enlightenment-devel — Enlightenment development discussion

Re: [E-devel] Fwd: Followup to Bug report for imlib2-1.12.5

[Kim W. <ki...@wo...>]

On 2025-12-02 23:47, Brian Lindholm wrote:
> As a follow-up to my previous bug report, I'm also sending in a minor 
> enhancement to the _DrawGradient() function in src/lib/grad.c.  The 
> logic in today's code results in significant portions of the screen 
> being filled in with a uniform color as opposed to a gradient when an 
> angle is used.  The attached patch "stretches" the gradient from corner 
> to corner, providing a gradient across the entire screen.  [And like my 
> previous patch, a some casts to (unsigned long int) were required to 
> avoid integer overflow issues.]
> 
Hello Brian,

I have pushed a patch, basically doing as suggested, but dealing with 
the i variable range limiting a bit differently (+ uint64_t for 32 bit 
arches).

Thanks :)

/Kim

Re: [E-devel] Fwd: Bug report for imlib2-1.12.5

[Kim W. <ki...@wo...>]

On 2025-12-02 23:46, Brian Lindholm wrote:
> I've stumbled across an integer overflow bug in imlib2-1.12.5, which I 
> discovered while experimenting with the "hsetroot" utility (from 
> https://github.com/himdel/hsetroot; it utilizes imlib2 internally).
> 
> The problem can be demonstrated with the following command line:
> 
> hsetroot -add black -add white -add black -add white -add black -add \
>    white -add black -add white -add black -add white -add black -add \
>    white -gradient 45
> 
> With straight imlib2-1.12.5, I get a bar on the left side of the screen 
> that is not filled correctly (per the attached orig.png file).  It only 
> happens on my QHD monitor (2560x1440 pixels), and I'm pretty sure it's 
> an integer overload issue.  If I strategically cast (int) to (unsigned 
> long int) in a few places so that 64-bit ints are used, the problem goes 
> away.
> 
Hello Brian,

I have pushed a fix, basically as suggested, but using uint64_t instead 
of unsigned long. This should fix things on 32 bit arches too.

I'll be back on the other issue when time permits.

Thanks :)

/Kim

[E-devel] Fwd: Followup to Bug report for imlib2-1.12.5

[Brian L. <bri...@gm...>]

As a follow-up to my previous bug report, I'm also sending in a minor 
enhancement to the _DrawGradient() function in src/lib/grad.c.  The 
logic in today's code results in significant portions of the screen 
being filled in with a uniform color as opposed to a gradient when an 
angle is used.  The attached patch "stretches" the gradient from corner 
to corner, providing a gradient across the entire screen.  [And like my 
previous patch, a some casts to (unsigned long int) were required to 
avoid integer overflow issues.]

Brian Lindholm
bri...@gm...

[E-devel] Fwd: Bug report for imlib2-1.12.5

[Brian L. <bri...@gm...>]

I've stumbled across an integer overflow bug in imlib2-1.12.5, which I 
discovered while experimenting with the "hsetroot" utility (from 
https://github.com/himdel/hsetroot; it utilizes imlib2 internally).

The problem can be demonstrated with the following command line:

hsetroot -add black -add white -add black -add white -add black -add \
   white -add black -add white -add black -add white -add black -add \
   white -gradient 45

With straight imlib2-1.12.5, I get a bar on the left side of the screen 
that is not filled correctly (per the attached orig.png file).  It only 
happens on my QHD monitor (2560x1440 pixels), and I'm pretty sure it's 
an integer overload issue.  If I strategically cast (int) to (unsigned 
long int) in a few places so that 64-bit ints are used, the problem goes 
away.

Brian Lindholm
bri...@gm...

Re: [E-devel] Enlightenment 0.27.1 ROCKS!

[Pierre C. <pi...@co...>]

Oh yes ! Fine enlightenment !

On 7/29/25 20:28, jos...@me... wrote:
> Niltze [Hello]-
>
> Maybe forgot to let you know but since Enlightenment 0.27.1 under 
> Debian Bookworm has been very stable. Now, I am upgrading one instance 
> to upcoming Debian Trixie and still my E 0.27.1 build carried over 
> from Bookworm continues to be very stable. Screenshot shows only a few 
> minutes of uptime,
> < http://www.enlightenment.org/ss/e-68890dac94dbb1.68570887.jpg >
>
>  but I had already been using Trixie to test my reiser4 5.17.14-1 
> kernel GCC14 build in this wireless unit:
> 'Installation-reports: Successful Metztli Reiser4 / Debian Bookworm 
> netboot install on AceMagic AX16'
> < 
> https://metztli.blog/amatl/amatl-main/install-reports/acemagic-ax16-ryzen-7 
> >
>  where I had 'a priori' hacked into a kernel module the rtw89 
> backported driver as unit has no ethernet port but WI-FI 6 feature.
>
> Thank you for the wonderful and stable Enlightenment Window Manager!
>

[E-devel] Enlightenment 0.27.1 ROCKS!

