Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Dantrell's GNOME Without Systemd Project (Part 2)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 10, 11, 12 ... 14, 15, 16  Next  
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Unsupported Software
View previous topic :: View next topic  
Author Message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Fri Dec 14, 2018 8:07 pm    Post subject: Reply with quote

Cuong Nguyen wrote:
What is profile dantrell-gnome:default/linux/amd64/17.0/desktop/gnome+plasma.

It ensures clean emerge output on a freshly extracted stage 3 tarball when simultaneously installing gnome-base/gnome and kde-plasma/plasma-meta.

Cuong Nguyen wrote:
When should I use this profile?

When you want both GNOME and KDE desktop environments and/or applications.

Cuong Nguyen wrote:
Why it mixes GNOME with KDE?

Because Gentoo is about choice. :wink:
_________________
Dantrell B.
Back to top
View user's profile Send private message
Cuong Nguyen
Tux's lil' helper
Tux's lil' helper


Joined: 18 Jan 2018
Posts: 130

PostPosted: Sat Dec 15, 2018 9:34 am    Post subject: Reply with quote

dantrell wrote:

Cuong Nguyen wrote:
Why it mixes GNOME with KDE?

Because Gentoo is about choice. :wink:


Thank you for good works, my choice is non-systemd gnome :D

I deployed your patch sets since 3.24 or 3.26. I actually created a local profile that points to your /extended profile and joined with gentoo /hardened/selinux profile. Now I just need to get your gnome+plasma profile as parent, because it points to the /extended profile.
/var/lib/layman/dantrell-gnome/profiles/default/linux/amd64/17.0/desktop/gnome+plasma/parent
Code:

gentoo:default/linux/amd64/17.0/desktop/plasma
dantrell-gnome-3-30:default/linux/amd64/17.0/desktop/gnome/3.30/extended
../../../..


My last build of 3.30 gnome failed at run-time with gnome-shell segfault calling libmozjs from spidermonkey:60. Maybe because I used lots of optimization (like -O3 -ffast-math and graphite, lto things). I am now rebuilding with relaxed optimization flags.
Back to top
View user's profile Send private message
Ultimator
n00b
n00b


Joined: 23 Jun 2018
Posts: 18

PostPosted: Sat Dec 15, 2018 2:30 pm    Post subject: Reply with quote

Cuong Nguyen wrote:
I actually created a local profile that points to your /extended profile and joined with gentoo /hardened/selinux profile.

Be careful when you do this. The hardened/selinux profile also includes the whole default profile stuff which may overwrite changes from dantrell's profile. You can check the order in which your profiles are loaded with this python code (got it from here)
Code:
import portage
for p in portage.settings.profiles:
    print("%s" % p)
Back to top
View user's profile Send private message
Cuong Nguyen
Tux's lil' helper
Tux's lil' helper


Joined: 18 Jan 2018
Posts: 130

PostPosted: Sat Dec 15, 2018 3:21 pm    Post subject: Reply with quote

Ultimator wrote:
Cuong Nguyen wrote:
I actually created a local profile that points to your /extended profile and joined with gentoo /hardened/selinux profile.

Be careful when you do this. The hardened/selinux profile also includes the whole default profile stuff which may overwrite changes from dantrell's profile. You can check the order in which your profiles are loaded with this python code (got it from here)
Code:
import portage
for p in portage.settings.profiles:
    print("%s" % p)


Thank you, I got messed with the order of combined profiles many times and I usually ended up with hardcoded use.mask, package.use.mask, mask.defaults in /etc/portage/profile folder.

