Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
kdbus in the kernel
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3 ... 9, 10, 11 ... 25, 26, 27  Next  
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sun Mar 15, 2015 5:24 pm    Post subject: Reply with quote

@tclover Glad we're past that. :-)

supervision-scripts looks interesting. Clean sh, which makes a change.

Only had one thing to tweak, but forgot what it was now, and off out. No bugs :-)

Glancing again, I notice a couple of portability issues:
Code:
[ -n "${1}" -a -n "${2}" -a -d "${CG_ROOT}/${1}" ] || return 0
local ctrl="${1}" group head=$(cgroup_find_path "${1}")
group="${CG_ROOT}/${1}${head}${CG_NAME}_${SVC_NAME}"

POSIX recommends against the -a form, and it's obsoleted. (cf: Application Usage.)
Also, you shouldn't assign in 'local' for portability, especially with multiple names (last one came up on dash, iirc.)
So, it's much easier just to treat it like C89, and forward declare, as you then have the field splitting defined not to happen for the word following the '='. So splitting it out, I'd do this:
Code:
[ -n "$1" ] && [ -n "$2" ] || return 0
[ -d "$CG_ROOT/$1" ] || return 0
local ctrl group head
head=$(cgroup_find_path "$1") && ctrl=$1 || return 0 # or use an if
group=$CG_ROOT/$1$head${CG_NAME}_$SVC_NAME

I tend to split out parameter-checking to keep it obvious by inspection. Not sure what you're doing with ctrl.

Anyhow, I'll clone it now, and take a look at the rest in more-depth soon.

Nice work.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Sun Mar 15, 2015 7:59 pm    Post subject: Reply with quote

@mods: Sorry for another off topic ride ;-) but I got to answer to this with a short answer first.
@steveL: *Fake* issues serving as discussion threads could do the trick. A short POSIX off topic won't hurt this thread after many, certainly.

First, I was aware of the 'local' POSIX compatibility issue. It's just... convenient to be able use the same relevant variable name in different *local* scope to keep consistency/efficiency on variable naming scheme. Without 'local' keyword, naming *local* scope variable is problematic and could lost possible readers. Unset can be used to unset *local* scope variables but not masking *global* ones--this is the issue here and could lead to confusion.

And after scripting a great deal for BusyBox/ash,--which has not a proper POSIX compliency,--I just tend to _abuse_ the use of 'local' for consistency sake.

steveL wrote:
Glancing again, I notice a couple of portability issues:
Code:
[ -n "${1}" -a -n "${2}" -a -d "${CG_ROOT}/${1}" ] || return 0
local ctrl="${1}" group head=$(cgroup_find_path "${1}")
group="${CG_ROOT}/${1}${head}${CG_NAME}_${SVC_NAME}"

POSIX recommends against the -a form, and it's obsoleted. (cf: Application Usage.)
Also, you shouldn't assign in 'local' for portability, especially with multiple names (last one came up on dash, iirc.)
So, it's much easier just to treat it like C89, and forward declare, as you then have the field splitting defined not to happen for the word following the '='. So splitting it out, I'd do this:
Code:
[ -n "$1" ] && [ -n "$2" ] || return 0
[ -d "$CG_ROOT/$1" ] || return 0
local ctrl group head
head=$(cgroup_find_path "$1") && ctrl=$1 || return 0 # or use an if
group=$CG_ROOT/$1$head${CG_NAME}_$SVC_NAME

I tend to split out parameter-checking to keep it obvious by inspection. Not sure what you're doing with ctrl.


Second, assigning while declaring variable makes the code concise but... the lack of a comma operator could bring in error prone declarations. I do take some care when doing so, however, you're right that I should refrain from it for compatibility reasons--with declaration first, and the assignment. I should take a look (on POSIX doc) for 'declare' counterpart because many shell has this builtin.

Regarding the '-a' (AND) and '-o' (OR), the related section uses some examples like `test "$1" -a "$2"' with '$1' asigned to '!' and '$2' to NULL. Such a nice example! I usually do not use that syntax but only with bash! for obvious reasons when relying on an unknown sh[ell] (POSIX compliant or not.) I always use test operand like '-n', '-z' ... for that making the pointed issue almost irrelevant. Sometimes I _do_ use the syntax you're highlighting on the previous example but it become heavy in scripts using that syntax quite often. So, a preference goes to '-a'/'-o' on plain sh with the previous cautious usage.

