Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
SystemD free system/gentoo
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3  
Reply to topic    Gentoo Forums Forum Index Gentoo Chat
View previous topic :: View next topic  
Author Message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Wed Jul 25, 2018 2:35 pm    Post subject: Reply with quote

Not it's still a fault :)
You should use a virtual/udev::local that is hack to accept eudev-1.10 as valid choice in its RDEPEND
Code:
        !systemd? ( || ( >=sys-fs/eudev-2.1.1 >=sys-fs/udev-217 ) )

And portage wouldn't have to try anything else but accept eudev-1.10 is good.
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Wed Jul 25, 2018 3:08 pm    Post subject: Reply with quote

I've masked anything greater than virtual/udev-215 which happily accepts anything greater than eudev 1.3, but all these reside in local. ;)

But krinn is correct, if needed I've hacked ebuilds before because of (r)depend constraints that devs thought were needed but really weren't.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2280
Location: Adendorf, Germany

PostPosted: Wed Jul 25, 2018 3:32 pm    Post subject: Reply with quote

The lowest version of eudev in the tree is 3.1.5. Why would someone use eudev-1.10 then?

If there was a good reason, it could have been preserved. It wasn't. You want something not in the tree any more -> your responsibility. No question. But I am curious, why that ancient eudev? What is wrong with the newer?

Edit: I just had a look at the 1.10 release: "Aug 22, 2014, 614 commits to master since this tag" That's quite some change I daresay...
_________________
Important German:
  1. "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
  2. "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Wed Jul 25, 2018 3:47 pm    Post subject: Reply with quote

krinn wrote:
Not it's still a fault :)
You should use a virtual/udev::local that is hack to accept eudev-1.10 as valid choice in its RDEPEND
[
Good idea. If i ever get an alternate, I can just change my local ebuild. Actually, mdev works fine but some other packages want the hardware data base that udev generates.

As I think about such a package, let's call it smalldev, would hard depend on mdev and have a little code to run after boot and hot plugging to generate the table. It wouldn't have all those udev rules, but I don't use them anyway.

However, I'm getting far from the main thread.
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Wed Jul 25, 2018 3:48 pm    Post subject: Reply with quote

Yamakuzure wrote:
The lowest version of eudev in the tree is 3.1.5. Why would someone use eudev-1.10 then?


Um, because it provides everything that is needed by my system and works perfectly well with all the software on my system.

I don't throw away old things just because they aren't being created/updated any more.
I still have systems running XP and an old version of win98SE (because of one old game :) )

I don't know that there is anything wrong with newer eudev as I haven't kept up with the development and what changes it brings to the system.

So bottom line, I suppose it's a personal choice.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Wed Jul 25, 2018 4:05 pm    Post subject: Reply with quote

Yamakuzure wrote:
The lowest version of eudev in the tree is 3.1.5. Why would someone use eudev-1.10 then?

I masked everything greater than 1.99 but didn't make a note why. Was there a 2.x series with problems? The only thing I see in the ebuild is
Code:

        ewarn
        ewarn "As of 2013-01-29, ${P} provides the new interface renaming functionality,"
        ewarn "as described in the URL below:"
        ewarn "https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames"
        ewarn
        ewarn "This functionality is enabled BY DEFAULT because eudev has no means of synchronizing"
        ewarn "between the default or user-modified choice of sys-fs/udev.  If you wish to disable"
        ewarn "this new iface naming, please be sure that /etc/udev/rules.d/80-net-name-slot.rules"
        ewarn "exists: touch /etc/udev/rules.d/80-net-name-slot.rules"
        ewarn
And i take care of that with " net.ifnames=0" on the kernel command line anyway.
I do use those rules to rename eth0 and eth1 to lan0 and wan0 on my router box. I run 1.10 there also so apparently the capability was there already.
So i really can't answer your question. Can you answer "Why would someone use anything later than eudev-1.10 then?" ?
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2280
Location: Adendorf, Germany

PostPosted: Thu Jul 26, 2018 7:41 am    Post subject: Reply with quote

So, if I understand you two correctly, you both brought the virtual/udev annoyance upon yourselves without any benefit in return, right?

I must admit that I am a bit disappointed now. I thought I'd be given some insight about problems with eudev 2+ that I wasn't aware of...

But of course the reason "personal choice" is always sound and valid, no matter who else might find it irresponsible, unlogical or whatever. ;-)

Tony0945 wrote:
Can you answer "Why would someone use anything later than eudev-1.10 then?" ?