Other gentoo-based distros ignored original gentoo profiles and uses theirown profiles and eclass (Calculate Linux) or different pkg manager (Sabayon).
Back to top
View user's profile Send private message
Shibotto
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jun 2015
Posts: 122
Location: Italy

PostPosted: Wed Dec 19, 2018 12:20 am    Post subject: Reply with quote

dantrell wrote:
That's fine since GNOME Without Systemd also supports GNOME With Systemd. :twisted:

>inb4 README-MAYBE.md
O:-)


I think I'm done. Proposed upstream and upstreamer. I'm not so confident about the ebuild, which I'm not that accustomed to yet, so feel free to take what's good and throw away everything else :D
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Thu Dec 20, 2018 12:34 am    Post subject: Reply with quote

Shibotto wrote:
I think I'm done. Proposed upstream and upstreamer. I'm not so confident about the ebuild, which I'm not that accustomed to yet, so feel free to take what's good and throw away everything else :D

Nice work! :D

I'm a bit preoccupied so I won't be getting around to it today but when I do I'll be using the patch as-is where applicable.
_________________
Dantrell B.
Back to top
View user's profile Send private message
Shibotto
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jun 2015
Posts: 122
Location: Italy

PostPosted: Fri Dec 21, 2018 12:29 am    Post subject: Reply with quote

dantrell wrote:
Nice work! :D

I'm a bit preoccupied so I won't be getting around to it today but when I do I'll be using the patch as-is where applicable.

Thank you :D
Take your time and don't worry, on my machines everything's already been patched anyway ahhahaha 8)

Well, for something completely different I found out that upstream gdk-pixbuf-thumbnailer (x11-libs/gdk-pixbuf-2.38.0) is completely broken for animated GIF. :lol:
Code:
shiba ~ > gdk-pixbuf-thumbnailer supercoolanimation.gif ohnoes.png

(gdk-pixbuf-thumbnailer:8894): GdkPixbuf-CRITICAL **: 00:29:43.335: gdk_pixbuf_save: assertion 'error == NULL || *error == NULL' failed

** (gdk-pixbuf-thumbnailer:8894): WARNING **: 00:29:43.336: Couldn't save the thumbnail 'ohnoes.png' for file 'supercoolanimation.gif': Not all frames of the GIF image were loaded.

This commit solves the issue. I find it curiously fortunate that despite 2.38.0 being out for several months, this problem has been spotted and fixed only a week ago, just in time to save me at least an hour of cross-distro testing and bisecting. 8O

Anyway, I think there are three viable solutions: mask 2.38.0, patch 2.38.0-r1, do nothing and wait for 2.38.1 (which hopefully won't be much further away, given a broken thumbnailer and a current stable release going back to 1 Sep).
Back to top
View user's profile Send private message
Cuong Nguyen
Tux's lil' helper
Tux's lil' helper


Joined: 18 Jan 2018
Posts: 130

PostPosted: Fri Dec 21, 2018 8:08 am    Post subject: Reply with quote

Strange behavior emerging dev-util/glib-utils-2.58.2::dantrell-gnome

No binary files created?

https://paste.pound-python.org/show/dHxKMGEHva3AGtQtMiLP/

Please help, thank you.

Code:

>>> Install glib-utils-2.58.2 into /var/tmp/portage/dev-util/glib-utils-2.58.2/image/ category dev-util
>>> Completed installing glib-utils-2.58.2 into /var/tmp/portage/dev-util/glib-utils-2.58.2/image/

 * Final size of build directory: 0 KiB
 * Final size of installed tree:  0 KiB

>>> Done.

>>> Installing (1 of 1) dev-util/glib-utils-2.58.2::dantrell-gnome
 * removing unneeded *.la files
!!! Unable to set SELinux security labels
>>> Auto-cleaning packages...

>>> No outdated packages were found on your system.


EDIT glib-utils-2.58.2.ebuild file is empty??!!

Code:
# Distributed under the terms of the GNU General Public License v2

EAPI="6"

DESCRIPTION="Build utilities for GLib using projects"

SLOT="0"
KEYWORDS="*"

IUSE=""

RDEPEND=""
Back to top
View user's profile Send private message
Shibotto
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jun 2015
Posts: 122
Location: Italy

PostPosted: Fri Dec 21, 2018 7:52 pm    Post subject: Reply with quote

Cuong Nguyen wrote:
glib-utils-2.58.2.ebuild file is empty??!!