NOTE: I prefer by far using '(( $+var ))'/'[[ $+var ]]' when scripting for zsh.

So, the main issue is rather declaration & assignment because my guess is that many modern shells support '-[ao]' for test bultin. I have to think again for the 'local' usage--it won't be that easy to get concise/breve/efficient code otherwise.

But lets discuss this on the issue tracker to not go too far off here.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 4213
Location: Dallas area

PostPosted: Sun Mar 15, 2015 9:05 pm    Post subject: Reply with quote

Seems like GKH is trying to earn his money from RH

Quote:
Add kdbus branch to linux-next?

Hi Stephen,

Can you add the:
https://git.kernel.org/cgit/linux/kernel/git/gregkh/char-misc.git/ kdbus

tree to linux-next so that it gets the normal amount of testing that
linux-next provides?

The code has been reworked and reviewed many times, and this last round
seems to have no objections, so I'm queueing it up to be merged for
4.1-rc1.

thanks,

greg k-h


What a whore!!!

Edit to add:
A whore being someone that sells part of themselves for money,
and IMO Greg has sold his soul for "filthy lucre"
_________________
Asus m5a99fx, FX 8320 - nouveau, oss4, rx550 for qemu passthrough
Acer laptop E5-575, i3-7100u - i965, alsa
---both---
5.0.13 zen kernel, profile 17.1 (no-pie & modified) amd64-no-multilib
gcc 8.2.0, eudev, openrc, openbox, palemoon
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sun Mar 15, 2015 10:40 pm    Post subject: Reply with quote

tclover wrote:
A short POSIX off topic won't hurt this thread after many, certainly.

Indeed; it's as on-topic as anything, and this is a chat forum. I don't tend to use web-interfaces these days; I used to love them, but everything's all 5.0-rc_beta1-alpha01-jboss-jquery-argle now.
Quote:
First, I was aware of the 'local' POSIX compatibility issue. It's just... convenient to be able use the same relevant variable name in different *local* scope to keep consistency/efficiency on variable naming scheme. Without 'local' keyword, naming *local* scope variable is problematic and could lost possible readers. Unset can be used to unset *local* scope variables but not masking *global* ones--this is the issue here and could lead to confusion.

You've misunderstood; I never said "don't use local". I just said "use it portably."
Quote:
Second, assigning while declaring variable makes the code concise but... the lack of a comma operator could bring in error prone declarations.

Only if you don't know the language; as with any language learn the idioms before you judge anything.
If the idioms change every 3 or 4 years, the language isn't ready.

In some cases (eg if it changes significantly every 3 or 4 years for over 30 years) the designers aren't really designers; they "just want it to work", which is how they threw the language together.
Quote:
Regarding the '-a' (AND) and '-o' (OR), the related section uses some examples like `test "$1" -a "$2"' with '$1' asigned to '!' and '$2' to NULL. Such a nice example! I usually do not use that syntax but only with bash! for obvious reasons when relying on an unknown sh[ell] (POSIX compliant or not.) I always use test operand like '-n', '-z' ... for that making the pointed issue almost irrelevant. Sometimes I _do_ use the syntax you're highlighting on the previous example but it become heavy in scripts using that syntax quite often.

Huh? The variant I put up is much lighter in terms of following what's happening.
Quote:
So, a preference goes to '-a'/'-o' on plain sh with the previous cautious usage.

Not following that at all. If you got rid of all the redundant braces, you'd be able to see what you're doing a lot better, as well as giving others (including yourself 6 months later) less cognitive load.

Still do what you want; I'm sincerely not bothered at all.
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Sun Mar 15, 2015 10:53 pm    Post subject: Reply with quote

Quote:
Add kdbus branch to linux-next?
Can you add the:
https://git.kernel.org/cgit/linux/kernel/git/gregkh/char-misc.git/ kdbus

tree to linux-next so that it gets the normal amount of testing that
linux-next provides?

Yes, because "this is all about us being good guys, and doing everything the right way", see we're humble really.
It's all about getting the right testing done, cos gosh-darn it, we've got the Right Stuff (this isn't a corporate game at all, oh no.. this is the Waltons.)