First reason would be bugfixes. There certainly have been issues fixed in the meantime.
Second reason would be to not have such annoyances with dependency hell I had to manually fix, keep and maintain in a local overlay without getting anything in return.

If there is nothing wrong with the newer version, I see no reason to keep the old one forcefully installed with such a high cost.
But as stated above, personal choice always matters more. So if you two trust your choice back then, even if you can remember the reasoning, and are willing to "keep the pieces", then I say: "Why not?" :-)
_________________
Important German:
  1. "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
  2. "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Thu Jul 26, 2018 12:55 pm    Post subject: Reply with quote

Yamakuzure wrote:
First reason would be bugfixes. There certainly have been issues fixed in the meantime.

It would be nice to see these in the package notes instead of just "version bump". At least a link to upstream revision history.

This is one reason why I locked openrc. There were feature updates that I didn't like. Like automatically loading all modules. Why bother building something as a module if it's always going to be loaded? And tmpfiles. Adding systemd locations. The init system is important. It's not just another application where unexpected quirks can be tolerated. Similarly with bash that I've locked because an unannounced change broke my scripts. I did keep that version after reading a forum discussion on how to fix it. Then I locked it. Too important for surprises.

So really there are more than two reasons to change a packages code:

1. New capability
2. Bugfixes
3. Upstream developer just decided to recode everything his way.
4. Upstream developer decided to change the API.
5. To make it more compatible with systemd because "wave of the future"
Upstream meaning Gentoo dev for gentoo projects.

#2 is always valid.
#1 is valid if the new capability is wanted or not harmful or optional.
#3 and #4 are rarely valid and should really be a forked package rather than just a version change.
#5 for me is never valid, YMMV

I could take it on faith, but past experience has taught me not to. I don't want to name names and get into a flame war, but their are three gentoo devs whose motives and skill I don't trust at all. There are two more that I don't think have bad motives but have made too many mistakes in the past through either inattention or poor process that I always wait a month to see if someone is going to report a problem in the forums, but I think their hearts are pure. The rest of the devs I think are top notch and I'm glad we have them.

EDIT:
So, again, "What changed between eudev 1 and eudev 3 and why?" I would think that our eudev devs were following the upstream udev changes. I trust RedHat as much as (or even less than) Microsoft, so why should I blindly accept their changes?
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Thu Jul 26, 2018 1:33 pm    Post subject: Reply with quote

Yamakuzure wrote:
So, if I understand you two correctly, you both brought the virtual/udev annoyance upon yourselves without any benefit in return, right?


Annoyance? I never said I was annoyed with any of it. virtual/udev-215 is still in the gentoo repository, true I did copy it to my local and have it masked past that point, but I routinely do that with lots of packages, and over time it's kept me from starting threads asking why xyz broke when stuff got promoted to stable that shouldn't have been. *shrugs*


As for the rest of your smarmy post ... whatever :roll:

Edit to add: Not even once have I had a conflict that I've had to resolve because of locking either openrc or eudev.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Thu Jul 26, 2018 2:43 pm    Post subject: Reply with quote

Tony0945 wrote:
This is one reason why I locked openrc. There were feature updates that I didn't like. Like automatically loading all modules.

INSTALL_MASK="/usr/lib/modules-load.d" ;)


Anon-E-moose wrote:
I did copy it to my local and have it masked past that point

Code:
equery d virtual/udev
dev-libs/libinput-1.11.1 (virtual/udev)
media-video/vlc-3.0.3-r1 (udev ? virtual/udev:0)
sys-apps/hwids-20180518 (udev ? virtual/udev)
sys-auth/consolekit-1.2.1 (acl ? >=virtual/udev-200)
sys-block/gparted-0.29.0 (hfs ? virtual/udev)
sys-fs/udev-init-scripts-32 (>=virtual/udev-217)
sys-fs/udisks-2.7.6 (virtual/udev)
sys-libs/libblockdev-2.18-r1 (lvm ? virtual/udev)
sys-power/upower-0.99.7 (kernel_linux ? virtual/udev)
virtual/dev-manager-0-r1 (virtual/udev)

you see, they depends on virtual/udev only, or depends on a version but always >=
following JRG's tips for guys with a brain :)
you could name it virtual/udev-9999::local making the mask effective while not having any mask set ; which will remove the need to update it each time a new virtual/udev-version-higher-than-yours enter the tree.
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Thu Jul 26, 2018 2:58 pm    Post subject: Reply with quote

I could rename it to -9999, but I haven't really had a need to.
It's been a long time since I've had to modify package.mask for that particular reason. *shrugs*
I find it's just not that big a deal.