It seems to be included in dev-libs/glib. If I got this right, it got split for Gentoo stage3 needs. It doesn't concern an overlay though, so maybe that's the reason.
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Sat Dec 22, 2018 3:59 am    Post subject: Reply with quote

dantrell wrote:
I'm a bit preoccupied so I won't be getting around to it today but when I do I'll be using the patch as-is where applicable.

Shibotto wrote:
Thank you :D

I patched it but these changes won't be going live just yet.

Shibotto wrote:
Well, for something completely different I found out that upstream gdk-pixbuf-thumbnailer (x11-libs/gdk-pixbuf-2.38.0) is completely broken for animated GIF. :lol:

You have been on a roll lately. :o

Shibotto wrote:
Anyway, I think there are three viable solutions: mask 2.38.0, patch 2.38.0-r1, do nothing and wait for 2.38.1 (which hopefully won't be much further away, given a broken thumbnailer and a current stable release going back to 1 Sep).

This issue was corrected (I patched 2.38.0 until 2.38.1 rolls out).

Sync up and you should be good to go.

Cuong Nguyen wrote:
No binary files created?

Cuong Nguyen wrote:
glib-utils-2.58.2.ebuild file is empty??!!

This is correct since dev-util/glib-utils as provided by the project overlay is basically just a virtual package.

Shibotto wrote:
It seems to be included in dev-libs/glib. If I got this right, it got split for Gentoo stage3 needs.

Indeed, it's still included in dev-libs/glib on my end.

Shibotto wrote:
It doesn't concern an overlay though, so maybe that's the reason.

I do a full review against Gentoo a couple times a year but even then I don't adapt every change.
_________________
Dantrell B.
Back to top
View user's profile Send private message
Cuong Nguyen
Tux's lil' helper
Tux's lil' helper


Joined: 18 Jan 2018
Posts: 130

PostPosted: Sat Dec 22, 2018 8:09 am    Post subject: Reply with quote

dantrell wrote:

Cuong Nguyen wrote:
glib-utils-2.58.2.ebuild file is empty??!!

This is correct since dev-util/glib-utils as provided by the project overlay is basically just a virtual package.


Does it mean I have to emerge dev-util/glib-utils::gentoo before ::dantrell-gnome? Because some latter packages require a tool from /usr/bin of dev-uitl/glib-utils?

I actually get into dantrell-gnome repos right after installing stage3, updated toolchain and switch to dantrell-gnome-3-30 profile. Then I ran emerge -uDN @world.

Should I start with plain gentoo desktop/gnome profile before?

Thank you
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Sun Dec 23, 2018 4:30 pm    Post subject: Reply with quote

dantrell wrote:
I patched it but these changes won't be going live just yet.

It went live for GNOME 3.30 at least.

Cuong Nguyen wrote:
Does it mean I have to emerge dev-util/glib-utils::gentoo before ::dantrell-gnome? Because some latter packages require a tool from /usr/bin of dev-uitl/glib-utils?

As was already stated, dev-util/glib-utils was not spliced from dev-libs/glib in the project overlays.

If you experience an actual issue then just oneshot dev-libs/glib to work around it and then report the issue to me.

Cuong Nguyen wrote:
I actually get into dantrell-gnome repos right after installing stage3, updated toolchain and switch to dantrell-gnome-3-30 profile. Then I ran emerge -uDN @world.

This follows along the line of the project documentation.

Cuong Nguyen wrote:
Should I start with plain gentoo desktop/gnome profile before?

This doesn't and is unsupported.
_________________
Dantrell B.
Back to top
View user's profile Send private message
Shibotto
Tux's lil' helper
Tux's lil' helper


Joined: 19 Jun 2015
Posts: 122
Location: Italy

PostPosted: Mon Dec 24, 2018 12:14 am    Post subject: Reply with quote

dantrell wrote:
You have been on a roll lately. :o

Yeah, I have some sort of discipline problem when I'm assigned too many jobs I care very little about :lol: :lol: :roll:

