Blog of Vasiliy Kirilichev

Before the next F-Spot release

2008-09-08T23:52:00.000+05:00

After official end of gsoc I have done one feature more — xicc support — you can set "Try to use the system display profile" and f-spot will try to use it, If xicc is not null. And it is not all :)

F-Spot: color profile support, final week

2008-08-11T21:20:00.000+05:00

Hi all!
I didn't write here for a long time, sorry. But I have good news, the latest F-Spot+color management is full functionality and it work faster than previous versions. Now I'm writing some documentation and clean up the code.

F-Spot: color profile support, progress 7

2008-07-13T18:37:00.000+05:00

Hi! It was a nice coding week, I can say that alpha version is done. So all big bugs with color profiles were fixed:

effects: all, but without soft focus now

rotated jpg pictures

and I don't apply the color profile to picture with transparent color as the lcms work incorrect with it.

And following was done:

save the current color profile settings in gconf

siderbar icons

So next week I plan to debug some part of my code and to correct the several insignificant bugs and.. I'll try to find new bugs :)

F-Spot: color profile support, progress 6

2008-07-08T01:54:00.000+05:00

So the following must be done for color profile support: (bugs)

one active icon in Filmstrip.cs has always color strips (when I use the color profile)
the color profile work incorrect with any picture with transparent color

save/load color profile settings on hdd

anything to make the apply color profile faster?

Here is the short screencast which can try to show how slowly is it.

F-Spot: color profile support, progress 5 (midpoint)

2008-07-07T01:19:00.000+05:00

So it's time to summarize what I have already done!

Following steps are done:
(apply the color profile to)
1) main picture
2) icons: main icons, small icons below it and icons from FilmStrip.cs, small icon below the main image in the "edit image" mode
3) fullscreen view, slide view

draft:
* printer profile

and it must be done:
1) save the current profiles in file and load it after program/system reboot.
2) icon in sidebar (troubles: when the icon is set automatically -- it take the image without color profile but
when the user set the icon (with change tag icon menu) it take the image WITH color profile, it must be fixed :)
3) fix bugs and optimize the code

(now a big bug -- when I apply profile to transparent color on image)

---

modified file list:
1) AsyncPixbufLoader.cs
2) ColorDialog.cs
3) PreferenceDialog.cs
4) F-Spot.glage
5) Photoview.cs
6) MainWindow.cs
7) IconView.cs
8) Global.cs (
9) PixbufUtils.cs
10) PrintOperation.cs (draft variant, only to understand is it right pixbuf copy)
11) TagView.cs
12) FilmStrip.cs
13) TextureDisplay.cs
14) TagSelectionWidget.cs
15) Makefile.am (drag the TagView.cs to other file group -- before it was a namespace error. any other way to fix it?)
16)* ColorManagement.cs (I plan to use this for all color profile features)

Progress 4

2008-06-30T17:50:00.000+05:00

Salut! This week there were so much things! The football championship (euro'08) has ended.. We have lost, but it was interesting to see it on tv, my congratulations to Spaniards, great game!
Now work: I am engaged the color profile to be applied correctly to icons.
There are allready basic ideas of realization, but there are some complexities with its realization, and now I'm doing it:) And one more: I have looked how (and where) I can apply the printer profile. The draft variant is already exist and it is work. But it is very complex to see how it is — right or not right — you can't see the histogram here! (or you can, but it is complex:) also I structured all changes made by myself,
have made the list of that it is necessary to make first of all and some optimized the code written earlier.

Progress 3

2008-06-23T17:03:00.000+05:00

Last 2 weeks I was engaged in code-optimization of a color profile in the picture loader,
now the profile is applied "on the fly", and also has made more logical in my part of application.

I have continued to apply the color profile to preview-icons, but it still demands some work.
This and the conservation of adjustments are one of the last that it is necessary to make up with
the monitor profiles to the beta-version. I'm going to start to be engaged in support
of the printer-profiles this week.

Progress 2

2008-06-08T19:40:00.000+05:00

So the work is moving and on this week the following has been made by me:
1) There is no designated place to store device profiles on Linux, /usr/share/color/icc/ has become a de facto standard, used by several applications, by F-Spot now too :) and this path: "/usr/local/share/color/icc " too. (FileChooserDialog - not now:) And I think that it is not a final version to store the profiles, I'd like to read more about ICC directories in Linux.
2) The profile is set to X screen.
3) The color profile support was added into the loader code.

But there are some errors, the code must be improve.

Progress 1

2008-06-02T09:37:00.000+05:00

Last week I have unhide the color management frame in f-spot preferences (it must be edit f-spot.glade by hand to unhide this) and read some about how can I do it (about glade).

Now I'm editing the PreferenceDialog.cs, I want it to look like this:

At the same time I'm continue studing the general structure of sources, and methods to use the color profiles.

Compiling and acquaintance with sources

2008-05-18T14:22:00.000+05:00

Hi, so I think that it is good to describe some questions that I recieve in the F-Spot-introduce time.

First — it was my biggest surprise — the MonoDevelop Projectfile is out of date :) It is really unusual for me, and today I explore the Project with MonoDev and compiles with "make all" in the console.

But it is also not so easy to compile the Project in the console: if I do everything according to "How to build from head" in my Ubuntu 8.04 I become Makefile not right. Report after launch "make all":

...
Making all in libeog
make[2]: Entering directory `/home/wasja/development/f-spot/trunk_new/libeog'
( --prefix=libeog_marshal ./libeog-marshal.list --header > libeog-marshal.tmp \
&& mv libeog-marshal.tmp libeog-marshal.h ) \
|| ( rm -f libeog-marshal.tmp && exit 1 )
/bin/bash: --prefix=libeog_marshal: command not found
make[2]: *** [libeog-marshal.h] Error 1
make[2]: Leaving directory `/home/wasja/development/f-spot/trunk_new/libeog'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/wasja/development/f-spot/trunk_new'
make: *** [all] Error 2

For the decision of this problem it is necessary to write in the .../libeog/Makefile:

GLIB_GENMARSHAL = *`pkg-config --variable=glib_genmarshal glib-2.0`*

and in the end of file it:

libeog-marshal.h: libeog-marshal.list
(*$(GLIB_GENMARSHAL) * --prefix=libeog_marshal $(srcdir)/libeog-marshal.list --header > libeog-marshal.tmp \
&& mv libeog-marshal.tmp libeog-marshal.h ) \
|| ( rm -f libeog-marshal.tmp && exit 1 )

libeog-marshal.c: libeog-marshal.h
(*$(GLIB_GENMARSHAL) * --prefix=libeog_marshal $(srcdir)/libeog-marshal.list --body > libeog-marshal.tmp \
&& mv libeog-marshal.tmp libeog-marshal.c ) \
|| ( rm -f libeog-marshal.tmp && exit 1 )
# Tell versions [3.59,3.63) of GNU make to not export

in .../libfspot/Makefile too.

thanks to sde — a more preferable alternative — to analyse the autogen log. (ldconfig helped me :)

Now it is work properly. I study the Sources, especially LCMS-Wrapper and how one can work with them.

Hello, World!

2008-05-09T19:26:00.000+05:00

So, I've been accepted onto the Google Summer of Code program and I'm going to be working on F-Spot and adding the functionality of being able to use a color profiles. You can read my proposal over here. My mentor is Lawrence Ewing and it is very cool :)