Re. your list of those needing virtual/udev, I have 2 packages out of it installed on my system.

But thanks for the link, if I need to go the 9999 route, I'll certainly do so.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
Hu
Moderator
Moderator


Joined: 06 Mar 2007
Posts: 21489

PostPosted: Fri Jul 27, 2018 1:20 am    Post subject: Reply with quote

Tony0945 wrote:
Similarly with bash that I've locked because an unannounced change broke my scripts.
What change was this? What broke?
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Fri Jul 27, 2018 1:58 am    Post subject: Reply with quote

Hu wrote:
Tony0945 wrote:
Similarly with bash that I've locked because an unannounced change broke my scripts.
What change was this? What broke?

It's somewhere in the forums. Last year? two years ago? Some default. The fix was to edit it back.
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2280
Location: Adendorf, Germany

PostPosted: Fri Jul 27, 2018 7:55 am    Post subject: Reply with quote

Tony0945 wrote:
Hu wrote:
Tony0945 wrote:
Similarly with bash that I've locked because an unannounced change broke my scripts.
What change was this? What broke?

It's somewhere in the forums. Last year? two years ago? Some default. The fix was to edit it back.
...or to fix your scripts to not depend on uncertain defaults?
_________________
Important German:
  1. "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
  2. "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
Back to top
View user's profile Send private message
krinn
Watchman
Watchman


Joined: 02 May 2003
Posts: 7470

PostPosted: Fri Jul 27, 2018 11:18 am    Post subject: Reply with quote

could be time to quickpkg the packages, release the mask, test new version and switch back to your "stuck" version if you don't like what you get no?
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Fri Jul 27, 2018 12:17 pm    Post subject: Reply with quote

Yamakuzure wrote:
Tony0945 wrote:
Hu wrote:
Tony0945 wrote:
Similarly with bash that I've locked because an unannounced change broke my scripts.
What change was this? What broke?

It's somewhere in the forums. Last year? two years ago? Some default. The fix was to edit it back.
...or to fix your scripts to not depend on uncertain defaults?


I'm wondering - Are you trying to be unpleasant?
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Fri Jul 27, 2018 1:22 pm    Post subject: Reply with quote

I blocked bash-completion and bash for a while because the latest "stable" version did some weirdness on my system,
don't remember the details as it was a year or two back.

Just because something is released stable doesn't mean it can't/won't cause problems.

One of the font packages, for several releases, has given me fuzzy looking fonts on both the console and urxvt,
and the fonts that I use aren't from that package.
I never have figured out why, as the easy answer was just to block at a particular version of that package.
I even googled when it first showed and some people on other distros were having the same problem.

Yes, I could fix it by not having the fonts on my system, but the older worked fine, so I went with that option.

As far as less than sanguine advice like "fix your scripts to not depend on uncertain defaults"
I consider that very boorish trolling, maybe it's something they're putting in the water in that country. *shrugs*

Yes, gentoo is all about choice even when a few think that others shouldn't exercise personal choice because ... well, they know better than the end user on their own system :roll:
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
berferd
Tux's lil' helper
Tux's lil' helper


Joined: 13 May 2004
Posts: 117

PostPosted: Fri Jul 27, 2018 2:19 pm    Post subject: Reply with quote

Yamakuzure wrote:
...or to fix your scripts to not depend on uncertain defaults?


How hard could it be? Tony couldn't have written that many lines of bash, he's not a Gentoo dev!

And besides he's surely got infinite time to deal with the latest whim of the Gentoo developer playpen.

EDIT: I'm savoring the delicious irony of you accusing people of spoiling for a fight.
Back to top
View user's profile Send private message
tld
Veteran
Veteran


Joined: 09 Dec 2003
Posts: 1811

PostPosted: Fri Jul 27, 2018 2:36 pm    Post subject: Reply with quote

Anon-E-moose wrote:
I blocked bash-completion and bash for a while because the latest "stable" version did some weirdness on my system,
don't remember the details as it was a year or two back.
While may have been longer ago than that, I recall some really awful changes to bash tab completion regarding variables, for example if you had a variable set to a path. At one point it hitting tab with a variable in your command started escaping the $ with \. They fixed that, however I believe that at one point the tab would expand such a variable, and that behavior never came back.