And that's precisely why I found another one :twisted:
While investigating something completely different (at the end of this post) I stumbled upon a suspicious /lib/udev/rules.d/rules.d/61-gnome-settings-daemon-rfkill.rules. A very quick look and an even simpler fix:
Code:
diff --git a/gnome-base/gnome-settings-daemon/gnome-settings-daemon-3.30.1.2-r1.ebuild b/gnome-base/gnome-settings-daemon/gnome-settings-daemon-3.30.1.2-r1.ebuild
index 578ea40..72046ba 100644
--- a/gnome-base/gnome-settings-daemon/gnome-settings-daemon-3.30.1.2-r1.ebuild
+++ b/gnome-base/gnome-settings-daemon/gnome-settings-daemon-3.30.1.2-r1.ebuild
@@ -124,7 +124,7 @@ src_prepare() {
 
 src_configure() {
        local emesonargs=(
-               -D udev_dir="$(get_udevdir)"/rules.d
+               -D udev_dir="$(get_udevdir)"
                -D alsa=true
                -D gudev=$(usex udev true false)
                -D color=$(usex colord true false)

Apparently the syntax changed when they transitioned to Meson (figures), requiring the path to udev base dir and only subsequently merging it when building the rfkill plugin.

---

What I was researching instead, was my problem with remote thumbnails, which I was able to solve. It turned out I didn't load the FUSE module, which GVfs relies on to provide a local path to applications requiring it, so adding it solved the problem. Since I built my kernel config starting from two systemd based distros, I guess there is something in systemd automating the process.

Mh... I think the last time I had to manually load FUSE was in the MODULES string of /etc/rc.conf. Nostalgic memories of an Arch Linux in a distant past without systemd, a decade ago. :roll:
Back to top
View user's profile Send private message
Spargeltarzan
Apprentice
Apprentice


Joined: 23 Jul 2017
Posts: 293

PostPosted: Mon Dec 24, 2018 1:11 pm    Post subject: Reply with quote

dantrell wrote:

If you have even a small suspicion that the combined profile is the cause then you need to uninstall it (and don't just switch profiles, you need to completely delete or otherwise dereference the project overlays) and test Plasma by itself. If the problem persists then you have the your answer.

I migrated to the official Plasma profile as you described and Plasma still freezes. So your profile is not the cause. I changed back and use your GNOME without systemd fully again. My PC runs stable with it (Desktop PC) :).

My only (little) issue what I have (still) is when I login my gnome keyring does not become autounlocked. So when I boot my PC firstly I login, then nextcloud wants a password and finally gnome-keyring...

Some months ago I tried GNOME without systemd also on my notebook, but I didn't manage to get the same battery life time as with the official systemd. I tried tlp from the overlay, laptop-mode-tools, powertop, played with the CPU kernel settings (governers, etc.). The lifetime is amazingly different, I mostly run all day long on my notebook (so much, that I even can't say how long) and with openrc after 2-3 hours I needed the power supply.
_________________
___________________
Regards

Spargeltarzan

Notebook: Lenovo YOGA 900-13ISK: Gentoo stable amd64, GNOME systemd, KVM/QEMU
Desktop-PC: Intel Core i7-4770K, 8GB Ram, AMD Radeon R9 280X, ZFS Storage, GNOME openrc, Dantrell, Xen
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Tue Dec 25, 2018 6:21 pm    Post subject: Reply with quote

Shibotto wrote:
Yeah, I have some sort of discipline problem when I'm assigned too many jobs I care very little about :lol: :lol: :roll:

Wait, that sounds like...

Shibotto wrote:
And that's precisely why I found another one :twisted:

I knew it. :o

Shibotto wrote:
A very quick look and an even simpler fix [...]

This issue has been corrected.

As always, thanks for the feedback.

Spargeltarzan wrote:
My only (little) issue what I have (still) is when I login my gnome keyring does not become autounlocked. So when I boot my PC firstly I login, then nextcloud wants a password and finally gnome-keyring...

This issue sounds vaguely familiar.