Just don't talk about things we don't understand like modularity, cohesion and coupling, let alone how to write network apps (we're desktop developers^W^W "kernel-experts", remember.) Or we'll get stroppy, throw a tantrum, and then just wait a week or two and act as if nothing happened.
Quote:
The code has been reworked and reviewed many times, and this last round seems to have no objections,

You mean no-one cba repeating the same old arguments over and over; funnily enough they think you're going to actually rethink what you're doing, now that the glaring userland problems have been pointed, over 2 major rounds this time, and indeed over the last 5 or so years.

Silly rabbits; they expect good faith. Instead they're going to get another code-dump of most awful ordure, then be told it's their fault as "it's in the kernel now," so let's you all just hurry up and make it work so I can sell the sizzle not the sausage, and we marketing bods can take all the money while we're about it. After all, no sales, no business.
You work for us now, not the other way round. "Like my Lamborghini? No you can't touch it, back to work, prole!"
Anon-E-moose wrote:
What a whore!!!

Edit to add:
A whore being someone that sells part of themselves for money,
and IMO Greg has sold his soul for "filthy lucre"

Yeah, like Mark Thomas used to wear a T-shirt saying "meeja hore" on stage.

Still best watch out, the new-age thought-police are on your case.. CoC-2.0(the liars refrain) will be on your tail soon.. ;-)

Huxley was an optimist.

"Good night, Jim bob!"
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Mon Mar 16, 2015 12:43 am    Post subject: Reply with quote

Quote:
The code has been reworked and reviewed many times, and this last round
seems to have no objections, so I'm queueing it up to be merged for
4.1-rc1.


So all the important questions/queries about the design--ioctl vs. syscals and BIG (meta)data to name the fundamental ones--have been wiped out on cosmetic documentation rewording without any notice on the supposedly NEXT round of the patchset?! WTF, so the *review* is done? And I thought a review was about discussing technical points instead of wording a few sentences.

And where is the ChangeLog of the modifications per review suggestions?! Down in the (hi)story hole?

Is this a bad joke... with last minute important fixes of the USER-SPACE test suite on the LAST round of the "review"?

EDIT(LATS-MINUT-IMPORTANT-FIX[irony]): GOD BLESS RH!
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/


Last edited by tclover on Mon Mar 16, 2015 1:02 am; edited 1 time in total
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Mon Mar 16, 2015 12:57 am    Post subject: Reply with quote

steveL wrote:
Huh? The variant I put up is much lighter in terms of following what's happening.


True if it was a complicated test chain. It's just a test of the controller and parameter name followed by the actual control (supervision) group which should be non NULL to be processed. Well, your splitting may improve reading a supposedly order but... any of those three test is fatal whatsoever might be the other arguments.

Quote:
Quote:
So, a preference goes to '-a'/'-o' on plain sh with the previous cautious usage.

Not following that at all. If you got rid of all the redundant braces, you'd be able to see what you're doing a lot better, as well as giving others (including yourself 6 months later) less cognitive load.


Just stating a preference of using '-a/o' cautiously as stated above because of... yet the braces introduces this heaviness I am talking about. Just decided to go with it since the design embraced a "keep it simple and sane!" approach which brings in quite a few useful features with simple helpers costing a very few lines.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
mv
Watchman
Watchman


Joined: 20 Apr 2005
Posts: 6293

PostPosted: Mon Mar 16, 2015 6:50 am    Post subject: Reply with quote

tclover wrote:
I always use test operand like '-n', '-z' ... for that making the pointed issue almost irrelevant

You are wrong: Even if using -n and -z properly, it is impossible to combine it with -a reliably, unless you have checked the content of the variables in advance.
Here is an example:
Code:
A='='
[ -n "${A}" -a -n "${A}" ]

Your result may depend on the shell; try that e.g. in bash-4.2_p53.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Mon Mar 16, 2015 7:48 am    Post subject: Reply with quote

mv wrote:
tclover wrote:
I always use test operand like '-n', '-z' ... for that making the pointed issue almost irrelevant

You are wrong: Even if using -n and -z properly, it is impossible to combine it with -a reliably, unless you have checked the content of the variables in advance.
Here is an example:
Code:
A='='
[ -n "${A}" -a -n "${A}" ]

Your result may depend on the shell; try that e.g. in bash-4.2_p53.

Right... bash-4.3_p30 fails, BusyBox-1.22.1/ash fails, zsh-5.0.5 (with or without `emulate sh') fails...

EDIT: However... I may or may not update the related lines because... the relevant line is meant to fails for _such_ bad arguments. Phew, that's quite some lucky turns of it.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Mon Mar 16, 2015 3:44 pm    Post subject: Reply with quote

tclover wrote:
However... I may or may not update the related lines because... the relevant line is meant to fails for _such_ bad arguments. Phew, that's quite some lucky turns of it.