[<jos...@me...>]

Niltze [Hello]-

Maybe forgot to let you know but since Enlightenment 0.27.1 under Debian 
Bookworm has been very stable. Now, I am upgrading one instance to 
upcoming Debian Trixie and still my E 0.27.1 build carried over from 
Bookworm continues to be very stable. Screenshot shows only a few 
minutes of uptime,
< http://www.enlightenment.org/ss/e-68890dac94dbb1.68570887.jpg >

  but I had already been using Trixie to test my reiser4 5.17.14-1 kernel 
GCC14 build in this wireless unit:
'Installation-reports: Successful Metztli Reiser4 / Debian Bookworm 
netboot install on AceMagic AX16'
< 
https://metztli.blog/amatl/amatl-main/install-reports/acemagic-ax16-ryzen-7 
 >
  where I had 'a priori' hacked into a kernel module the rtw89 backported 
driver as unit has no ethernet port but WI-FI 6 feature.

Thank you for the wonderful and stable Enlightenment Window Manager!

-- 
Best Professional Regards.

--
Jose R R
http://metztli.it
---------------------------------------------------------------------------------------------
Download Metztli Reiser4: Debian Bookworm w/ Linux 5.17.13-2 AMD64
---------------------------------------------------------------------------------------------
feats ZSTD compression https://sf.net/projects/metztli-reiser4/
-------------------------------------------------------------------------------------------
Official current Reiser4 resources: https://reiser4.wiki.kernel.org/

Re: [E-devel] What are E developers thoughts on the Arcan display server?

[Carsten H. <ra...@ra...>]

On Wed, 16 Jul 2025 00:12:07 -0300 Vinícius dos Santos Oliveira
<vin...@gm...> said:

> What are E developers thoughts on the Arcan display server?

None - it's not relevant to E (short version). Other than it basically competes
in many ways or is orthogonal. 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

[E-devel] What are E developers thoughts on the Arcan display server?

[Vinícius d. S. O. <vin...@gm...>]

What are E developers thoughts on the Arcan display server?

-- 
Vinícius dos Santos Oliveira

Re: [E-devel] [EGIT] [efl] 01/01: Windows: fix buffer overflow in test_flip2()

[Carsten H. <ra...@ra...>]

On Fri, 2 May 2025 15:35:17 +0200 Vincent Torri <vin...@gm...> said:

> look at the 2ndusage of buf:
> 
> 
>    snprintf(buf, sizeof(buf),
>             "This is an entry widget in this window that<br/>"
>             "uses markup <b>like this</> for styling and<br/>"
>             "formatting <em>like this</>, as well as<br/>"
>             "<a href=X><link>links in the text</></a>, so enter text<br/>"
>             "in here to edit it. By the way, links are<br/>"
>             "called <a href=anc-02>Anchors</a> so you will need<br/>"
>             "to refer to them this way.<br/>"
>             "<br/>"
> 
>             "Also you can stick in items with (relsize + ascent): "
>             "<item relsize=16x16 vsize=ascent
> href=emoticon/evil-laugh></item>" " (full) "
>             "<item relsize=16x16 vsize=full
> href=emoticon/guilty-smile></item>" " (to the left)<br/>"
> 
>             "Also (size + ascent): "
>             "<item size=16x16 vsize=ascent href=emoticon/haha></item>"
>             " (full) "
>             "<item size=16x16 vsize=full href=emoticon/happy-panting></item>"
>             " (before this)<br/>"
> 
>             "And as well (absize + ascent): "
>             "<item absize=64x64 vsize=ascent
> href=emoticon/knowing-grin></item>" " (full) "
>             "<item absize=64x64 vsize=full
> href=emoticon/not-impressed></item>" " or even paths to image files on disk
> too like: " "<item absize=96x128 vsize=full
> href=file://%s/images/sky_01.jpg></item>"
>             " ... end."
>             , elm_app_data_dir_get()
>            );
> 
> 
> ok, not buffer overflow, but anyway, buf should be longer for this

oh i didnt look at the 2nd use... so yeah - then its ok :)

> Vincent
> 
> On Fri, May 2, 2025 at 10:08 AM Carsten Haitzler <ra...@ra...> wrote:
> >
> > On Thu, 01 May 2025 23:32:04 -0700 Enlightenment Git
> > <no-...@en...> said:
> >
> > actually... the code was correct. that doesn't buffer overflow at all - it
> > simply cuts off the path at 259 chars (plug 0 byte) and thus fails to load
> > the image. if the max path is truly 260 and the file is somehow in a place
> > in the filesystem where it cannot be addressed due to this limit... then
> > making the buffer bigger even though the max path is only 260 isn't going
> > to help.
> >
> > the alternative is that the max path is not 260 and is actually longer and
> > 260 is some hold-over from dos/fat days and thus we should be defining
> > PATH_MAX as something bigger and that is the solution. PATH_MAX is used all
> > over the place so these kinds of commits will then have to be done in 100's
> > of places.
> >
> > > This is an automated email from the git hooks/post-receive script.
> > >
> > > git pushed a commit to branch master
> > > in repository efl.
> > >
> > >
> > > View the commit online.commit 7629a786523d6fce3f1090bde0cb2fe7692835b5
> > > Author: Vincent Torri <vt...@ou...>
> > > AuthorDate: Fri May 2 08:30:52 2025 +0200
> > >
> > >     Windows: fix buffer overflow in test_flip2()
> > >
> > >     MAX_PATH is 260 on Windows, while buf is used for a long message
> > > ---
> > >  src/bin/elementary/test_flip.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/src/bin/elementary/test_flip.c
> > > b/src/bin/elementary/test_flip.c index 4e63ebae47..14f6cfe6df 100644
> > > --- a/src/bin/elementary/test_flip.c
> > > +++ b/src/bin/elementary/test_flip.c
> > > @@ -271,7 +271,7 @@ void
> > >  test_flip2(void *data EINA_UNUSED, Evas_Object *obj EINA_UNUSED, void
> > > *event_info EINA_UNUSED) {
> > >     Evas_Object *win, *bg, *bx, *bx2, *fl, *o, *bt, *tb, *li, *en;
> > > -   char buf[PATH_MAX];
> > > +   char buf[4096];
> > >
> > >     win = elm_win_add(NULL, "flip2", ELM_WIN_BASIC);
> > >     elm_win_title_set(win, "Flip 2");
> > >
> > >
> > > --
> > > To stop receiving notification emails like this one, please contact
> > > the administrator of this repository.
> >
> >
> > --
> > ------------- Codito, ergo sum - "I code, therefore I am" --------------
> > Carsten Haitzler - ra...@ra...
> >
> >
> >
> > _______________________________________________
> > enlightenment-devel mailing list
> > enl...@li...
> > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> 
> 
> _______________________________________________
> enlightenment-devel mailing list
> enl...@li...
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

