| View previous topic :: View next topic |
| Author |
Message |
ferringb
Developer
Joined: 03 Apr 2003
Posts: 355
Location: USA
|
 Posted: Sun Nov 26, 2006 7:45 pm Post subject: |
 |
|
| ciaranm wrote: | | nightmorph wrote: | You dredged up a minor bug from 2003. How typically petty. Granted, you always have to be right, so in this case, it looks like you are . . . . <storytime mode>once upon a time, boys and girls, Portage needed a few minor updates to take advantage of new Python 2.3 functionality</storytime mode>.
This is the best you can come up with? Go back to coding your favorite package manager; you're better at that than defending your own statements.  |
You know, Python doesn't update major versions that often... 2.2 to 2.3 is the last one for which I was paying attention. That was back in the days when I did user support... |
Interesting claim.
| ciaranm wrote: | | ferringb wrote: | What internal standards body backs up ruby?
...
Out of curiousity, who ratified libxml, which y'all use? Or libebt? Where is the independant (secondary) implementation of it? |
We have a soft dep upon libxml that is very carefully designed not to break even if someone compiles against it and then royally fucks up their libxml install.
libebt used to be part of the Paludis code. It was split out because a few other projects use it. It is, however, well and truly under our control.
So basically, the only artificial external non-standard dependency introduced by Paludis is bash, which is required for the tree anyway. |
Where however is the international standard backing it? You're dodging the fact you leveled that crazy con at portage, when a form of it hits the core of paludis just the same.
Seriously, that was a fairly loony claim.
| ciaranm wrote: | | ferringb wrote: | | Meanwhile, python has compliance tests, which PyPy, jython, and ironpython rely on (jython is a somewhat shitty example since they stopped at python 2.2). |
Yet they still can't ship a simple crypto library without screwing it up... |
It would help if you had a clue what you were talking about; pycrypto isn't actually part of python.
It's an extension module; would be like blaming paludis for libebt being whacky.
Meanwhile, python 2.5 does ship a new crypto lib, one that's far less suckage then pycrypto.
| ciaranm wrote: | | ferringb wrote: | | Line of bullshit, that one; if you're going to claim it changes API arbitrarily (implying whenever, semi-frequently), better back it up with more then a single example from 3 years back. |
No no, I'm claiming they can, and that they have in the past. |
So by that redefinition, c++ changes ABI arbitrarily. Fair enough.
| ciaranm wrote: | | ferringb wrote: | | Meanwhile, c++ abi changes *are* a disadvantage. Going from 4.x to 3.4 can break existing binaries; |
Downgrading pretty much anything has that kind of issue. Vendors rarely guarantee backwards portability. |
Funny that, portage doesn't suffer it for downgrading to 2.5 to 2.4.
| ciaranm wrote: | | ferringb wrote: | | Figured you'd pull up an example of portage relying on incorrect behaviour from 3+ years back; portage relying on a daft bug in 2.2 isn't much of an example; further it's not actually backing up the "newer python releases"; lemme guess, the bug 1600 is next on the list? |
You said that it's never had problems when you'd been around. I demonstrated that it had. You could do the honourable thing and say "oh, I forgot about that, well remembered". |
Just barely in range on that one; well played master troll, you bested me on a bug that running stable, I never saw when I was getting going.
You're also leaving out that that particular change shifted portage into a version neutral location to prevent further issues; not conducive to your claim I realize.
| ciaranm wrote: | | ferringb wrote: | | Where's the new examples? Say transition 2.3 to 2.4, 2.4 to 2.5? |
Changing your requirements, eh? How long until it's "find something that will be broken by 2.6"? |
Nope; you stated portage was held up by python for long periods of time; 2.3 to 2.4, no issue, 2.4 to 2.5, no issue. You brought up 2.2 to 2.3... which looks to have been a 2 month period (extra month of stabling). So in the last 3 versions, one issue.
You could do the honourable thing, and say "Oh, I was exagerating from an old example that wasn't really that long of a delay, thanks for correcting me".
| ciaranm wrote: | | ferringb wrote: | | Or are you going to stick to picking out ancient examples as proof? If that's the case, you really don't have much of a leg to stand on when trying to defend the gcc 3.3 to 3.4, double standards being a bit of bullshit and all. |
gcc 3.3 doesn't include a C++ compiler, so Paludis doesn't care about it. |
Christ, you're claiming I'm changing the requirements?
gcc 3.3 doesn't have a c++ compiler you're willing to deal with would be the honorable thing to state there; else if you're going to stick to that claim, guess a whole lot of people got screwed over by a nonexistant 3.3 to 3.4 abi breakage for gcc's c++ symbol mangling.
