Awesome Privacy

Recent Commits

• Pascal Obry (04 Oct 24)

  Merge pull request #17597 from victoryforce/translations-241004 Translations update: en@truecase and Ukrainian

• Victor Forsiuk (04 Oct 24)

  Ukrainian translation update

• Victor Forsiuk (04 Oct 24)

  en@truecase update

• Pascal Obry (04 Oct 24)

  Update French translation.

• Pascal Obry (04 Oct 24)

  Update darktable.pot for translators.

• ralfbrown (01 Oct 24)

  parallelize colorspace conv, add NULL checks

• Pascal Obry (04 Oct 24)

  Update integration test sub-module.

• Pascal Obry (04 Oct 24)

  Merge branch 'gaussian_improvements' of https://github.com/jenshannoschwalm/darktable

• ralfbrown (03 Oct 24)

  add IMAGE.TAGS.HIERARCHY substitution variable

• Mario Zimmermann (04 Oct 24)

  update export module gui for style from preset (#17590) * update export module gui for style from preset The chosen style and style mode are saved in the preset but the gui is not updated accordingly. * save style name in dt config * remove duplicate g_free

• Hanno Schwalm (04 Oct 24)

  Implement dt_gaussian_fast_blur_cl_buffer() and dt_gaussian_fast_blur() and maintenance Although the standard gaussian blur code for cl buffers or CPU is pretty fast we can do even better for small sigmas. Using a simple NxN kernel is faster because we don't have to fiddle around with sorting data and especially on cards with lots of processing units and fast cl_mem or on CPU cached memory the algorithm is clearly faster due to cache locality. Both (OpenCL and CPU) fast gaussian variants support 1, 2 or 4 channels and run on a 9x9 coeff matrix suitable for a sigma up to ~1.5 (acceptable up to 2.0), performance if roughly doubled compared with the standard gaussian blurs in cpu memory or cl buffers. If sigma is larger than the acceptable threshold we fallback to standard algorithm. For details threshold, dual demosaicing, mask blurring and segmentation gradients we used special gaussian blurs, now the public dt_gaussian_fast_blur_cl_buffer() and dt_gaussian_fast_blur() functions are used. As both functions now do proper calculations at the borders the mentioned algos have less artifacts at the outermost 4 pixels. Maintenance while being here: make use of dt_opencl_enqueue_kernel_1d_args() in gaussian OpenCL code.

• Pascal Obry (03 Oct 24)

  Merge pull request #17582 from ralfbrown/view_alloc_checks [maint] check memory allocs in views

• Pascal Obry (03 Oct 24)

  Merge pull request #17580 from ralfbrown/lib_alloc_checks [maint] check memory allocs in utility modules

• Tomasz Golinski (03 Oct 24)

  Fix compilation without OpenEXR Commit e38d2c8db4f457a0e45624d646244446352de609 "add more magic bytes, use only appropriate loader when magic matches" introduced surrogate loader functions for file types whose libraries are not available. It was missing a surrogate for OpenEXR resulting in compilation fail if the library was not enabled. This commit adds the missing surrogate.

• Pascal Obry (03 Oct 24)

  Minor code reformatting.

• Pascal Obry (03 Oct 24)

  Merge pull request #17568 from darktable-org/po/filmstrip-sel change way to interact with filmstrip

• Pascal Obry (03 Oct 24)

  channelmixrrgb: Clear chroma adaptation pointer on clean-up. Continued work for #17368

• Pascal Obry (02 Oct 24)

  Simplify select single call (no need for a list). And make sure the single selection occurs after the edited image to fix the range selection.

• Pascal Obry (01 Oct 24)

  Do not use hard-coded doubl-click-time, request it from settings.

• Pascal Obry (30 Sept 24)

  Change filmstrip UX to match what lighttable does. So for the selection: - Click : select a single image - Ctrl+Click : toggle image selection - Shift+Click : select range And so now to change image in darkroom: - Double-click : change edited image Note that the tricky part is that a double-click does raise: GDK_BUTTON_PRESS GDK_BUTTON_PRESS GDK_2BUTTON_PRESS So we do resort on a timeout action for single click (GDK_BUTTON_PRESS) selection which gets remove if a double click is actually done. Part of #16275.

• Pascal Obry (30 Sept 24)

  Use dt_is_valid_imgid(id).

• Pascal Obry (30 Sept 24)

  Minor code reformatting.

• Pascal Obry (02 Oct 24)

  Revert part of 0a6cc9404802b2fb3b75583160441778ace60889. This commit has introduced a regression where reseting the histroy stack would break the actual chroma and a reset of the color calibration is then needed to get back the proper colors. This can be reproduced by: - Reset histroy on lighttable - Enter darkroom - Zoom on the picture and see the colors/chroma changing

• Pascal Obry (02 Oct 24)

  RELEASE_NOTES.md: undo on filmstrip.

• Pascal Obry (30 Sept 24)

  darkroom: Allow undo/redo support for copy/paste history in filmstrip.

• ralfbrown (02 Oct 24)

  check memory allocations for failure in view.c

• ralfbrown (02 Oct 24)

  check memory allocations for failure in tethering view

• ralfbrown (02 Oct 24)

  check memory allocations for failure in map view

• ralfbrown (02 Oct 24)

  check memory allocations for failure in tagging module

• ralfbrown (02 Oct 24)

  check memory allocations for failure in recent collections module