Re: [E-devel] [EGIT] [efl] 01/01: Windows: fix buffer overflow in test_flip2()

[Vincent T. <vin...@gm...>]

look at the 2ndusage of buf:


   snprintf(buf, sizeof(buf),
            "This is an entry widget in this window that<br/>"
            "uses markup <b>like this</> for styling and<br/>"
            "formatting <em>like this</>, as well as<br/>"
            "<a href=X><link>links in the text</></a>, so enter text<br/>"
            "in here to edit it. By the way, links are<br/>"
            "called <a href=anc-02>Anchors</a> so you will need<br/>"
            "to refer to them this way.<br/>"
            "<br/>"

            "Also you can stick in items with (relsize + ascent): "
            "<item relsize=16x16 vsize=ascent href=emoticon/evil-laugh></item>"
            " (full) "
            "<item relsize=16x16 vsize=full href=emoticon/guilty-smile></item>"
            " (to the left)<br/>"

            "Also (size + ascent): "
            "<item size=16x16 vsize=ascent href=emoticon/haha></item>"
            " (full) "
            "<item size=16x16 vsize=full href=emoticon/happy-panting></item>"
            " (before this)<br/>"

            "And as well (absize + ascent): "
            "<item absize=64x64 vsize=ascent href=emoticon/knowing-grin></item>"
            " (full) "
            "<item absize=64x64 vsize=full href=emoticon/not-impressed></item>"
            " or even paths to image files on disk too like: "
            "<item absize=96x128 vsize=full
href=file://%s/images/sky_01.jpg></item>"
            " ... end."
            , elm_app_data_dir_get()
           );


ok, not buffer overflow, but anyway, buf should be longer for this

Vincent

On Fri, May 2, 2025 at 10:08 AM Carsten Haitzler <ra...@ra...> wrote:
>
> On Thu, 01 May 2025 23:32:04 -0700 Enlightenment Git
> <no-...@en...> said:
>
> actually... the code was correct. that doesn't buffer overflow at all - it
> simply cuts off the path at 259 chars (plug 0 byte) and thus fails to load the
> image. if the max path is truly 260 and the file is somehow in a place in the
> filesystem where it cannot be addressed due to this limit... then making the
> buffer bigger even though the max path is only 260 isn't going to help.
>
> the alternative is that the max path is not 260 and is actually longer and 260
> is some hold-over from dos/fat days and thus we should be defining PATH_MAX as
> something bigger and that is the solution. PATH_MAX is used all over the place
> so these kinds of commits will then have to be done in 100's of places.
>
> > This is an automated email from the git hooks/post-receive script.
> >
> > git pushed a commit to branch master
> > in repository efl.
> >
> >
> > View the commit online.commit 7629a786523d6fce3f1090bde0cb2fe7692835b5
> > Author: Vincent Torri <vt...@ou...>
> > AuthorDate: Fri May 2 08:30:52 2025 +0200
> >
> >     Windows: fix buffer overflow in test_flip2()
> >
> >     MAX_PATH is 260 on Windows, while buf is used for a long message
> > ---
> >  src/bin/elementary/test_flip.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/src/bin/elementary/test_flip.c b/src/bin/elementary/test_flip.c
> > index 4e63ebae47..14f6cfe6df 100644
> > --- a/src/bin/elementary/test_flip.c
> > +++ b/src/bin/elementary/test_flip.c
> > @@ -271,7 +271,7 @@ void
> >  test_flip2(void *data EINA_UNUSED, Evas_Object *obj EINA_UNUSED, void
> > *event_info EINA_UNUSED) {
> >     Evas_Object *win, *bg, *bx, *bx2, *fl, *o, *bt, *tb, *li, *en;
> > -   char buf[PATH_MAX];
> > +   char buf[4096];
> >
> >     win = elm_win_add(NULL, "flip2", ELM_WIN_BASIC);
> >     elm_win_title_set(win, "Flip 2");
> >
> >
> > --
> > To stop receiving notification emails like this one, please contact
> > the administrator of this repository.
>
>
> --
> ------------- Codito, ergo sum - "I code, therefore I am" --------------
> Carsten Haitzler - ra...@ra...
>
>
>
> _______________________________________________
> enlightenment-devel mailing list
> enl...@li...
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Re: [E-devel] [EGIT] [efl] 01/01: Windows: fix buffer overflow in test_flip2()

[Carsten H. <ra...@ra...>]

On Thu, 01 May 2025 23:32:04 -0700 Enlightenment Git
<no-...@en...> said:

actually... the code was correct. that doesn't buffer overflow at all - it
simply cuts off the path at 259 chars (plug 0 byte) and thus fails to load the
image. if the max path is truly 260 and the file is somehow in a place in the
filesystem where it cannot be addressed due to this limit... then making the
buffer bigger even though the max path is only 260 isn't going to help.

the alternative is that the max path is not 260 and is actually longer and 260
is some hold-over from dos/fat days and thus we should be defining PATH_MAX as
something bigger and that is the solution. PATH_MAX is used all over the place
so these kinds of commits will then have to be done in 100's of places.