Tom
Back to top
View user's profile Send private message
Ant P.
Watchman
Watchman


Joined: 18 Apr 2009
Posts: 6920

PostPosted: Fri Jul 27, 2018 4:21 pm    Post subject: Reply with quote

I vaguely remember bash-completion had a really bad habit of screwing up input and forcing me to run `stty sane` every few minutes, but I haven't seen that happen for months now.
Back to top
View user's profile Send private message
Hu
Moderator
Moderator


Joined: 06 Mar 2007
Posts: 21489

PostPosted: Sat Jul 28, 2018 12:10 am    Post subject: Reply with quote

I asked my question because I was hoping for details on specific broken versions, what bad behaviors those had, and so on - not to start a fight over whether any particular change was good or whether it was proper for users to depend on old defaults. I too recall the annoying terminal-goes-to-no-echo mode problem, and I hadn't thought about it until Ant P.'s post, but it seems like it hasn't hit me in a while, after a long period where it was quite annoying. Also, like tld mentions, the newer rules for tab interaction with variables are worse than the older ones.
Back to top
View user's profile Send private message
Tony0945
Watchman
Watchman


Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA

PostPosted: Sat Jul 28, 2018 1:21 am    Post subject: Reply with quote

Hu, I have to go away for the weekend, but will research it on my return.
Back to top
View user's profile Send private message
Anon-E-moose
Watchman
Watchman


Joined: 23 May 2008
Posts: 6095
Location: Dallas area

PostPosted: Sat Jul 28, 2018 8:13 pm    Post subject: Reply with quote

Speaking of bash and bash completion :lol:

If I run something like this "tar xfC /usr/portage/distfiles/gtk+-3.22.19.tar.xz /tmp"

then I do something like "less /tmp/gtk+-3.22.19/*"

I get this on stderr

sh: _gtar: line 60: syntax error near unexpected token `('
sh: _gtar: line 60: ` --file | -f | -!(-*)f)'
sh: error importing function definition for `_gtar'
sh: __tar_detect_ext: line 7: syntax error near unexpected token `('
sh: __tar_detect_ext: line 7: ` ?(-)*[cr]*f)'
sh: error importing function definition for `__tar_detect_ext'
/bin/bash: _gtar: line 60: syntax error near unexpected token `('
/bin/bash: _gtar: line 60: ` --file | -f | -!(-*)f)'
/bin/bash: error importing function definition for `_gtar'
/bin/bash: __tar_detect_ext: line 7: syntax error near unexpected token `('
/bin/bash: __tar_detect_ext: line 7: ` ?(-)*[cr]*f)'
/bin/bash: error importing function definition for `__tar_detect_ext'
/bin/bash: _gtar: line 60: syntax error near unexpected token `('
/bin/bash: _gtar: line 60: ` --file | -f | -!(-*)f)'
/bin/bash: error importing function definition for `_gtar'
/bin/bash: __tar_detect_ext: line 7: syntax error near unexpected token `('
/bin/bash: __tar_detect_ext: line 7: ` ?(-)*[cr]*f)'
/bin/bash: error importing function definition for `__tar_detect_ext'


and if I do "set |less" I will find lots of functions _gtar, __tar, etc until I log out (if console) or kill urxvt (if in an X session)

This has been going on for a while, and I'm curious if anyone else has noticed it.

I've updated both bash and bash-completion to the latest stable release.
It's not a show stopper, just a major annoyance to log out or restart a shell to not see the messages.


Edit to add: if I pull the tar command from history OR type it out completely without using tab-completion I don't get the weird behavior.
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2280
Location: Adendorf, Germany

PostPosted: Mon Jul 30, 2018 10:35 am    Post subject: Reply with quote

Tony0945 wrote:
Yamakuzure wrote:
...or to fix your scripts to not depend on uncertain defaults?

I'm wondering - Are you trying to be unpleasant?

berferd wrote:
EDIT: I'm savoring the delicious irony of you accusing people of spoiling for a fight.
No. I am sorry if this sounded too absolute.
_________________
Important German:
  1. "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
  2. "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Gentoo Chat All times are GMT
Goto page Previous  1, 2, 3
Page 3 of 3

 
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