I haven't tested it in a long time but last I checked the steps on ArchWiki were correct. The caveat was that it would only work for a keyring named "Login".
_________________
Dantrell B.
Back to top
View user's profile Send private message
saboya
Guru
Guru


Joined: 28 Nov 2006
Posts: 448
Location: Brazil

PostPosted: Wed Dec 26, 2018 4:27 pm    Post subject: Reply with quote

Spargeltarzan wrote:
My only (little) issue what I have (still) is when I login my gnome keyring does not become autounlocked. So when I boot my PC firstly I login, then nextcloud wants a password and finally gnome-keyring...

dantrell wrote:
This issue sounds vaguely familiar.


I had this issue before, and had it again with my most recent install. My guess is that the issue is pam_gnome_keyring.so is loaded twice by pam if you use GDM, because it's present in /etc/pam.d/system-login and /etc/pam.d/gdm-password. Removing this line

Code:
session                optional        pam_gnome_keyring.so auto_start


from /etc/pam.d/system-login fixed the issue for me.
Back to top
View user's profile Send private message
simonvanderveldt
Tux's lil' helper
Tux's lil' helper


Joined: 26 Jan 2016
Posts: 105

PostPosted: Fri Dec 28, 2018 8:30 pm    Post subject: Reply with quote

Just ran into a compilation error with media-sound/mixxx-2.2.0 which tries to use up_client_get_devices2 from UPower when version UPower version 0.99.8 is installed (Bug filed here https://bugs.gentoo.org/show_bug.cgi?id=673890).
I have 0.99.8.20171128-r2::dantrell-gnome installed, but it doesn't seem to provide that function because the ebuild for UPower still uses the old 0.99.3 sources.
There is a patch for it in the dentrell-gnome overlay (https://github.com/dantrell/gentoo-overlay-dantrell-gnome/blob/master/sys-power/upower/files/upower-0.99.8-0015-lib-add-a-new-version-of-up-client-get-devices-which.patch) but that isn't applied in the ebuild for 0.99.8 (see the patches applied here https://github.com/dantrell/gentoo-overlay-dantrell-gnome/blob/master/sys-power/upower/upower-0.99.8.20171128-r2.ebuild#L68).

@dantrell is there any reason you're not using the releases from UPower's gitlab (https://gitlab.freedesktop.org/upower/upower/tags)?
That's also the URL that the ebuilds in the gentoo tree are using, see https://github.com/gentoo/gentoo/blob/master/sys-power/upower/upower-0.99.8-r1.ebuild#L13
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Fri Dec 28, 2018 10:47 pm    Post subject: Reply with quote

simonvanderveldt wrote:
Just ran into a compilation error with media-sound/mixxx-2.2.0 which tries to use up_client_get_devices2 from UPower when version UPower version 0.99.8 is installed [...]

Use 0.99.9.20181025 which is slotted in ~.

P.S. Please don't file bugs with Gentoo if you didn't reproduce the same issue with their ebuilds.

simonvanderveldt wrote:
@dantrell is there any reason you're not using the releases from [...]

You forgot that I support multiple GNOME release versions and implementation (e.g GNOME 3.14 with ConsoleKit).
_________________
Dantrell B.
Back to top
View user's profile Send private message
simonvanderveldt
Tux's lil' helper
Tux's lil' helper


Joined: 26 Jan 2016
Posts: 105

PostPosted: Fri Dec 28, 2018 11:01 pm    Post subject: Reply with quote

dantrell wrote:
simonvanderveldt wrote:
Just ran into a compilation error with media-sound/mixxx-2.2.0 which tries to use up_client_get_devices2 from UPower when version UPower version 0.99.8 is installed [...]

Use 0.99.9.20181025 which is slotted in ~.

P.S. Please don't file bugs with Gentoo if you didn't reproduce the same issue with their ebuilds.

OK, I'll give that a try. Though tbh I do think it's not entirely correct to have to use a newer version since up_client_get_devices2 was added in 0.99.8.

Andyeah, stupid mistake on my end, I missed that it was because of a dep on UPower :x
Still found an issue with the ebuild, so at least we managed to pull an enhancement from it :)

dantrell wrote:
simonvanderveldt wrote:
@dantrell is there any reason you're not using the releases from [...]

