Hi Frank,
> Just recently I started using darktable and I really like it a lot. It has
> almost everything I need. It would be very cool if one could adjust the view
> to certain information from the metadata. One example is to show for the
> current filmroll only pictures with ISO 100 or in a
OK, pushed to git with those 2 includes (should have checked that
myself).
On Mon, 02 Sep 2013 21:52:52 +0200 (CEST)
Richard Levitte wrote:
> That worked like a charm! (which shows that gdb on my system is
> allowed to attach to other processes, obviously)
>
> BTW, I had to add the following t
That worked like a charm! (which shows that gdb on my system is
allowed to attach to other processes, obviously)
BTW, I had to add the following two lines in src/common/darktable.c to
have waitpid declared properly:
#include
#include
Cheers,
Richard
In message <20130902210410.20d88...@darkho
Everyone,
Just recently I started using darktable and I really like it a lot. It has
almost everything I need. It would be very cool if one could adjust the view
to certain information from the metadata. One example is to show for the
current filmroll only pictures with ISO 100 or in a range be
Am 02.09.2013 21:35, schrieb Alexander Rabtchevich:
> It has no sense.
Of course it has. The problem is, to be really useful the lightness
channel should be blended, too, but independent from the colour. The
color adjustment blend mode is primary intended (at least as far as i
know) to reduce the
IMHO the mask should define the whole effect area - in the other case
there should be another especial mask for blending, which is excessive.
If I use a mask for some area (a face for example), and try to adjust it
with a tonal curve, I definitely don't want to apply the lightness
effect to th
Yup, UI'll try it out asap.
In message <20130902210410.20d88...@darkhorse.paraf.in> on Mon, 2 Sep 2013
21:04:10 +0400, parafin said:
parafin> I have another theory about what's causing the hang, can you please try
parafin> the attached patch? it replaces popen/read/write with fork/exec, so
para
Le 02/09/2013 19:32, johannes hanika a écrit :
> hm. i have something that makes my pictures look acceptable (much more
> highlights reconstructed than with the superold version and less
> clipping than the old version). i guess a couple more samples wouldn't
> hurt before i deploy this.
Sound pro
Hi,
whoever is working on gentoo behind a http-proxy may have a look
at the slightly modified ebuild. (attached)
http instead of git protocol.
naming: -9998 instead of -
have fun
Wolfgang
p.s.:
tar xfz dt.ebuild.tgz
ebuild media-gfx/darktable/darktable-9998.ebuild merge
dt.ebuild.tgz
D
hm. i have something that makes my pictures look acceptable (much more
highlights reconstructed than with the superold version and less clipping
than the old version). i guess a couple more samples wouldn't hurt before i
deploy this.
-jo
On Tue, Sep 3, 2013 at 1:25 AM, johannes hanika wrote:
>
I have another theory about what's causing the hang, can you please try
the attached patch? it replaces popen/read/write with fork/exec, so
that there's no I/O deadlock possible between DT and gdb processes.
On Mon, 2 Sep 2013 15:41:39 +0200
johannes hanika wrote:
> does your system allow gdb to
Wolfgang Goetz wrote:
> Hello,
just a follow-up. ;-)
> tried to compile on Centos6 @work... FAIL!
message of today:
CMake Error at cmake/modules/FindGlib.cmake:33 (message):
Glib version check failed. Version 2.22.5 was found, at least version 2.30
is required
@Simon: please please ple
Hi Rob,
> Do you have options on how you provide your SAN volume? I’ve found
> that providing a SAN volume through ISCSI instead of NFS or CIFS gives
> huge performance benefits when you are executing 1000s of very small
> file operations as the ISCSI volume is treated as a local disk and get
does your system allow gdb to attach to processes? some disable that by
default, which might lead to weird behaviour.
oh, also i'm quite sure we don't restrict ourselves to async-signal-safe
functions in the sigseg handler (man 7 signal). so i guess every once in a
million this is quite prone to d
yeah, sorry that might be my fault. i think i have some, but if you want to
make extra sure i know what you mean an example image would certainly help.
-jo
On Mon, Sep 2, 2013 at 1:40 PM, Moritz Moeller wrote:
> I just noticed that on DT 1.2.2 (Official OS X build) the HR does not
> seem to wor
Hi!
When darktable crashes, or bugs out, or whatever you want to call it,
it starts gdb with a set of commands. Trouble is, this simply does
nothing. Or rather, it just freezes. Processes look like this:
: ; ps auxw | grep darktable
levitte 13570 28.6 10.7 4171824 877656 pts/7 tl 14:08 10
2013/9/2 johannes hanika :
> yes, that should be fine. at least if the xmp is written while holding a
> lock on the dt_image_t struct (dt_image_cache_write_release() writes the xmp
> if called with the right flag). explicitly calling dt_exif_xmp_write does
> not lock anything and will probably soon
I just noticed that on DT 1.2.2 (Official OS X build) the HR does not
seem to work any more, as expected. I use a preset I created a year ago
and which always worked a treat, it get auto-applied to all my DNGs:
method [Reconstruct in LCh]
clipping threshold [1.0]
This completely blo
On Mon, Sep 2, 2013 at 12:18 PM, Pascal Obry wrote:
> 2013/9/2 johannes hanika :
> > how did you trigger the xmp writing? mark 2k images and tag them at the
> same
> > time?
>
> Right. Doing just that.
>
> > that should indeed go into a bg job, where the bg job executes the
> > tagging and will l
Hi Pascal,
Do you have options on how you provide your SAN volume? I've found that
providing a SAN volume through ISCSI instead of NFS or CIFS gives huge
performance benefits when you are executing 1000s of very small file operations
as the ISCSI volume is treated as a local disk and gets t
2013/9/2 johannes hanika :
> how did you trigger the xmp writing? mark 2k images and tag them at the same
> time?
Right. Doing just that.
> that should indeed go into a bg job, where the bg job executes the
> tagging and will lock the image only for a short time while it writes the
> xmp? we'll n
On Mon, Sep 2, 2013 at 12:02 PM, Pascal Obry wrote:
> 2013/9/2 johannes hanika :
> > no, that's exactly the race condition horror scenario i described
> earlier..
> > if we write to that file we want to be sure we do that now and only in
> one
> > thread at a time. if you just put it in a job you
2013/9/2 johannes hanika :
> no, that's exactly the race condition horror scenario i described earlier..
> if we write to that file we want to be sure we do that now and only in one
> thread at a time. if you just put it in a job you'll need to provide
> wrappers which force it to happen now and pu
On Mon, Sep 2, 2013 at 11:21 AM, Jose Carlos Garcia Sogo
wrote:
> To be more specific:
>
> All these modules are calling dt_image_synch_xmp which at the end
> calls dt_exif_xmp_write, which does the actual writing of the file. So
> probably we should call this a job.
>
no, that's exactly the race
To be more specific:
All these modules are calling dt_image_synch_xmp which at the end
calls dt_exif_xmp_write, which does the actual writing of the file. So
probably we should call this a job.
On Mon, Sep 2, 2013 at 11:11 AM, Jose Carlos Garcia Sogo
wrote:
> More than a background job it should
More than a background job it should be done in a thread different to
the UI thread.
On Mon, Sep 2, 2013 at 10:41 AM, Pascal Obry wrote:
> Changing a keyword or metadata trigger the writing of the xmp with
> updated information. Using this on some large collection can take a
> very long time if t
well, the usual.. data safety. we're writing these to be sure stuff on disk
is synched, so if you defer that to some time later in a bg thread it might
not be quite as safe any more. maybe that's fine if we only do it for those
insanely large bulks of images you're talking about.
we would need to
Changing a keyword or metadata trigger the writing of the xmp with
updated information. Using this on some large collection can take a
very long time if the files are on an external drive. During this time
darktable is not responsive to any action.
It wouldn't be an issue if the waiting was 1 or 2
Hi everyone
thank's for the very good work.
I have been trying to contribute by profiling my camera, an olympus epl3
which files do you want me to ship over exactly, and to which email?
best
remy
--
Learn the latest--V
Hi there,
i had a nice idea for a further tool in darktable.
Take a look at the fantastic tool at
http://sourceforge.net/projects/antimidges/
Description:
Antimidges is simple program which allows you to filter all moving
objects from photos and keep only the move less objects. It is useful
when
Il 31/08/2013 17:41, johannes hanika ha scritto:
> could who ever of you can reproduce this reliably try current master
> and tell me whether it's still a problem?
>
> cheers,
> jo
>
Hi jo,
the master now works on my system. Great!
Thank you very much!
Cheers,
Ivan
31 matches
Mail list logo