The fun part about arguing with you is that you occasionally know some interesting tidbits; the annoying part about arguing with you is that you're willing to lie your ass off at any given moment (the 3.3 to 3.4 example above comes to mind)... mind you it's never "lieing" persay, it's miscommunication with the recipient of course being at fault (we all are such gross idiots one wonders why you even indulge us).
_________________
I don't want to be buried in a pet cemetery. ~Ramones |
|
| Back to top |
|
 |
ciaranm
Retired Dev
Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding
|
| Posted: Sun Nov 26, 2006 8:01 pm Post subject: |
|
|
| ferringb wrote: | | Interesting claim. |
You may not know this, but in the good old days me and rac were the user support people...
Unfortunately, seemant decided we were no longer allowed to banninate anyone using love-sources, so I stopped.
| ferringb wrote: | | Where however is the international standard backing it? You're dodging the fact you leveled that crazy con at portage, when a form of it hits the core of paludis just the same. |
The only core Paludis component over which we do not have control and over which there is no international standard is bash.
| ciaranm wrote: | | It's an extension module; would be like blaming paludis for libebt being whacky. |
That'd be a fair claim. We wouldn't use libebt where it whacky.
| ferringb wrote: | | So by that redefinition, c++ changes ABI arbitrarily. Fair enough. |
Actually, the ABI is standardised. It's a vendor standard (and not from the gcc people), not an international one, which doesn't help...
| ferringb wrote: | | Funny that, portage doesn't suffer it for downgrading to 2.5 to 2.4. |
Try downgrading to 2.2.
| ferringb wrote: | | Just barely in range on that one; well played master troll, you bested me on a bug that running stable, I never saw when I was getting going. |
So when I'm right I'm a troll?
| ferringb wrote: | | ciaranm wrote: | | ferringb wrote: | | Or are you going to stick to picking out ancient examples as proof? If that's the case, you really don't have much of a leg to stand on when trying to defend the gcc 3.3 to 3.4, double standards being a bit of bullshit and all. |
gcc 3.3 doesn't include a C++ compiler, so Paludis doesn't care about it. |
Christ, you're claiming I'm changing the requirements? |
No no. 3.3 won't build Paludis and never did build Paludis.
| ferringb wrote: | | gcc 3.3 doesn't have a c++ compiler you're willing to deal with would be the honorable thing to state there; else if you're going to stick to that claim, guess a whole lot of people got screwed over by a nonexistant 3.3 to 3.4 abi breakage for gcc's c++ symbol mangling. |
A C++ compiler can compile this:
| Code: |
template <typename T_>
struct S
{
static const int s = 0;
};
template <typename T_>
struct S<const T_>
{
};
template <typename T_>
struct S<const T_ &>
{
static const int t = 0;
};
int main(int, char *[])
{
return S<int &>::s + S<const int &>::t;
}
|
3.3 can't compile that, thus 3.3 is not a C++ compiler. Paludis relies upon something considerably more complex (but that triggers the same issues) working correctly. |
|
| Back to top |
|
|
Zinic
n00b
Joined: 19 Dec 2004
Posts: 9
Location: Texas
|
| Posted: Sun Nov 26, 2006 8:16 pm Post subject: |
|
|
You know, in all honesty, after reading through all of these posts I find myself strangely interested in Paludis.
_________________
/* I'm not sure what this function does or why the hell it's even in here but every time I take it out, the damn thing won't compile. I've asked around and no one else knows what it does either. I think we'd better keep it in...
-John */ |
|
| Back to top |
|
|
boniek
Guru
Joined: 26 Mar 2005
Posts: 373
|
| Posted: Sun Nov 26, 2006 8:41 pm Post subject: |
|
|
I recommend trying it out. I have converted stage3 to paludis and installed whole system from there. WFM not counting some circular dependency errors, which were easy to overcome and thats not really Paludis fault. Portage and Paludis work for me simultaneously on the same system. From user perspective most striking difference is in its speed. Paludis is very fast at syncing. Also there are some very interesting features in svn (uninstalling with all dependencies, or only with dependencies that are not used by other packages). Can't wait for paludis 0.8.5 
_________________
[HOWTO]New freetype subpixel font rendering for lcd monitors |
|
| Back to top |
|
|
ciaranm
Retired Dev
Joined: 19 Jul 2003
Posts: 1719
Location: In Hiding
|
| Posted: Sun Nov 26, 2006 8:46 pm Post subject: |
|
|
| boniek wrote: | | I recommend trying it out. I have converted stage3 to paludis and installed whole system from there. WFM not counting some circular dependency errors, which were easy to overcome and thats not really Paludis fault. |
Not really? Surely you mean 'not at all'. 