You forgot that I support multiple GNOME release versions and implementation (e.g GNOME 3.14 with ConsoleKit).

Ohw, so it's because the same version/ebuild is being used by different versions of GNOME that you support?
Wouldn't it be better to use the sources matching the version (i.e. 0.99.8 uses the 0.99.8 sources instead of 0.99.3) and apply patches to that though?
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Sat Dec 29, 2018 12:15 am    Post subject: Reply with quote

simonvanderveldt wrote:
OK, I'll give that a try. Though tbh I do think it's not entirely correct to have to use a newer version since up_client_get_devices2 was added in 0.99.8.

Andyeah, stupid mistake on my end, I missed that it was because of a dep on UPower :x
Still found an issue with the ebuild, so at least we managed to pull an enhancement from it :)?

You are misinterpreting the version numbers. They are git releases and not tarball releases. 0.99.8.20171128 doesn't have what you were looking for because 0.99.8 wasn't tarballed until 2018-06-19.

In short, there wasn't an issue with the ebuild. It was appropriately named, versioned and slotted in ~ (where, for the record, media-sound/mixxx-2.2.0 is also slotted).

simonvanderveldt wrote:
Ohw, so it's because the same version/ebuild is being used by different versions of GNOME that you support?

Yep.

simonvanderveldt wrote:
Wouldn't it be better to use the sources matching the version (i.e. 0.99.8 uses the 0.99.8 sources instead of 0.99.3) and apply patches to that though?

"Better" is subjective. I think the word you are looking for is "cleaner".

I used that method that around 0.99.1 until a significant amount of code got ported to GDBus. Since then it's simpler to use my current method.

Granted, I could forward-port the deprecated code but I consider that a very low priority issue.
_________________
Dantrell B.
Back to top
View user's profile Send private message
simonvanderveldt
Tux's lil' helper
Tux's lil' helper


Joined: 26 Jan 2016
Posts: 105

PostPosted: Sat Dec 29, 2018 1:45 pm    Post subject: Reply with quote

dantrell wrote:
simonvanderveldt wrote:
OK, I'll give that a try. Though tbh I do think it's not entirely correct to have to use a newer version since up_client_get_devices2 was added in 0.99.8.

Andyeah, stupid mistake on my end, I missed that it was because of a dep on UPower :x
Still found an issue with the ebuild, so at least we managed to pull an enhancement from it :)?

You are misinterpreting the version numbers. They are git releases and not tarball releases. 0.99.8.20171128 doesn't have what you were looking for because 0.99.8 wasn't tarballed until 2018-06-19.

In short, there wasn't an issue with the ebuild. It was appropriately named, versioned and slotted in ~ (where, for the record, media-sound/mixxx-2.2.0 is also slotted).

Are you sure? I only see a SRC_URI linking to an old release tarball https://github.com/dantrell/gentoo-overlay-dantrell-gnome/blob/695c664a4839b8011d2819acc2eaa8e1cb025859/sys-power/upower/upower-0.99.8.20171128-r2.ebuild#L10. Nothing related to git in there.
Apart from that shouldn't the version then be 0.99.3-<date>, like the ebuild writing manual (https://devmanual.gentoo.org/ebuild-writing/file-format/index.html) states?
Quote:
When packaging a snapshot of a source repository, there are two commonly used formats. The first treats the snapshot as a patch to the previous version, and so the ebuild version is in the format $(last-released-version)_pYYYYMMDD. Alternatively, the snapshot may be treated as a pre-release to an upcoming version, usually used when a release is anticipated but not out yet. The format for this is $(upcoming-version)_preYYYYMMDD.


dantrell wrote:
simonvanderveldt wrote:
Wouldn't it be better to use the sources matching the version (i.e. 0.99.8 uses the 0.99.8 sources instead of 0.99.3) and apply patches to that though?

"Better" is subjective. I think the word you are looking for is "cleaner".