That's nonsense. Just check the arguments before you use them if they're not script literals; that's what shell functions and case are for.

In that way you can just notify the user properly, instead of the script looking borked.

It's like writing any code: you have to check your pre-conditions since, as you say, that is what allows us to fail early and fail hard, so the user can simply sort out whatever the problem is, and try again/restart the service etc.

The first thing my boss made me work on when I started learning bash, was supporting routines to verify our inputs. Most often than not that finds mistakes in my script much more than it does in the inputs.

And you're right, it really does help when you can bail out with decent error messages before the event, and catches our bugs before they happen; and when they do we always get a stack trace(bash), and the user always knows that it is a bug, which saves everyone time.

Hmm thinking on it, we do deliberately allow for some things to fail similarly in configure space, but they're script-internals that are always meant to be integral, not user input.
So we would get error messages if they happen to be rewritten by the application script, to be non-numeric, but we don't check them everywhere, we just use them, specifically because they are not user input, but script-internal (and we want an ugly error message if the script borks them.)

Thus I do see where you're coming from; but you have to draw a distinction between what belongs to the script (because we set it ourselves) and everything else (user inputs, configuration, data from the file-system.) The former is only what we explicitly assigned, or have verified. perl's bless/taint is a nice metaphor for this.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Mon Mar 16, 2015 5:01 pm    Post subject: Reply with quote

steveL wrote:
That's nonsense.

Maybe not...

You're almost right on every single detail you're covering afterward, stilll...

In this case '$1' is an internal simple '[:alpha:]' word--controller name,--so the 1st/3rd test are done on plain word ('[:alpha:]' & '/'). Now, the only risky one is '$2' which comes from configuration files and normally the test would almost pass if '$2' is !NULL and may start with garbage like '[:punct:]' which is the source of... the issue here. If it's the case--a user configured a controller e.g. 'CGROUP_CPU="= !"' --the following 'while' loop would just die right away and nothing would be done with that garbage--which is the right way to treat this for such a voluntary borkage.

So, leaving it as is for now--I don't see any reason to bother for such a (user) voluntary borkage.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Tue Mar 17, 2015 6:46 am    Post subject: Reply with quote

tclover wrote:
You're almost right on every single detail you're covering afterward, still...

Which part was factually incorrect?
Quote:
So, leaving it as is for now--I don't see any reason to bother for such a (user) voluntary borkage.

Look how much time you've just spent, reasoning out why you might be okay possibly in this instance.

I prefer write once, and forget (until I want to rework it.)

Write robust, so you can reuse without worry.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Tue Mar 17, 2015 8:27 am    Post subject: Reply with quote

steveL wrote:
I prefer write once, and forget (until I want to rework it.)

Write robust, so you can reuse without worry.


A good approach certainly.

But it's almost the case here, the risky argument is tested beforehand with '[ -n "$arg" ]' so risky '[:punct:]' won't be coming. Though I did not write the whole thing myself,--this is qnikst's contribution to OpenRC/CGroup,--I've just cleaned it and improved a few things, and then integrated it to supervision service model. I don't bit the robust thing _first_ unless there is a reason to... it. I prefer by far the simple/efficient approach first and then compose what might arise afterwards and THIS put my time on efficient usage instead of looking to perfection and robustness pointlessly--it might come along the way or _never_ (no problem with me.)

I've just took the time to respond to your well meaning and useful inquiries and then found out that it was right to leave it 'as is' unless _actual_ issues arise.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
Yamakuzure
Advocate
Advocate


Joined: 21 Jun 2006
Posts: 2273
Location: Bardowick, Germany

PostPosted: Tue Mar 17, 2015 2:10 pm    Post subject: Reply with quote

Oh my god, just update that instead of writing again and again why updating it would be a good idea but you won't do it because it is not a problem (now). You certainly are a "yes, but..."-person, right? It only wastes your precious time as a developer. Do it, say "thank you" and move on.
_________________
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
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Tue Mar 17, 2015 5:33 pm    Post subject: Reply with quote

tclover wrote:
Though I did not write the whole thing myself,--this is qnikst's contribution to OpenRC/CGroup

Actually most of the "newer" cgroup code is my script.