> This is an automated email from the git hooks/post-receive script.
> 
> git pushed a commit to branch master
> in repository efl.
> 
> 
> View the commit online.commit 7629a786523d6fce3f1090bde0cb2fe7692835b5
> Author: Vincent Torri <vt...@ou...>
> AuthorDate: Fri May 2 08:30:52 2025 +0200
> 
>     Windows: fix buffer overflow in test_flip2()
>     
>     MAX_PATH is 260 on Windows, while buf is used for a long message
> ---
>  src/bin/elementary/test_flip.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/src/bin/elementary/test_flip.c b/src/bin/elementary/test_flip.c
> index 4e63ebae47..14f6cfe6df 100644
> --- a/src/bin/elementary/test_flip.c
> +++ b/src/bin/elementary/test_flip.c
> @@ -271,7 +271,7 @@ void
>  test_flip2(void *data EINA_UNUSED, Evas_Object *obj EINA_UNUSED, void
> *event_info EINA_UNUSED) {
>     Evas_Object *win, *bg, *bx, *bx2, *fl, *o, *bt, *tb, *li, *en;
> -   char buf[PATH_MAX];
> +   char buf[4096];
>  
>     win = elm_win_add(NULL, "flip2", ELM_WIN_BASIC);
>     elm_win_title_set(win, "Flip 2");
> 
> 
> -- 
> To stop receiving notification emails like this one, please contact
> the administrator of this repository.


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

Re: [E-devel] What's the state of Eo?

[Carsten H. <ra...@ra...>]

On Mon, 28 Apr 2025 13:59:27 -0300 Vinícius dos Santos Oliveira
<vin...@gm...> said:

eo itself is fine. everything in efl is built on it pretty much - including
legacy api's on top. eo efl api's (beta) are still beta.

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

[E-devel] What's the state of Eo?

[Vinícius d. S. O. <vin...@gm...>]

Is it stable already?

-- 
Vinícius dos Santos Oliveira
https://vinipsmaker.github.io/

Re: [E-devel] imlib2-1.12.5

[<jos...@ju...>]

Beautiful, ... so long ago.

---------- Original Message ----------
From: Kim Woelders <ki...@wo...>
To: e-announce <enl...@li...>, e-release <enl...@li...>, e-users <enl...@li...>, e-devel <enl...@li...>
Subject: [E-devel] imlib2-1.12.5
Date: Mon, 7 Apr 2025 08:34:28 +0200

imlib2-1.12.5:

A couple of fixes:
- Fix scaling regression in 1.12.4
- Fix libyuv detection when no .pc file is provided
- Make RTLD_GLOBAL requirement optional (default as in pre-1.12.4)

Source packages:
https://sourceforge.net/projects/enlightenment/files/imlib2-src/1.12.5/

/Kim


_______________________________________________
enlightenment-devel mailing list
enl...@li...
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Re: [E-devel] Mount volumes

[Quelrond <que...@gm...>]

Finally... Everything works 'out-of-the-box', I just had some strange 
problems on the PC where I tested it.

Starting bsdisks (FreeBSD analog of Udisks2), I get DBUS messages about 
changing of devices status. Enlightenment sees them, so activating two 
options in E settings -  "Icons on Desktop" (File Manager Settings -> 
Display)  and "Show device icons on desktop" (File Manager Settings -> 
Device) I've got icons appeared on desktop and in EFM. It is possible to 
mount/unmount volumes now. So, neither eeze is needed for this, not 
udev. And udisks is active by default, so FreeBSD port is built with 
this option.

