncommited modification.
>
> Ulrich
>
>
>
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Ulrich Pegelow]

Am 03.01.2017 um 17:29 schrieb Roman Lebedev:

On Tue, Jan 3, 2017 at 7:22 PM, Ulrich Pegelow

What now?

$ git checkout -f darktable-2.2.x



Yep, works. But then:

$ git checkout master
M   src/external/rawspeed
Switched to branch 'master'

Which means we generate a pseudo modification in src/external/rawspeed:

diff --git a/src/external/rawspeed b/src/external/rawspeed
--- a/src/external/rawspeed
+++ b/src/external/rawspeed
@@ -1 +1 @@
-Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6
+Subproject commit 2452e1afd18e06710decce6961a122f43120bfb6-dirty


If we always need to force a checkout there is quite some risk of 
accidentally loosing changes. Forcing a checkout will just delete any 
uncommited modification.


Ulrich


___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] compile DT with new exiv2 version

[code]

On 01/03/2017 03:22 AM, Matthias Bodenbinder wrote:
> Hello,
> 
> I am running Mint LMDE and can not easily install exiv2 0.25 from testing 
> because it has too many dependencies.
> I compiled version 0.25 by myself and installed it in /home/software/exiv2. 
> In parallel I still have the default exiv2 version 0.24.
> 
> How can I tell my self compiled darktable to use the new exvi2 0.25? It is 
> still grabing the old version:
> libexiv2.so.13 => /usr/lib/x86_64-linux-gnu/libexiv2.so.13

Does the following work? (Change filepaths and filenames as needed.)

$ export LD_PRELOAD=/home/software/exiv2/libexiv2.so.13
$ /opt/darktable/bin/darktable

> Your help is appreciated.
> Thank you
> Matthias
> 
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
> 

___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Roman Lebedev]

On Tue, Jan 3, 2017 at 7:22 PM, Ulrich Pegelow
 wrote:
> Am 02.01.2017 um 22:27 schrieb Roman Lebedev:
>>
>> Hi everyone.
>>
>> Starting with today, darktable git master tracks rawspeed
>> library as a submodule.
>>
>> Which means, after updating darktable repo (git pull/git fetch),
>> and before building, you now need to make sure that the
>> submodule is updated to.
>>
>
> Hmm...
>
> $ git pull
> $ git submodule init
> $ git submodule update
> $ ./build ... etc -> all OK
> $ git checkout darktable-2.2.x
> error: The following untracked working tree files would be overwritten by
> checkout:
> src/external/rawspeed/.gitattributes
> src/external/rawspeed/.gitignore
> src/external/rawspeed/CMakeLists.txt
> src/external/rawspeed/LICENSE
> src/external/rawspeed/README.md
> [ ... many more ... ]
> src/external/rawspeed/RawSpeed/MosDecoder.h
> src/external/rawspeed/RawSpeed/MrwDecoder.cpp
> src/external/rawspeed/RawSpeed/MrwDeco
> Aborting
> $
>
> What now?
$ git checkout -f darktable-2.2.x

> Ulrich
Roman.

>>
>> And day-to-day, after git pull for darktable
>> (i'm not 100% sure about the exact command,
>> someone with more knowledge please do chime in)
>> $ cd src/external/rawspeed && git pull && \
>>git checkout -f && git submodule update -f
>>
>> Roman.
>>
>> ___
>> darktable developer mailing list
>> to unsubscribe send a mail to
>> darktable-dev+unsubscr...@lists.darktable.org
>>
>>
>
> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

[darktable-dev] Re: HiPPI support on MacOs

[Bastian A.]

Hi everyone,

the updated release (2.2.1+3) resolves low resolution issue I reported
earlier.

Thanks,
Bastian

On Tue, Dec 27, 2016 at 9:26 PM, Bastian A.  wrote:

> Hi everyone,
>
> thanks for the Christmas update.
> I am using darktable 2.2.0 on MacOs 10.12.2; since the latest update parts
> of the Ui  are rendered in low resolution similar to how things were
> rendered back in version 1.4. I've deleted the saved application state
> folder and the darktablerc file to see if there was a problem with settings
> left over from the older version, but this had no effect on the Ui
> resolution.
>
> Has anyone else experienced this issue?
>
> Best,
> Bastian
>

___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Terry Duell]

On Tue, 03 Jan 2017 19:23:43 +1100, Roman Lebedev   
wrote:



On Tue, Jan 3, 2017 at 2:07 AM, Terry Duell  wrote:
On Tue, 03 Jan 2017 09:46:23 +1100, Wolfgang Goetz  


wrote:


Terry Duell wrote:



[snip]



That does the trick...but the build still fails here for another reason.
Fedora 25 is still on LLVM 3.8.0 but 3.9 is now required.

It is not *required*. It is optional.
It is used to test-compile opencl kernels. That is all.


OK, thanks, I'll see if I can find why that error is being reported.

Cheers,
--
Regards,
Terry Duell
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Roman Lebedev]

On Tue, Jan 3, 2017 at 1:46 AM, Wolfgang Goetz  wrote:
> MacOS/brew broken:
> ...
> [  2%] Building CXX object
> src/external/rawspeed/RawSpeed/CMakeFiles/rawspeed.dir/ArwDecoder.cpp.o
> In file included
> from 
> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/ArwDecoder.cpp:1:
>  
> /Users/wolfganggoetz/repos/darktable/src/external/rawspeed/RawSpeed/StdAfx.h:97:10:
> fatal error: 'jpeglib.h' file not found #include "jpeglib.h" ^
> 1 error generated.
Hmm, you do have jpeg or jpeg-turbo formula installed?
Because travis-ci osx builds do not fail, e.g.
https://travis-ci.org/darktable-org/rawspeed/jobs/188323509

> kind regards
> Wolfgang
Roman.
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

Re: [darktable-dev] Warning: git master now uses rawspeed as submodule !

[Roman Lebedev]

On Tue, Jan 3, 2017 at 2:07 AM, Terry Duell  wrote:
> On Tue, 03 Jan 2017 09:46:23 +1100, Wolfgang Goetz 
> wrote:
>
>> Terry Duell wrote:
>
>
> [snip]
>
>>> Well that seems to have buggered things up :-)
>>> I suspect that the above commands might not be right...attempting to
>>> build results in a message that src/external/rawspeed has no
>>> CMakeLists.txt, and in fact it is an empty dir.
>>> Hopefully, someone who does know how to do this can educate us poor
>>> ignorant sods real soon.
>>>
>>> Cheers,
>>
>>
>>
>> Greetings,
>> this wfm:
>>
>> git pull
>> git submodule init
>> git submodule update
>
>
> That does the trick...but the build still fails here for another reason.
> Fedora 25 is still on LLVM 3.8.0 but 3.9 is now required.
It is not *required*. It is optional.
It is used to test-compile opencl kernels. That is all.

> Thanks for your help.
>
> Cheers,
> --
> Regards,
> Terry Duell
Roman.

> ___
> darktable developer mailing list
> to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
>
___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org

[darktable-dev] compile DT with new exiv2 version

[Matthias Bodenbinder]

Hello,

I am running Mint LMDE and can not easily install exiv2 0.25 from testing 
because it has too many dependencies.
I compiled version 0.25 by myself and installed it in /home/software/exiv2. In 
parallel I still have the default exiv2 version 0.24.

How can I tell my self compiled darktable to use the new exvi2 0.25? It is 
still grabing the old version:
libexiv2.so.13 => /usr/lib/x86_64-linux-gnu/libexiv2.so.13

Your help is appreciated.
Thank you
Matthias

___
darktable developer mailing list
to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org