The borkage is ofc all Hubbs, who really should be forcibly retired if he refuses to get off the pot.

Especially since he refused to correct the copyright, which I raised with qnikst about a year ago.

No I won't take it up with the same spineless leeches who refused to do anything about his libel.

And no, I'm not interested in suing Gentoo Foundation Inc., either.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Tue Mar 17, 2015 7:42 pm    Post subject: Reply with quote

steveL wrote:
tclover wrote:
Though I did not write the whole thing myself,--this is qnikst's contribution to OpenRC/CGroup

Actually most of the "newer" cgroup code is my script.


Well then you can follow a clean up attempt patch discussion here then. Though, they seem to be notoriously playing dead sometimes, I've set a month duration on bpaste for the patch. And WH found a funny ("tried to apply the patch manually and almost have to edit everything at hand") excuse which is blatantly false for the first patch... I've just played an apparently deaf/(false-)polite hear on it. And then included that single line blocker on the second round.
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
saellaven
Guru
Guru


Joined: 23 Jul 2006
Posts: 526

PostPosted: Tue Mar 17, 2015 8:03 pm    Post subject: Reply with quote

steveL wrote:

And no, I'm not interested in suing Gentoo Foundation Inc., either.


Unfortunately, that might be exactly what it takes to knock williamh off his high horse where he's been abusing his position for a long time now while everyone that can do something about it twiddles their thumbs.
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Wed Mar 18, 2015 4:11 am    Post subject: Reply with quote

steveL wrote:
And no, I'm not interested in suing Gentoo Foundation Inc., either.

saellaven wrote:
Unfortunately, that might be exactly what it takes to knock williamh off his high horse where he's been abusing his position for a long time now while everyone that can do something about it twiddles their thumbs.

I understand where you're coming from, but it really wouldn't do that; it would just cause more division, and wagon-circling, none of which would help us make better software.

Besides which, I really don't have the time for it. It's hard enough right now to clear any time at all for Gentoo work, due to deadlines and because of the amount I've wasted on it over the last few years (I've been told in no uncertain terms. ;)
Basically I can get time for Gentoo, including at work and using it for work stuff, so long as I maintain my current hiatus from interaction with any official developer channels, apart from the ones I've always hung out in, on IRC. eg I haven't used bugzilla for the best part of a year.

And I have to admit that prospect is rather appealing.
It's not like I ever got any joy out of it, only a load of abuse from whoever felt like it at whatever point, along with accusations of whinging if I let off steam about it elsewhere, or trolling if I answered back (along with a pile-in of vipers abusing official positions).
Too old for that sh1t now.

Thinking back I do recall not quite believing the people who told me how crappy the developer list was, when I first joined the forums. It took me months to first post, but have to admit: they were right.

The freeks took over the show, and they're utter saps when it come to politics in the real-world; the ones who aren't just selling out to corporates as a good career move, usually because they've just left college.

Still who cares, so long as they churn out ebuilds? ;) Eventually they'll learn the basics.

Perhaps that's why there's such a large "silent majority" amongst the developers, as I've been told so often.

I don't understand the mindset that can expect any sort of respect for that, though; and certainly not ego-massages just to behave as well as any 13 year-old girl does on a daily basis -- under much more trying circumstances, in the main.

Plus ca change. ;-)
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Wed Mar 18, 2015 4:13 am    Post subject: Reply with quote

tclover wrote:
I've just took the time to respond to your well meaning and useful inquiries and then found out that it was right to leave it 'as is' unless _actual_ issues arise.

No it's not. What Yamakuzure said; just look how much time you've wasted solely to argue and reason that a less robust, less generically useful variant is somehow "right". Which you knew was the case before you wasted even more time on it.

The time to sort out the fundamentals is before you distribute it, not after you've wiped the user's hard-drive. (That's a C reference to UB, or undefined behaviour; don't take it literally and start arguing about how this couldn't possibly blah blah blah.)
Back to top
View user's profile Send private message
mv
Watchman
Watchman


Joined: 20 Apr 2005
Posts: 6293

PostPosted: Wed Mar 18, 2015 7:05 am    Post subject: Reply with quote

steveL wrote:
The time to sort out the fundamentals is before you distribute it

This is not the idea of open software. Instead: publish as soon as possible so that others can contribute and provide such minor patches. (OTOH, I do not understand why rejecting such a patch which really does no harm and in some corner case might improve things.)
The situation is different, of course, when you mark a release as "stable".
Back to top
View user's profile Send private message
steveL
Watchman
Watchman


Joined: 13 Sep 2006
Posts: 5153
Location: The Peanut Gallery

PostPosted: Wed Mar 18, 2015 3:44 pm    Post subject: Reply with quote

steveL wrote:
The time to sort out the fundamentals is before you distribute it

mv wrote:
This is not the idea of open software. Instead: publish as soon as possible so that others can contribute and provide such minor patches.