Peter


On 08/04/2025 16:26, Carsten Haitzler wrote:
> On Tue, 8 Apr 2025 13:24:00 +0000 Quelrond <que...@gm...> said:
>
>> Hello,
>>
>> I would like to enable mount/unmount volumes support in Enlightenment
>> under FreeBSD.
>>
>> I see three options in meson_options.txt of Enlightenment:
>>
>> - 'device-udev'
>> - 'mount-udisks'
>> - 'mount-eeze'
>>
>> It seems that eeze needs libmount that is not present in FreeBSD, so,
>> probably, the last option cannot work.
>>
>> udev and udisks are present and enables by default, but Enlightenment
>> does not show devices on desktop (yes, this option is activated in EFM
>> settings).
>>
>> How should the components work together and how can I debug them?
> at this point i'm not going to touch these bits of efm... why?
>
> https://git.enlightenment.org/raster/efm2
>
> this is a temporary place to build efm2 code until it's fully featured enough
> to merge in and replace current efm.
>
> it's a new design with a backend process (per dir) that you speak to over
> stdin/out - the idea is you can replace the backend with any other script
> (bash, python or anything or any binary that can do stdin/out). i alrady made a
> custom view bash script:
>
> https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/table/open
>
> and this is the default standard filesystem one:
>
> https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/default
>
> the backend is responsible for generating of thumbnails as well as metadata
> saving/loading and tracking status of background operations. the thumbnailer is
> pretty fully featured there. does all the usual and also can download images
> from the internet like movie posters and music album art and so on. it guesses
> on what to do. for theme files it can render a mini screen of content i the
> theme, thumbnail ttf fonts too. for paged content (pdf, ps, doc etc.) it will
> pre-render the first 32 pages into the thumbnail for movies/video it also
> renderes multiple frames in the file into the thumbnail ... i was going to have
> it later also maybe do some kind of waveform render for wav and other
> music/audio files. i need to make a thumbnail command kit that can be called to
> examine and modify (insert/remove) images from the thumbnail files. still -
> this would be necessary if you replace the default thumbnailer with your own
> in your own vfs backend. but that's not hard - easy to do - all the code is
> there.
>
> the reason i point to this is... this is meant to be a lot more modular and
> easier to extend and change, which when i get to handling mounting of
> filesystems - i'll follow a similar design with some process that sits around
> and probably writes some files to disk to indicate if there is something you
> can mount (or have fixed files for fixed mounts - be they local block devices or
> sshfs or smb or anything else). this should ultimately make it easier to
> support bsd.
>
> at this point i'm fleshing out the backend and frontend features. dnd currently
> works (it moves files). you can navigate dirs and it'll list files and
> generate/fetch thumbnails or album art/covers or movie posters etc. i don't have
> dnd with "copy" or "ask" done. copy & paste is not done (cut is actually the
> big problem - when you cut files ... what do i do with the cut files?
> conceptually i should move them into a special cut buffer dir - available for
> paste, but that's not sane. i could keep a special cut buffer list of file
> paths and all files when loaded by a backend filter out listing any files in
> the cut buffer list and pretend they are not there... not sure). i don't have
> any delete done (will have to support full unlink directly or trash can stuff).
> i need as trashcan vfs implementation too. i am recently working on mime
> handling and that now has me at the point of "what is the ui model for allowing
> the user to select any app/desktop to open a set of files with".
>
> anyway... efm2 is where i am putting effort as there isn't much point putting
> effort into efm now as it'll be replaced by efm2 (efm2 will migrate into the e
> tree then). so right now the best thing to do is to look at efm2 and how it
> works and make THAT work for you on bsd. first see if efm2 currently works fine
> and find any issues that are not "i just haven't done that yet".
>
> efm2 already has multiple icon views (free-placement - i.,e. icon stays where
> you manually dnd it like on the desktop with horizontal and vertical
> auto-arrange modes, auto icon grid view with horiz or vert layout, simple list
> mode and a detailed column list mode).
>
> https://www.enlightenment.org/ss/e-67df028d5a3310.10486448.jpg
> https://www.enlightenment.org/ss/e-67df00c3916aa9.36155616.jpg
> https://www.enlightenment.org/ss/e-67df0332df1af7.22084925.jpg
> https://www.enlightenment.org/ss/e-67df03ccd92de6.77421391.jpg
> https://www.enlightenment.org/ss/e-67df03aa7e3a98.79713538.jpg
> https://www.enlightenment.org/ss/e-67df039ae6b9b6.52391813.jpg
>
> columns are also dragable (click and drag the bar between columns) and sortable
> by clicking on the header.
>
> the ui is a test ui - the main scrollable icon view is what is being worked on
> -t he buttons around it are just for testing this. so don't assume this is what
> efm2 will look like. it's a bunch of test code.
>
> anyway - this is where you want to be looking and spending effort. as per your
> removable device + mounting problem... i haven't started solving that yet but
> as above, my ideas are:
>
> 1. some process that sits around and maybe can run other or binaries in a
> polling fashion or as a result of some system event. these would then put
> special files in some dir that indicate some fs you can open/mount
> 2. per mountable fs there would be some kind of mount and unmount handler
> "binary/script" called when the fs first needs mounting or called when not
> needed anymore (unmount time) as well as indicating a custom vfs if needed
> other than the default one.
>

Re: [E-devel] Mount volumes

[Quelrond <que...@gm...>]

Thanks for this clarification, Raster.
As for now - I do not suppose to touch the code of EFM or any other 
component, I would like just to understand how does it work in the 
current state (it seems to work in Linux).
Then - yes, I understand the idea of EFM2 and I suppose that with such 
approach the management of mount/unmount will be much more simple.
In the current state it seems that udev is not used in this process, but 
with udisks, even without eeze it should work somehow. udisk itself is 
available
https://www.freshports.org/sysutils/libudisks

...and it works (udisksctl shows correctly add/remove of usb disk).
Maybe, E needs eeze and libmount, maybe something else, but it does not 
show the device icon on desktop after insert.