| Quote: | | Can't wait for paludis 0.8.5 |
Will probably have to be 0.10.0, since it's an ABI break... |
|
| Back to top |
|
|
ferringb
Developer
Joined: 03 Apr 2003
Posts: 355
Location: USA
|
| Posted: Sun Nov 26, 2006 8:54 pm Post subject: |
|
|
| boniek wrote: | | Paludis is very fast at syncing. |
FEATURES="-metadata-transfer" disables the cache transfer (the slow part of it).
Helps to actually configure things the same when you're attempting a comparison
_________________
I don't want to be buried in a pet cemetery. ~Ramones |
|
| Back to top |
|
|
revenger
n00b
Joined: 23 Sep 2005
Posts: 23
Location: The Netherlands
|
| Posted: Sun Nov 26, 2006 8:58 pm Post subject: |
|
|
| what's with the 2.6% incompatibility with portage? |
|
| Back to top |
|
|
boniek
Guru
Joined: 26 Mar 2005
Posts: 373
|
| Posted: Sun Nov 26, 2006 8:59 pm Post subject: |
|
|
| ferringb wrote: | | boniek wrote: | | Paludis is very fast at syncing. |
FEATURES="-metadata-transfer" disables the cache transfer (the slow part of it).
Helps to actually configure things the same when you're attempting a comparison |
I didn't know about that. Thanks for the tip! Why it is not set as default in Portage?
_________________
[HOWTO]New freetype subpixel font rendering for lcd monitors |
|
| Back to top |
|
|
ferringb
Developer
Joined: 03 Apr 2003
Posts: 355
Location: USA
|
| Posted: Sun Nov 26, 2006 9:10 pm Post subject: |
|
|
| boniek wrote: | | ferringb wrote: | | boniek wrote: | | Paludis is very fast at syncing. |
FEATURES="-metadata-transfer" disables the cache transfer (the slow part of it).
Helps to actually configure things the same when you're attempting a comparison |
I didn't know about that. Thanks for the tip! Why it is not set as default in Portage? |
Because that step (localizing the cache) is required if you want to work with overlays in portage (which most people do).
If you don't localize the cache, non-root users can't update it for overlay eclasses; this is an issue paludis suffers also, although portage provides a metadata-overlay cache module to enable it to still work (pkgcore does something equivalent automatically for portage configuration).
For the general speed claim; I'll just point out that it's faster in proc execution, but for most ops you're going to hit IO as the main limiter (in which case the run times for both go fairly pair shaped and become equivalent, with portage slightly leading last I tested). Feel free to do your own testing, just make sure you look up how to flush the kernel pages/dentry cache to keep the comparison on the same footing.
_________________
I don't want to be buried in a pet cemetery. ~Ramones
Last edited by ferringb on Sun Nov 26, 2006 9:16 pm; edited 1 time in total |
|
| Back to top |
|
|
ferringb
Developer
Joined: 03 Apr 2003
Posts: 355
Location: USA
|
| Posted: Sun Nov 26, 2006 9:11 pm Post subject: |
|
|
| ciaranm wrote: | | ferringb wrote: | | Interesting claim. |
You may not know this, but in the good old days me and rac were the user support people...
Unfortunately, seemant decided we were no longer allowed to banninate anyone using love-sources, so I stopped. |
May not know this, but I was also user support in #gentoo for a good chunk of 06/03 through 01/05. Funny thing, I don't recall seeing you talking to much in there aside from harassing folk (and subsequently getting stripped of powers for harassment of users, bitching about it incessantly).
Meanwhile, the "interesting claim" was in regard to "python doesn't update major versions that often"; they only do it every year, which, yeah, I spose is not that often if you're claiming a monthly release cycle for language implementations is the norm.
| ciaranm wrote: | | ferringb wrote: | | Where however is the international standard backing it? You're dodging the fact you leveled that crazy con at portage, when a form of it hits the core of paludis just the same. |
The only core Paludis component over which we do not have control and over which there is no international standard is bash. |
Honorable thing to do here would be to point out that "yes, we too have to rely on others to not screw up api".