Not true; first you have to know the language you're using before you start unleashing "hello world"-2.0 onto the rest of us.

I'm really not interested in "open software" written by people who haven't learnt how to code yet; not when it comes to delivery, nor distribution.
Quote:
OTOH, I do not understand why rejecting such a patch which really does no harm and in some corner case might improve things.

Me neither, though we do run into this kind of thing on IRC all the time.

It's almost like it's a point of pride not to concede things could be done better, or something; really I think it's just part of being attached to your work, which is what you have to let go of.
It takes a long time to do that, ime, though some people have that attitude from the beginning.

It's natural for anyone who cares about what they're doing. The flipside is they also usually care about doing things right.

IME you just have to be straightforward, and occasionally blunt to get through.
Back to top
View user's profile Send private message
tclover
Guru
Guru


Joined: 10 Apr 2011
Posts: 516

PostPosted: Fri Mar 20, 2015 9:43 am    Post subject: Reply with quote

steveL wrote:
It's almost like it's a point of pride not to concede things could be done better, or something; really I think it's just part of being attached to your work, which is what you have to let go of.
It takes a long time to do that, ime, though some people have that attitude from the beginning.

It's natural for anyone who cares about what they're doing. The flipside is they also usually care about doing things right.

IME you just have to be straightforward, and occasionally blunt to get through.


Couldn't be truer for... OpenRC team/upstream. These guys threw the same *mantra* of FORM-BEFORE-CONTENT several times on me before closing bug/issue without taking a look at the patch.

Oh boys, are they just a bunch of Kings--that want anybody before them to kiss their feet before engaging in any meaningful discussion--or something?!

Just opened another bug then and wait their... reaction--just do not hold your breath. LULZ
_________________
home/:mkinitramfs-ll/:supervision/:e-gtk-theme/:overlay/
Back to top
View user's profile Send private message
depontius
Advocate
Advocate


Joined: 05 May 2004
Posts: 3389

PostPosted: Thu Mar 26, 2015 3:28 pm    Post subject: Reply with quote

Speaking of kdbus in the kernel, which we haven't been for a few posts now, there has been stunningly little feedback on the V4 release. Only a few minor suggestions, which have generally been accepted with thanks. No discussion of the design point, or anything like that. With absolutely no possible conflict of interest, Greg KH has directed that it be pulled into linux-next for inclusion as part of linux-4.1.

So at one end of the spectrum we have a subsystem flawed at the specification level getting fastpathed into the kernel. At the other end we have debate about fine points of shell scripting delaying a subsystem which I strongly suspect is already in far better shape than kdbus.

There has got to be a happy medium somewhere in there.
_________________
.sigs waste space and bandwidth
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 4213
Location: Dallas area

PostPosted: Thu Mar 26, 2015 6:23 pm    Post subject: Reply with quote

depontius wrote:
Greg KH has directed that it be pulled into linux-next for inclusion as part of linux-4.1.


Making it into linux-next doesn't mean that it will be "included" in the 4.1 release.

There were some performance questions, last week, but haven't seen any real answers back yet.
_________________
Asus m5a99fx, FX 8320 - nouveau, oss4, rx550 for qemu passthrough
Acer laptop E5-575, i3-7100u - i965, alsa
---both---
5.0.13 zen kernel, profile 17.1 (no-pie & modified) amd64-no-multilib
gcc 8.2.0, eudev, openrc, openbox, palemoon
Back to top
View user's profile Send private message
depontius
Advocate
Advocate


Joined: 05 May 2004
Posts: 3389

PostPosted: Mon Mar 30, 2015 3:53 pm    Post subject: Reply with quote

It's not April 1 yet, but...

The systemd Project Forks the Linux Kernel
http://distrowatch.com/weekly.php?issue=20150330#community
_________________
.sigs waste space and bandwidth
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Goto page Previous  1, 2, 3 ... 9, 10, 11 ... 25, 26, 27  Next
Page 10 of 27

 
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