Peter

On 08/04/2025 14:26, Carsten Haitzler wrote:
> On Tue, 8 Apr 2025 13:24:00 +0000 Quelrond <que...@gm...> said:
>
>> Hello,
>>
>> I would like to enable mount/unmount volumes support in Enlightenment
>> under FreeBSD.
>>
>> I see three options in meson_options.txt of Enlightenment:
>>
>> - 'device-udev'
>> - 'mount-udisks'
>> - 'mount-eeze'
>>
>> It seems that eeze needs libmount that is not present in FreeBSD, so,
>> probably, the last option cannot work.
>>
>> udev and udisks are present and enables by default, but Enlightenment
>> does not show devices on desktop (yes, this option is activated in EFM
>> settings).
>>
>> How should the components work together and how can I debug them?
> at this point i'm not going to touch these bits of efm... why?
>
> https://git.enlightenment.org/raster/efm2
>
> this is a temporary place to build efm2 code until it's fully featured enough
> to merge in and replace current efm.
>
> it's a new design with a backend process (per dir) that you speak to over
> stdin/out - the idea is you can replace the backend with any other script
> (bash, python or anything or any binary that can do stdin/out). i alrady made a
> custom view bash script:
>
> https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/table/open
>
> and this is the default standard filesystem one:
>
> https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/default
>
> the backend is responsible for generating of thumbnails as well as metadata
> saving/loading and tracking status of background operations. the thumbnailer is
> pretty fully featured there. does all the usual and also can download images
> from the internet like movie posters and music album art and so on. it guesses
> on what to do. for theme files it can render a mini screen of content i the
> theme, thumbnail ttf fonts too. for paged content (pdf, ps, doc etc.) it will
> pre-render the first 32 pages into the thumbnail for movies/video it also
> renderes multiple frames in the file into the thumbnail ... i was going to have
> it later also maybe do some kind of waveform render for wav and other
> music/audio files. i need to make a thumbnail command kit that can be called to
> examine and modify (insert/remove) images from the thumbnail files. still -
> this would be necessary if you replace the default thumbnailer with your own
> in your own vfs backend. but that's not hard - easy to do - all the code is
> there.
>
> the reason i point to this is... this is meant to be a lot more modular and
> easier to extend and change, which when i get to handling mounting of
> filesystems - i'll follow a similar design with some process that sits around
> and probably writes some files to disk to indicate if there is something you
> can mount (or have fixed files for fixed mounts - be they local block devices or
> sshfs or smb or anything else). this should ultimately make it easier to
> support bsd.
>
> at this point i'm fleshing out the backend and frontend features. dnd currently
> works (it moves files). you can navigate dirs and it'll list files and
> generate/fetch thumbnails or album art/covers or movie posters etc. i don't have
> dnd with "copy" or "ask" done. copy & paste is not done (cut is actually the
> big problem - when you cut files ... what do i do with the cut files?
> conceptually i should move them into a special cut buffer dir - available for
> paste, but that's not sane. i could keep a special cut buffer list of file
> paths and all files when loaded by a backend filter out listing any files in
> the cut buffer list and pretend they are not there... not sure). i don't have
> any delete done (will have to support full unlink directly or trash can stuff).
> i need as trashcan vfs implementation too. i am recently working on mime
> handling and that now has me at the point of "what is the ui model for allowing
> the user to select any app/desktop to open a set of files with".
>
> anyway... efm2 is where i am putting effort as there isn't much point putting
> effort into efm now as it'll be replaced by efm2 (efm2 will migrate into the e
> tree then). so right now the best thing to do is to look at efm2 and how it
> works and make THAT work for you on bsd. first see if efm2 currently works fine
> and find any issues that are not "i just haven't done that yet".
>
> efm2 already has multiple icon views (free-placement - i.,e. icon stays where
> you manually dnd it like on the desktop with horizontal and vertical
> auto-arrange modes, auto icon grid view with horiz or vert layout, simple list
> mode and a detailed column list mode).
>
> https://www.enlightenment.org/ss/e-67df028d5a3310.10486448.jpg
> https://www.enlightenment.org/ss/e-67df00c3916aa9.36155616.jpg
> https://www.enlightenment.org/ss/e-67df0332df1af7.22084925.jpg
> https://www.enlightenment.org/ss/e-67df03ccd92de6.77421391.jpg
> https://www.enlightenment.org/ss/e-67df03aa7e3a98.79713538.jpg
> https://www.enlightenment.org/ss/e-67df039ae6b9b6.52391813.jpg
>
> columns are also dragable (click and drag the bar between columns) and sortable
> by clicking on the header.
>
> the ui is a test ui - the main scrollable icon view is what is being worked on
> -t he buttons around it are just for testing this. so don't assume this is what
> efm2 will look like. it's a bunch of test code.
>
> anyway - this is where you want to be looking and spending effort. as per your
> removable device + mounting problem... i haven't started solving that yet but
> as above, my ideas are:
>
> 1. some process that sits around and maybe can run other or binaries in a
> polling fashion or as a result of some system event. these would then put
> special files in some dir that indicate some fs you can open/mount
> 2. per mountable fs there would be some kind of mount and unmount handler
> "binary/script" called when the fs first needs mounting or called when not
> needed anymore (unmount time) as well as indicating a custom vfs if needed
> other than the default one.
>

Re: [E-devel] Mount volumes

[Carsten H. <ra...@ra...>]

On Tue, 8 Apr 2025 13:24:00 +0000 Quelrond <que...@gm...> said:

> Hello,
> 
> I would like to enable mount/unmount volumes support in Enlightenment 
> under FreeBSD.
> 
> I see three options in meson_options.txt of Enlightenment:
> 
> - 'device-udev'
> - 'mount-udisks'
> - 'mount-eeze'
> 
> It seems that eeze needs libmount that is not present in FreeBSD, so, 
> probably, the last option cannot work.
> 
> udev and udisks are present and enables by default, but Enlightenment 
> does not show devices on desktop (yes, this option is activated in EFM 
> settings).
> 
> How should the components work together and how can I debug them?

at this point i'm not going to touch these bits of efm... why?