A forlorn concept to you I suspect.
| ciaranm wrote: | | ferringb wrote: | | It's an extension module; would be like blaming paludis for libebt being whacky. |
That'd be a fair claim. We wouldn't use libebt where it whacky. |
You left out "thus my FUD against python was actually bullshit"; understandable.
| ciaranm wrote: | | ferringb wrote: | | So by that redefinition, c++ changes ABI arbitrarily. Fair enough. |
Actually, the ABI is standardised. It's a vendor standard (and not from the gcc people), not an international one, which doesn't help... |
Point still stands, considering everyone uses gcc (don't even try the icc angle, you know as well as I they're the corner cases).
Also says a lot about your push for standards and specifications (where is the paludis spec btw, since you seemed so hell bent demanding prefix provide a spec before even writing code?).
| ferringb wrote: | | ferringb wrote: | | Funny that, portage doesn't suffer it for downgrading to 2.5 to 2.4. |
Try downgrading to 2.2. |
Depends on the portage version; these days 2.3 is the required min, although if you bump back to 2.0.51, should be able to bounce back to 2.2 without issue.
Now, if you were claiming that paludis has a minimal version, sure, I'd roll with that; course, the original point wasn't about paludis minimal version, it was that compiling with gcc 4.x effectively locks you at 4.x; python lacks no equivalent restriction there.
| ciaranm wrote: | | ferringb wrote: | | Just barely in range on that one; well played master troll, you bested me on a bug that running stable, I never saw when I was getting going. |
So when I'm right I'm a troll? |
Considering the points being claimed, yep. Doesn't mean you're wrong; just means that you hit a factual claim, and caught me in a blind spot.
As said, well played.
| ciaranm wrote: | | ferringb wrote: | | ciaranm wrote: |
gcc 3.3 doesn't include a C++ compiler, so Paludis doesn't care about it. |
Christ, you're claiming I'm changing the requirements? |
No no. 3.3 won't build Paludis and never did build Paludis. |
You stated it didn't include a C++ compiler, paludis inability to compile against it was not the claim (I could give a rats ass about paludis compilation). Re-read you're statement there. Now, if you misstated, and actually meant-
| ferringb wrote: | | gcc 3.3 doesn't have a c++ compiler you're willing to deal with would be the honorable thing to state there; else if you're going to stick to that claim, guess a whole lot of people got screwed over by a nonexistant 3.3 to 3.4 abi breakage for gcc's c++ symbol mangling. |
I'd shut up. But since you seem incapable of admitting that 3.3 actually had a c++ compiler (shitty though it may have been), not much point.
Note you also dropped the quote about lieing; prudent I suppose when you're busy redefining early statements.
| ciaranm wrote: | A C++ compiler can compile this:
<snip>
3.3 can't compile that, thus 3.3 is not a C++ compiler. |
Ahh... so qt for gcc 3.3 doesn't actually use a C++ compiler. Funny, thought they were c++ based. Oh well.
If the claim were "3.3 parses some crap incorrectly", sure, but the claim is a flat out "it's not a c++ compiler". One is true, one is the result of hitting the pipe one too many times, and/or basing the definition off of your own pet projects requirements (statement was not about paludis, was about gcc).
Meanwhile... hitting the point where discussion is going to go rather circular if past bouts are any indication (say, for example going by the mayhem you unleashed in your indefinite suspension bug), so I'm going to bow out and go back to work.
_________________
I don't want to be buried in a pet cemetery. ~Ramones |
|
| Back to top |
|
|
Lepaca Kliffoth
l33t
Joined: 28 Apr 2004
Posts: 737
Location: Florence, Italy
|
| Posted: Sun Nov 26, 2006 9:19 pm Post subject: |
|
|
Holy shit I never saw such a collection of trolls in my whole life. The amount of trolling and bickering going on between devs and OTW vips is mindbogging. On one side we have current devs crapping their pants in fear of some app a retired dev is programming, on the other we can witness said retired dev excercising the kind of sarcasm that everyone hates for his own amusement. In the middle, attempts at actually talking about Paludis by potential Paludis users are drowned by discussions about shit happening with python or c++ that nobody gives a flying rat's turd about, since in the end nobody really is a python or c++ user.
When I clicked the link I wanted to ask something about Paludis but doing that would water down all the trolls and since I like my drinks stiff and OTW threads at least 80% trollish I'll just leave this stupid post and laugh my way out of the room. You all suck, but in a very funny way. Thanks for keeping dumb guys like me entertained.
_________________
It isn't enough to win - everyone else must lose, and you also have to rub it in their face (maybe chop off an arm too for good measure).