Tbh whilst I do think it's definitely better to use sources matching the version we're building, I also understand the challenges with all the upstream changes and backwards compatbility.
In the end how we get to the correct sources, while relevant, doesn't matter as much as making sure the library we get when building is correct, for upower-0.99.8 that's not the case since it doesn't support the functions it's supposed to support for it's version. In other words you're not getting version 0.99.8 when it's clearly named/versioned as 0.99.8. This is obviously undesirable.

dantrell wrote:
I used that method that around 0.99.1 until a significant amount of code got ported to GDBus. Since then it's simpler to use my current method.

Granted, I could forward-port the deprecated code but I consider that a very low priority issue.

Hmm, that's kinda annoying. So the changes have become so large that applying simple patches to them won't work and it's easier to cherry-pick based on older versions? That sounds scary.
Anyway, the patch for up_client_get_devices2 seems to be backwards compatible. Can't it just be added to the 0.99.8 ebuild? Or otherwise version that ebuild as something lower than 0.99.8?
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Sat Dec 29, 2018 9:20 pm    Post subject: Reply with quote

simonvanderveldt wrote:
In other words you're not getting version 0.99.8 when it's clearly named/versioned as 0.99.8. This is obviously undesirable.

It's named 0.99.8 because you are getting 0.99.8. As stated previously, you are misinterpreting the version numbers.

0.99.8.20171128 follows up to commit cc37137 (the last commit on 2017-11-28 ). If you review commit itself, you'll note that a post release version bump to 0.99.8 took place.

Why that particular commit? Because the process is semi-automated and at the time 0.99.8.20171128 was added (i.e. before 2018-03-02), it was the most recent commit.

simonvanderveldt wrote:
Anyway, the patch for up_client_get_devices2 seems to be backwards compatible. Can't it just be added to the 0.99.8 ebuild?

UPower is on a forward only rolling release cycle on my end.

If you want the release tarballs (and you aren't using the ConsoleKit implementation), switch to Gentoo's version.
_________________
Dantrell B.
Back to top
View user's profile Send private message
simonvanderveldt
Tux's lil' helper
Tux's lil' helper


Joined: 26 Jan 2016
Posts: 105

PostPosted: Sat Dec 29, 2018 10:31 pm    Post subject: Reply with quote

dantrell wrote:
simonvanderveldt wrote:
In other words you're not getting version 0.99.8 when it's clearly named/versioned as 0.99.8. This is obviously undesirable.

It's named 0.99.8 because you are getting 0.99.8. As stated previously, you are misinterpreting the version numbers.

0.99.8.20171128 follows up to commit cc37137 (the last commit on 2017-11-28 ). If you review commit itself, you'll note that a post release version bump to 0.99.8 took place.

Why that particular commit? Because the process is semi-automated and at the time 0.99.8.20171128 was added (i.e. before 2018-03-02), it was the most recent commit.

Ah, now I get where the confusion is coming from.
That RELEASE file isn't the release notes, it's just some info for the maintainers what to do when they do a/the next release, so it/the commit you linked just updates those notes to include the next version that is going to be released (i.e. 0.99.8 in this case). The commit message isn't very clear about this unfortunately.
See the NEWS file (https://cgit.freedesktop.org/upower/tree/NEWS?id=cc37137637a7bf1e1b48ec328ae13e47b63aaa88 for the same commit you linked to) for the actual release info showing "Version 0.99.7 from 2017-11-28" or check the git log for it https://cgit.freedesktop.org/upower/log/NEWS.

dantrell wrote:
simonvanderveldt wrote:
Anyway, the patch for up_client_get_devices2 seems to be backwards compatible. Can't it just be added to the 0.99.8 ebuild?

UPower is on a forward only rolling release cycle on my end.

If you want the release tarballs (and you aren't using the ConsoleKit implementation), switch to Gentoo's version.

OK, cool. So gentoo's UPower should work as well, the only/main difference between that and the one in your overlay is the ConsoleKit support?
Back to top
View user's profile Send private message
dantrell
Guru
Guru


Joined: 01 Jun 2007
Posts: 580
Location: Earth

PostPosted: Sat Dec 29, 2018 11:15 pm    Post subject: Reply with quote

simonvanderveldt wrote:
Ah, now I get where the confusion is coming from. [...]