https://git.enlightenment.org/raster/efm2

this is a temporary place to build efm2 code until it's fully featured enough
to merge in and replace current efm.

it's a new design with a backend process (per dir) that you speak to over
stdin/out - the idea is you can replace the backend with any other script
(bash, python or anything or any binary that can do stdin/out). i alrady made a
custom view bash script:

https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/table/open

and this is the default standard filesystem one:

https://git.enlightenment.org/raster/efm2/src/branch/master/src/backends/default

the backend is responsible for generating of thumbnails as well as metadata
saving/loading and tracking status of background operations. the thumbnailer is
pretty fully featured there. does all the usual and also can download images
from the internet like movie posters and music album art and so on. it guesses
on what to do. for theme files it can render a mini screen of content i the
theme, thumbnail ttf fonts too. for paged content (pdf, ps, doc etc.) it will
pre-render the first 32 pages into the thumbnail for movies/video it also
renderes multiple frames in the file into the thumbnail ... i was going to have
it later also maybe do some kind of waveform render for wav and other
music/audio files. i need to make a thumbnail command kit that can be called to
examine and modify (insert/remove) images from the thumbnail files. still -
this would be necessary if you replace the default thumbnailer with your own
in your own vfs backend. but that's not hard - easy to do - all the code is
there.

the reason i point to this is... this is meant to be a lot more modular and
easier to extend and change, which when i get to handling mounting of
filesystems - i'll follow a similar design with some process that sits around
and probably writes some files to disk to indicate if there is something you
can mount (or have fixed files for fixed mounts - be they local block devices or
sshfs or smb or anything else). this should ultimately make it easier to
support bsd.

at this point i'm fleshing out the backend and frontend features. dnd currently
works (it moves files). you can navigate dirs and it'll list files and
generate/fetch thumbnails or album art/covers or movie posters etc. i don't have
dnd with "copy" or "ask" done. copy & paste is not done (cut is actually the
big problem - when you cut files ... what do i do with the cut files?
conceptually i should move them into a special cut buffer dir - available for
paste, but that's not sane. i could keep a special cut buffer list of file
paths and all files when loaded by a backend filter out listing any files in
the cut buffer list and pretend they are not there... not sure). i don't have
any delete done (will have to support full unlink directly or trash can stuff).
i need as trashcan vfs implementation too. i am recently working on mime
handling and that now has me at the point of "what is the ui model for allowing
the user to select any app/desktop to open a set of files with".

anyway... efm2 is where i am putting effort as there isn't much point putting
effort into efm now as it'll be replaced by efm2 (efm2 will migrate into the e
tree then). so right now the best thing to do is to look at efm2 and how it
works and make THAT work for you on bsd. first see if efm2 currently works fine
and find any issues that are not "i just haven't done that yet".

efm2 already has multiple icon views (free-placement - i.,e. icon stays where
you manually dnd it like on the desktop with horizontal and vertical
auto-arrange modes, auto icon grid view with horiz or vert layout, simple list
mode and a detailed column list mode).

https://www.enlightenment.org/ss/e-67df028d5a3310.10486448.jpg
https://www.enlightenment.org/ss/e-67df00c3916aa9.36155616.jpg
https://www.enlightenment.org/ss/e-67df0332df1af7.22084925.jpg
https://www.enlightenment.org/ss/e-67df03ccd92de6.77421391.jpg
https://www.enlightenment.org/ss/e-67df03aa7e3a98.79713538.jpg
https://www.enlightenment.org/ss/e-67df039ae6b9b6.52391813.jpg

columns are also dragable (click and drag the bar between columns) and sortable
by clicking on the header.

the ui is a test ui - the main scrollable icon view is what is being worked on
-t he buttons around it are just for testing this. so don't assume this is what
efm2 will look like. it's a bunch of test code.

anyway - this is where you want to be looking and spending effort. as per your
removable device + mounting problem... i haven't started solving that yet but
as above, my ideas are:

1. some process that sits around and maybe can run other or binaries in a
polling fashion or as a result of some system event. these would then put
special files in some dir that indicate some fs you can open/mount
2. per mountable fs there would be some kind of mount and unmount handler
"binary/script" called when the fs first needs mounting or called when not
needed anymore (unmount time) as well as indicating a custom vfs if needed
other than the default one.

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

[E-devel] Mount volumes

[Quelrond <que...@gm...>]

Hello,

I would like to enable mount/unmount volumes support in Enlightenment 
under FreeBSD.

I see three options in meson_options.txt of Enlightenment:

- 'device-udev'
- 'mount-udisks'
- 'mount-eeze'

It seems that eeze needs libmount that is not present in FreeBSD, so, 
probably, the last option cannot work.

udev and udisks are present and enables by default, but Enlightenment 
does not show devices on desktop (yes, this option is activated in EFM 
settings).

How should the components work together and how can I debug them?

Best regards,

Peter

[E-devel] imlib2-1.12.5

[Kim W. <ki...@wo...>]

imlib2-1.12.5:

A couple of fixes:
- Fix scaling regression in 1.12.4
- Fix libyuv detection when no .pc file is provided
- Make RTLD_GLOBAL requirement optional (default as in pre-1.12.4)

Source packages:
https://sourceforge.net/projects/enlightenment/files/imlib2-src/1.12.5/

/Kim

[E-devel] enlightenment 0.27.1

[Carsten H. <ra...@ra...>]

Release update to enlightenment to fix some bugs. 0.27.1. Fix the following:

  * cpufreq - fix some startup conditions with 0 sizes and counts
  * mixer - allow wheel on mixer
  * signal handlers - report more info in log
  * msgbus - fix free on shutdown on e in screen blank blockers
  * battery - limit to 100% health
  * bindings - fix lookup of null actions
  * dbus menu + systray - fix mem access on bad unref
  * client menu - dont show kbd shortcut if no desktop
  * valgrind - bring back working valgrind option
  * efm - fix shutdown of file monitor on file dir close


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

[E-devel] efl 1.28.1 release

[Carsten H. <ra...@ra...>]

Micro Release fixes some bugs in 1.28.0

  * Some typos in error debug prints
  * Fix build of ecore audio on windows with GUID changes
  * Fix edje player to not have black bg when alpha win used
  * Fix annoying warning on efreet desktop refs on exit
  * Fix access of uninit dpms state on displays with no dpms
  * Fix eina thread queue misalignment of data blocks

http://download.enlightenment.org/rel/libs/efl/efl-1.28.1.tar.xz

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

[E-devel] imlib2-1.12.4

[Kim W. <ki...@wo...>]

imlib2-1.12.4:

Mostly module loader/saver additions and improvements
- Added QOI saver
- Added HEIF saver
- Added AVIF loader/saver
- Y4M loader: Various improvements
- QOI loader: Various improvements
- SVG loader: Handle .svgz too
- scaling: Fix potential crash when scaling large images
- No longer link modules with libImlib2.la

Source packages:
https://sourceforge.net/projects/enlightenment/files/imlib2-src/1.12.4/

/Kim

Re: [E-devel] Malloc Pants - Key new feature!

[Carsten H. <ra...@ra...>]

On Wed, 12 Feb 2025 11:03:38 -0300 Felipe Magno de Almeida
<fel...@gm...> said:

> Hello Raster,
> 
> To have malloc on my desktop is something I'm looking forward to trying. It
> would be interesting if we could animate it.

PantsDB is a powerful database. It allows not just pngs but also animated
gifs, webps too as does the Pants module gadget.

Unfortunately the work required to have an animated Malloc in and not in pants
might be a bit too much work at this point, but PantsDB+Pants module are
capable of handling such an entry, should it be added.

Let the over-engineering continue!

:)

> Kind regards,
> 
> On Wed, Feb 12, 2025 at 11:01 AM Carsten Haitzler <ra...@ra...>
> wrote:
> 
> > In news of critical importance to all E users, we have a new feature in the
> > PantsDB of the Pants module:
> >
> > https://www.enlightenment.org/ss/e-67a761271e0612.43864909.png
> >
> > There is now an option of a Malloc with pants on and pants off. It's in git
> > master. I have considered this egregious lack of floof in Enlightenment of
> > late
> > and have worked hard to resolve this gap in Enlightenment by providing
> > such a
> > new feature.
> >
> > I hope this fills in such an important gap in Enlightenment's feature set
> > and
> > makes your days more manageable as a result.
> >
> > --
> > ------------- Codito, ergo sum - "I code, therefore I am" --------------
> > Carsten Haitzler - ra...@ra...
> >
> >
> >
> > _______________________________________________
> > enlightenment-devel mailing list
> > enl...@li...
> > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> >
> 
> 
> -- 
> Felipe Magno de Almeida
> Owner @ Expertise Solutions
> www: https://expertise.dev
> phone: +55 48 9 9681.0157
> LinkedIn: in/felipealmeida
> 
> _______________________________________________
> enlightenment-devel mailing list
> enl...@li...
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...

Re: [E-devel] Malloc Pants - Key new feature!

[Felipe M. de A. <fel...@gm...>]

Hello Raster,

To have malloc on my desktop is something I'm looking forward to trying. It
would be interesting if we could animate it.

Kind regards,

On Wed, Feb 12, 2025 at 11:01 AM Carsten Haitzler <ra...@ra...>
wrote:

> In news of critical importance to all E users, we have a new feature in the
> PantsDB of the Pants module:
>
> https://www.enlightenment.org/ss/e-67a761271e0612.43864909.png
>
> There is now an option of a Malloc with pants on and pants off. It's in git
> master. I have considered this egregious lack of floof in Enlightenment of
> late
> and have worked hard to resolve this gap in Enlightenment by providing
> such a
> new feature.
>
> I hope this fills in such an important gap in Enlightenment's feature set
> and
> makes your days more manageable as a result.
>
> --
> ------------- Codito, ergo sum - "I code, therefore I am" --------------
> Carsten Haitzler - ra...@ra...
>
>
>
> _______________________________________________
> enlightenment-devel mailing list
> enl...@li...
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
>


-- 
Felipe Magno de Almeida
Owner @ Expertise Solutions
www: https://expertise.dev
phone: +55 48 9 9681.0157
LinkedIn: in/felipealmeida

[E-devel] Malloc Pants - Key new feature!

[Carsten H. <ra...@ra...>]

In news of critical importance to all E users, we have a new feature in the
PantsDB of the Pants module:

https://www.enlightenment.org/ss/e-67a761271e0612.43864909.png

There is now an option of a Malloc with pants on and pants off. It's in git
master. I have considered this egregious lack of floof in Enlightenment of late
and have worked hard to resolve this gap in Enlightenment by providing such a
new feature.

I hope this fills in such an important gap in Enlightenment's feature set and
makes your days more manageable as a result.

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ra...@ra...