Animebox! |
|
| Back to top |
|
|
Lepaca Kliffoth
l33t
Joined: 28 Apr 2004
Posts: 737
Location: Florence, Italy
|
| Posted: Sun Nov 26, 2006 9:22 pm Post subject: |
|
|
| Lepaca Kliffoth wrote: | Holy shit I never saw such a collection of trolls in my whole life. The amount of trolling and bickering going on between devs and OTW vips is mindbogging. On one side we have current devs crapping their pants in fear of some app a retired dev is programming, on the other we can witness said retired dev excercising the kind of sarcasm that everyone hates for his own amusement. In the middle, attempts at actually talking about Paludis by potential Paludis users are drowned by discussions about shit happening with python or c++ that nobody gives a flying rat's turd about, since in the end nobody really is a python or c++ user.
When I clicked the link I wanted to ask something about Paludis but doing that would water down all the trolls and since I like my drinks stiff and OTW threads at least 80% trollish I'll just leave this stupid post and laugh my way out of the room. You all suck, but in a very funny way. Thanks for keeping dumb guys like me entertained. |
STFU N00B PALUDIS IS THA FUTUR3 I TELL YA JUST GIVE IT TIME CUZ PACKAGE MANAGERZ NEED HUMANZ TO GIVE THEM TIME AND TESTING AND SHIT
_________________
It isn't enough to win - everyone else must lose, and you also have to rub it in their face (maybe chop off an arm too for good measure).
Animebox! |
|
| Back to top |
|
|
Lepaca Kliffoth
l33t
Joined: 28 Apr 2004
Posts: 737
Location: Florence, Italy
|
| Posted: Sun Nov 26, 2006 9:24 pm Post subject: |
|
|
| Lepaca Kliffoth wrote: | | Lepaca Kliffoth wrote: | Holy shit I never saw such a collection of trolls in my whole life. The amount of trolling and bickering going on between devs and OTW vips is mindbogging. On one side we have current devs crapping their pants in fear of some app a retired dev is programming, on the other we can witness said retired dev excercising the kind of sarcasm that everyone hates for his own amusement. In the middle, attempts at actually talking about Paludis by potential Paludis users are drowned by discussions about shit happening with python or c++ that nobody gives a flying rat's turd about, since in the end nobody really is a python or c++ user.
When I clicked the link I wanted to ask something about Paludis but doing that would water down all the trolls and since I like my drinks stiff and OTW threads at least 80% trollish I'll just leave this stupid post and laugh my way out of the room. You all suck, but in a very funny way. Thanks for keeping dumb guys like me entertained. |
STFU N00B PALUDIS IS THA FUTUR3 I TELL YA JUST GIVE IT TIME CUZ PACKAGE MANAGERZ NEED HUMANZ TO GIVE THEM TIME AND TESTING AND SHIT |
WTF R U SAYIN NIGGA PALUDIS WILL NEVER BE SUPPORTED CUZ C++ BREAKZ AND PY DON'T GO F*K URSELV STAY THE F*K AWAY FROM PORTAGE
_________________
It isn't enough to win - everyone else must lose, and you also have to rub it in their face (maybe chop off an arm too for good measure).
Animebox! |
|
| Back to top |
|
|
Shadow Skill
Veteran
Joined: 04 Dec 2004
Posts: 1022
|
| Posted: Sun Nov 26, 2006 9:24 pm Post subject: |
|
|
I've had some problems with ebuilds not wanting to behave with Paludis other than that and some missing features it works fairly well and I think it is destined for great things.
_________________
Ware wa mutekinari.
Wa ga kage waza ni kanau mono nashi.
Wa ga ichigeki wa mutekinari.
"First there was nothing, so the lord gave us light. There was still nothing, but at least you could see it." |
|
| Back to top |
|
|
Lepaca Kliffoth
l33t
Joined: 28 Apr 2004
Posts: 737
Location: Florence, Italy
|
| Posted: Sun Nov 26, 2006 9:26 pm Post subject: |
|
|
There, I think that should kind of cover it. G'night
_________________
It isn't enough to win - everyone else must lose, and you also have to rub it in their face (maybe chop off an arm too for good measure).
Animebox! |
|
| Back to top |
|
|
amne
Bodhisattva
Joined: 17 Nov 2002
Posts: 6377
Location: Graz / EU
|
| Posted: Sun Nov 26, 2006 9:41 pm Post subject: |
|
|
| amne wrote: | | Ok, enough of this crap or i'll get the pony. |
There it is. *stomp*
_________________
Dinosaur week! (Ok, this thread is so last week) |
|
| Back to top |
|
|
|
|
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
|
|
Off the Wall