I think you are still confused.

As I said, the process is semi-automated. Here's how it works in pseudo commands:
  • "git pull" is performed
  • "configure.ac" is checked for the version (i.e. commit cc37137)
  • "git log" is checked for the date
If these steps were followed after 2017-11-28 but before 2018-03-02, what version.date combination do you get?

0.99.8.20171128.

simonvanderveldt wrote:
OK, cool. So gentoo's UPower should work as well, the only/main difference between that and the one in your overlay is the ConsoleKit support?

It's the main difference, yes, but not the only difference.

It should still work but I neither test nor support it.
_________________
Dantrell B.
Back to top
View user's profile Send private message
simonvanderveldt
Tux's lil' helper
Tux's lil' helper


Joined: 26 Jan 2016
Posts: 105

PostPosted: Thu Jan 03, 2019 10:09 pm    Post subject: Reply with quote

dantrell wrote:
simonvanderveldt wrote:
Ah, now I get where the confusion is coming from. [...]

I think you are still confused.

As I said, the process is semi-automated. Here's how it works in pseudo commands:
  • "git pull" is performed
  • "configure.ac" is checked for the version (i.e. commit cc37137)
  • "git log" is checked for the date
If these steps were followed after 2017-11-28 but before 2018-03-02, what version.date combination do you get?

0.99.8.20171128.

While I understand that this is technically what causes this version to be shown I don't think it's correct to release a package from the the first commit of what's going to be the next version versioned with that next version number, that will cause issues since it obviously doesn't contain what that version number is going to deliver.
Exaggerated example to highligh the problem: It's the same as when I'm writing some software for which I only want to release a 1.0.0, so I never make a release until that 1.0.0 is ready and on commit 1 I don't do anything else than add the build tooling which defines the version as 1.0.0 and then that single commit gets packaged as 1.0.0. That would obviously not make sense.
I have no clue why the UPower devs don't update the build info when they are actually cutting the release, but alas.

Anyway, if using the releases is too problematic I guess we can't really do much about it apart from using the correct versioning in the ebuild (i.e. in this case it should be 0.99.8_pre20171128 and potentially the build scripts could be patched so they are no longer defining the incorrect version or make sure that it offers all functionality that 0.99.8 is supposed to offer.
Not sure if you want to do that though, so I guess I'll just unmask 0.99.9 and hope it all works ;)
[edit] Scratch that, just noticed you already unmasked it :)

Now for something totally different, I'm a bit late with this, but just tried to update webkit-gtk to 2.22.5 and it gave me the following error:
Code:
In file included from /var/tmp/portage/net-libs/webkit-gtk-2.22.5/work/webkitgtk-2.22.5/Source/WebCore/platform/glib/EventHandlerGLib.cpp:38:0:
/var/tmp/portage/net-libs/webkit-gtk-2.22.5/work/webkitgtk-2.22.5/Source/WebCore/platform/PlatformWheelEvent.h:216:60: error: no 'bool WebCore::PlatformWheelEvent::isEndOfNonMomentumScroll() const' member function declared in class 'WebCore::PlatformWheelEvent'
 inline bool PlatformWheelEvent::isEndOfNonMomentumScroll() const
                                                            ^~~~~
/var/tmp/portage/net-libs/webkit-gtk-2.22.5/work/webkitgtk-2.22.5/Source/WebCore/platform/PlatformWheelEvent.h:221:67: error: no 'bool WebCore::PlatformWheelEvent::isTransitioningToMomentumScroll() const' member function declared in class 'WebCore::PlatformWheelEvent'
 inline bool PlatformWheelEvent::isTransitioningToMomentumScroll() const


In case it's relevant these are my USE flags for it:
Code:
net-libs/webkit-gtk -egl -geolocation -opengl -webgl -jit -libnotify
Back to top
View user's profile Send private message
Display posts from previous:   
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Unsupported Software All times are GMT
Goto page Previous  1, 2, 3 ... 10, 11, 12 ... 14, 15, 16  Next
Page 11 of 16

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum