diter' po/fr.po
>
> #, c-format
> msgid "edit `%s' for module `%s'"
> msgstr "éditer `%s' pour le module `%s'"
Même problème en anglais alors?
Je ne sais pas trop quoi faire dans ce cas! Que proposerais-tu?
--
Pasc
anging all those to standard forward quotes is
not a translation issue only!
Pascal.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
--
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Le 20/11/2013 11:59, Tobias Ellinghaus a écrit :
> Just "darktable" should never be capitalized.
Even when first word in a sentence?
- --
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagin
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Le 20/11/2013 18:48, Tobias Ellinghaus a écrit :
> Am Mittwoch, 20. November 2013, 18:12:35 schrieb Pascal Obry:
>> Le 20/11/2013 11:59, Tobias Ellinghaus a écrit :
>>> Just "darktable" should never be capitalized.
>
.name)
print("fmt is"..fmt)
end
dt.register_storage("module_print","view with
profile",store_callback,nil,supported_callback)
Any idea?
And if the callback returns true or false does not matters, all the
format are listed in the export dialog.
--
Pascal Obry / Mag
Jeremy,
> ok, normaly i fixed it. please test..
Not completely, it is a bit better as now *all* the format are always
listed even if the callback returns false.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
h
Jeremy,
> huh ?
>
> ok, i'll work on it ASAP.
Thanks a lot! I'm pleased to do the testing :)
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net
Jeremy,
> should work now...
>
> could you test again ?
Working fine now, thanks a lot!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key
g else.
I'm not using openSUSE so I'm not sure to be of any help but without
posting the actual error reported by the compiler it will be difficult
to know what's going wrong.
Pascal.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagi
le"
function
store_callback(storage,image,format,filename,num,total,high_quality,extra_data)
end
function supported_callback(storage,format)
return true
end
dt.register_storage("module_print","lua-export",store_callback,nil,supported_callback)
>>>
--
Pascal
Le 11/12/2013 15:48, jeremy rosen a écrit :
> I havn't really looked at that kind of use-cases, i'll look into it ASAP
Thanks a lot Jeremy!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.ob
currently not integrated into presets/styles.
* #402 fixes #9431 which is a very bad bug (removing wrong image)
this pull-request should be tested later today by the reporter
on my side I have tested it successfully
That's all for now.
--
Pascal Obry / Magny Les Hameaux (78)
The bes
2013/12/16 Halgeir Kjønås Rennehvammen :
> Blending is a bit like black magic for me. Why not write a good blog
> article about how blending works.
There is some video around. don't have links here...
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is
be to remove the history only when clicking on the
[compress history stack] button.
What do you think?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
sider
https://github.com/darktable-org/darktable/pull/406 for inclusion into
1.4?
This patch seems quite safe on the dt side :)
What do you think?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
2013/12/17 Pascal Obry :
> Looking at the lens.c evolution it seems that this loose search has
> never been activated.
> Don't know why, but it seems to be working just fine in my case. We
> had many report about
> lens mis-match on the mailing-list, is there good reasons to
2013/12/17 jeremy rosen :
> pascal, please keep it post-1.4
Sure. Marked as such on GitHub.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949B
Ok thanks for the feedback I have closed this pull-request!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
Le 23/12/2013 11:48, Moritz Moeller a écrit :
> See subject. 1.4rc1, OS X 10.9.1.
Indeed, and looks like there is something more broken even as the style
is not updated when changing preset.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
h
Le 23/12/2013 11:48, Moritz Moeller a écrit :
> See subject. 1.4rc1, OS X 10.9.1.
Looks like my issue is only for "none" which is saved as an empty string
into the settings and then not properly restored.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel i
Le 23/12/2013 11:48, Moritz Moeller a écrit :
> See subject. 1.4rc1, OS X 10.9.1.
I have fixed the issue with export presets without a style but the reset
still done nothing.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.
uot;does
nothing" visible.
I'm no expert in this area and I'd like the view of some other
developers before doing anything (if there is something to do).
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
htt
low definition (see
corresponding preference) of the image. So nothing wrong to me.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net
or all formats
- for LDR only
- for RAW and HDR images
Something better? Not sure LDR is making sense to our users...
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver
This is now fixed.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
s to be done to recover.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
s what I would try.
PLEASE BACKUP YOUR .config/darktable/library.db
First check on darktable the image id as displayed in the lightable.
Then:
echo DELETE FROM masks WHERE imgid= | sqlite3
.config/darktable/library.db
Hope this helps,
--
Pascal Obry / Magny Les Hameaux (78)
The best
Markus,
> so there are no cross-references from somewhere else to those entries in
> mask? I think about zombie-module-instances referring invalid shapes.
No, I don't think so.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imaginat
Le 11/01/2014 12:24, Roumano a écrit :
> Can you reproduce it ?
Indeed. I can reproduce. This happens only when applying a style from
the lighttable and not in the darkroom.
I'll have a look.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of ima
Hi Roumano,
> On my PC, now (tested with the last git version) darktable ecrase my
> already existing history stack when i apply a style.
Should be fixed now.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Tobias,
The French translation is not up-to-date on the darktable 1.4.x
branch. Do you want me to push the changes there or will you
cherry-pick them from master?
- --
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means
Le samedi 25 janvier 2014 à 18:38 +0100, Richard Levitte a écrit :
> It seems to have a bug, though... I can't change the job name
> "noname" to what I want, I get MMDD_noname (where MMDD is the
> current date) whatever I do...
This has been fixed some hours
Count me in, size M.
Thanks!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
somehow?
A script in "tools/clean-tagxtag.sh"
What do you think?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-ke
likely to be assigned. If no one uses and/or likes that feature we could as
> well drop it and get rid of all that code.
I do use quite a lot the tagging module and found this feature really
nice (I come
from Lightroom which has the same feature BTW). So I'm really against removing
this fe
e run only once after starting
darktable version 1.6.
UPDATE tagxtag
SET count=(SELECT count(*)
FROM tagged_images
WHERE tagid=id2
AND imgid IN (SELECT imgid FROM tagged_images WHERE tagid=id1));
UPDATE tagxtag SET count = count + 100 WHER
I have refrained from translating this word in French. Possible
translations would have been "annulation du scintillement" ou
"atténuation du scintillement". Those terms are lot too long for the
IOP. If someone has a better proposal let me know!
Pascal.
--
Pascal Obry
or the
2 controls. Note sure I'm clear, look at the actual GUI :)
Pascal.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv
cintillement" in brackets.
>
> Example:
> http://www.lesnumeriques.com/moniteur-ecran-lcd/benq-xl2720z-p17980/test.html
> (search for "flicker").
Right.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr
ld it be possible to insert export job with higher priority to avoid
them being removed?
Anyway I'm not familiar enough with this code to find a way to fix.
Any idea?
Thanks,
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
4.
Fixed, thanks for the heads-up.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
-
e to IGV give me the very
same image, at least close enough to not see any change on screen.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserve
tr=0x6d5750)
at /home/obry/dev/builds/darktable/src/src/control/control.c:816
#7 0x74283062 in start_thread (arg=0x7fffe2211700)
at pthread_create.c:312
#8 0x70d92a3d in clone ()
at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
(gdb)
--
Pascal Obry / Magny Les H
Le dimanche 16 mars 2014 à 11:48 +0100, Pascal Obry a écrit :
> I receive the following message every time I export to Picasa. Any idea?
Issue found and fixed.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
h
ange the database layout
> manually. Since we did quite a few changes in the meantime that would be a
> little risky though.
Or continue using a Git version of dt (a version before Feb 25 should
do).
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
Hello,
Given the recent updates to the basecure tool I'm wondering if
regenerating the basecurve will give better result? Does the
computation is actually more accurate?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.e
Hi Edouard,
> In all cases, i'm still looking forward feedback. The improvements
> should help get more accurate when using multiple images to determine
> the curves.
Ok, I'll try the new tool on my D800 samples then and I'll get back to you.
Thanks,
--
Pascal Obry
ing it gradually (especially recently using
> Robert Hutton's excellent videos).
>
>
> I really need some help and would be much appreciative.
Activating OpenCL if supported by your systems may help a lot I suppose.
Have you tried? Did you looked at the manual on how to do thi
e waiting for a printing module and I'd
like to see this happen in 2014 :)
Waiting for ideas... Contributors... Feedback...
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.
here is not an issue with this specific version
of Nvidia driver.
Any feedback?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
PACKAGE_VERSION}/build/uvm"
> CLEAN+="; make -C uvm clean"
# dkms remove nvidia-current/331.67 -k $(uname -r)
# dkms build nvidia-current/331.67
Then reboot, or:
# insmod /var/lib/dkms/nvidia-current/331.67/$(uname
-r)/x86_64/module/nvidia-current-uvm.ko
I'm now enjoying th
2014-05-02 10:29 GMT+02:00 Brainslug :
> Are you going to file a Debian bug report for this? It seems that the
> newer version (experimental, 334.21) is still missing the uvm module and
> therefore does not support OpenCL support.
Done: #746643
--
Pascal Obry / Magny Les Hameaux (7
developers.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
--
"Accel
I have just proposed a pull-request to fix this issue:
https://github.com/darktable-org/darktable/pull/568
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key
ntrol/jobs/control_jobs.c:1291
1291 mstorage->export_dispatched(mstorage);
(gdb) print mstorage->export_dispatched
$1 = (void (*)(struct dt_imageio_module_storage_t *)) 0x0
So, the storage is null.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imag
Jeremy,
> Ok, I just commited a fix. can you test that it works for you ?
Yes, working fine on my side too. Thanks!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.
Issue resolved for Nvidia 331.79-1 on Debian unstable.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
So nothing to add there!
Cheers,
Pascal.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --
enerate the final
> TIFF (correct size, correct color transformations, ...) and pass that to
> cupsPrintFile() or one of the similar functions [0].
Will look, thanks for the pointer!
Cheers,
Pascal.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagi
deliver best possible print quality for the given hardware we should
> allow the use of turboprint.
Difficult to answer as I don't know turboprint. Where does turboprint
come into action? How is it configured and setup? Does it works with
CUPS or something else?
--
Pascal Obry / M
Some news...
I have been able to create a page out of an image with Gutenprint
(outside of dt). At this point I think I have most of the Gutenprint
knowledge to create the printing module.
So this is moving, slowly but moving...
--
Pascal Obry / Magny Les Hameaux (78)
The best way to
ult
and maybe clipped gamut. All this won't fit a simple dialog.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver ke
with experience with this?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F94
for dt. I'm still not sure
how to apply a printer ICC profile in-memory while
creating the cairo surface, but I suppose this is possible.
How does that sounds?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
Magick.
Probably so, and I just discovered that Gtk has support for this:
gtk_print_job_get_surface ()
This routine returns a cairo surface to be printed.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.ob
; ==
> == Internal ==
> ==
>
> The internal code will be mostly on those files:
>
> src/views/print.c - the print module itself
>
> src/libs/print_settings.c - the print dialog as seen on the right
>
> An important point is to rememb
of the cairo surface to
be printed and we can do that without any color managed support from
cairo (again as done in the develop module).
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg
hange
of profile before. This is the recommended way to print on many systems.
At least when I was working on Windows the proper solution was to let
the application (Lightroom at this time) do the work and be sure to
*disable* any color management in the printer driver.
--
Pascal Obry / Magny Les H
ering
>- AdobeRGB for the intermediate rendering
>
>
> ==
> == Internal ==
> ==
>
> The internal code will be mostly on those files:
>
> src/views/print.c - the print module itself
>
> src/libs/print_settings.c - the print dialog as seen on
After discussion on IRC and with Pascal de Bruijn here are updated notes.
The important changes are in the exported/intermediate format where we need
to be able to control the intent/profile.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http
dt will evolve, but I really doubt
something like this will be supported.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.ne
Hi there!
Another iteration on this print module design.
After IRC discussions we are really leaning toward using the Gtk+ print
API.
I have added a section about Color Management and the current questions
which remains to answer.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to
added an alignment widget (same as in watermark module) to
specify the alignment of the picture inside the print area.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net
that the
> applied style is simply added to the existing.
> I then need to compress the history in order to clean-up the true sequence.
>
> Is this as intended?
Yes, you have an option in the lightable to select whether a style
replaces or is added to the existing history.
--
Pa
s.
Thanks!
-
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
--
Want fast and e
> What is the question?
Indeed not a question more a remark that having the tonecurve replacing
the basecurve is misleading as you need to *manually* deactivate the
basecurve which is well the *base* curve applied to all RAW
automatically by darktable. This usage seems broken to me.
--
P
hard to see a difference when looking at the
corresponding jpeg.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
of the pipe, that will deprecate basecurve
> iop.
Ok, anyway for now I have output the basecurve from dt-curve-tool and
will be using this.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --key
on my side both
> improvements, but I was still unsatisfied of color saturation with the
> single tonecurve applied. Colors were lacking some pop.
Using the basecurve I have a bit too much saturation compared to the
D800 camera jpeg. So I cannot confirm your experiment.
--
Pascal Obry
Le lundi 04 août 2014 à 18:51 -0400, Patrick Shanahan a écrit :
> How to correct, bug report, ???
Yes, this will be the best way to record the exchange and possible
solution if others have the same problem.
Thanks,
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is
> What happens here?
Not sure, maybe not enough memory?. This could explain why using
multiple threads is not working. How much RAM do you have allocated to
the virtual machine?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.
t;PRId64"'"),
(gint64)((1024 * 1024 * 512) * factor));
It seems that the " just after % confused gettext!
Any idea?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.ne
; that gets translated. We need to move it outside _(), together
> with % symbol, I guess to a second snprintf.
Thanks for your help. Now fixed!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg -
f 4*sizeof(float) */
-} __attribute__((packed));
+} __attribute__((packed,aligned(16)));
// last resort mem alloc for dead images. sizeof(dt_mipmap_buffer_dsc) +
dead image pixels (8x8)
// __m128 type for sse alignment.
Is that it?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to
nice is to be able to tell dt to rebuild the cache
for all images. I could launch this task and go to sleep :)
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key
Thinking about this maybe Lua can help here.
Jeremy, do you think there is way in Lua to iterate over all images in
the database and ask for rebuilding the preview in the cache?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http
_MIPMAP_0, DT_MIPMAP_BEST_EFFORT);
if(buf.buf)
dt_mipmap_cache_read_release(darktable.mipmap_cache, &buf);
So that's 2 functions, a struct and some constants (DT_MIPMAP_*).
How does that sounds?
Have a good day!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to tra
it machines, so let's try to
> make compiler do the padding.
>
> commit ffbec14de21b0d3cf0b77e040308134c49632c22
> Author: Roman Lebedev
> Date: Fri Sep 5 23:04:22 2014 +0400
>
> Use size_t to store buffer sizes. With this (and probably following), DT
> can o
/mipmaps-62d87cbcad56aa2b4a21cd680776714d70e93ef6'
failed!
Starting again dt most of my mipmap were gone and needed a rebuild!
So there is definitely still something wrong with the mipmap cache!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagin
As a follow-up to my previous message.
2014-09-16 8:38 GMT+02:00 Pascal Obry :
> Starting again dt most of my mipmap were gone and needed a rebuild!
But after that it seems that the mipmap cache is ok.
That means that there is an incompatibility between the old mipmap and
the new one. Is t
After a full day to rebuild my mipmap cache (40% was done):
[mipmap_cache] serialization to
`/home/obry/.cache/darktable/mipmaps-62d87cbcad56aa2b4a21cd680776714d70e93ef6'
failed!
Shouldn't we consider reverting the recent patches on this code?
--
Pascal Obry / Magny Les H
oks like a buffer overflow or a alignment issue...) and
then next time you start dt it rebuild the cache.
I'm not sure how to help to reproduce!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
. start dt
5. start dt again
As it is in step 4 (while exiting) that the cache is corrupted.
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
---
I have reverted the recent mipmap changes and will report. I'll be
testing on my two machines.
Things is that I have read and read again this code and cannot see
something wrong! I'm also using GCC 4.9.1 from Debian/sid BTW.
--
Pascal Obry / Magny Les Hameaux (78)
The best way
an impact on the mipmap level 0).
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
--
yesterday (I have
copy of my latest working mipmap cache, this speed up the testing).
So definitely one of those commits trigger the issue!
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg
send the spec of my other machine later.
My proposal is to revert the four listed commits in my previous message.
Rethink them and propose a new version. I'm ready to test them as I have
cache invalidation on two different computers. Easy to test for me.
How does that sound?
--
Pascal Obry
2014-09-18 6:53 GMT+02:00 Pascal Obry :
> My machine: Debian/sid / Intel core i7-4800MQ (8 cores) / 16Gb RAM.
>
> I'll send the spec of my other machine later.
My second machine is an Intel i5-2520 (4 cores) / 8Gb RAM.
--
Pascal Obry / Magny Les Hameaux (78)
The best way t
2014-09-18 9:16 GMT+02:00 Pascal Obry :
> 2014-09-18 6:53 GMT+02:00 Pascal Obry :
>> My machine: Debian/sid / Intel core i7-4800MQ (8 cores) / 16Gb RAM.
>>
>> I'll send the spec of my other machine later.
>
> My second machine is an Intel i5-2520 (4 cores) / 8Gb
float here is puzzling...
Any cache expert to shed some light?
--
Pascal Obry / Magny Les Hameaux (78)
The best way to travel is by means of imagination
http://v2p.fr.eu.org
http://www.obry.net
gpg --keyserver keys.gnupg.net --recv-key F949BD3B
1 - 100 of 633 matches
Mail list logo
