| View previous topic :: View next topic |
| Author |
Message |
jerco@chile3000.com
n00b
Joined: 29 Dec 2002
Posts: 14
|
 Posted: Tue Dec 31, 2002 11:14 pm Post subject: Kylix 3 problem's... |
 |
|
Hi all.
As soon as I open the Save As dialog box it locks up.
I probe with put "kylix/bin" in "PATH" and in "LD_LIBRARY_PATH", read documents, i start "XFS", kernel it's fine ( 2.4.19-r10), lib's OK (names and links), but nothing, well not nothing, now the "first" time i probe "open a Save as Dialog" this function fine, but later this locks up the KYLIX or BCB. 
Well, new ideas to probe???  |
|
| Back to top |
|
 |
474
l33t
Joined: 19 Apr 2002
Posts: 714
|
| Posted: Wed Jan 01, 2003 9:11 am Post subject: |
|
|
| Try a vanilla kernel. Not a very deductive answer I know but I think it makes sense to do this with closed-source software because they'll probably have written it with either the vanilla or RedHat kernel in mind, and let's face it - gentoo-r10 is heavily patched. If it corrects the problem, then at least you know it's something to do with your kernel. Also, have you emerged libcompat? |
|
| Back to top |
|
|
jerco@chile3000.com
n00b
Joined: 29 Dec 2002
Posts: 14
|
| Posted: Wed Jan 01, 2003 1:01 pm Post subject: licompat?? |
|
|
I don't see this lib in my system!!!
within which package this ?
pd: I'm compilining vanilla now... , tank's |
|
| Back to top |
|
|
jerco@chile3000.com
n00b
Joined: 29 Dec 2002
Posts: 14
|
| Posted: Wed Jan 01, 2003 3:05 pm Post subject: KYLIX 3 it's working fine now!!! |
|
|
Yep, i compiling vanilla 2.4.20 and they WORK!!!!!!! 
yes.. tank's
tank's a lot... |
|
| Back to top |
|
|
ghetto
Guru
Joined: 10 Jul 2002
Posts: 369
Location: BC, Canada
|
| Posted: Mon Jan 06, 2003 4:57 pm Post subject: |
|
|
Im glad to see stuff like that work, I tried to use Kylix before but i had a similar problem.
But at the same time this is a rather unfortunate fix, after all who uses a stock kernel now a days???
Which pretty much means that if you want to use Kylix you have to either reboot into the stock kernel or else resign yourself to using it fulltime which im really not willing to do.
I will be so glad when 2.6 comes out.. finallly a stock kernel that might be runable without patching the hell out of it 
_________________
Blizzard you suck. |
|
| Back to top |
|
|
474
l33t
Joined: 19 Apr 2002
Posts: 714
|
| Posted: Mon Jan 06, 2003 5:41 pm Post subject: |
|
|
| ghetto wrote: | | But at the same time this is a rather unfortunate fix, after all who uses a stock kernel now a days??? |
Who indeed. My angle on these things has changed recently however. While, certainly on a desktop, I would probably not run a vanilla kernel without a few patches, I have reached the conclusion that a lot of the kernels that seem to be popular at the moment are quite simply over-patched. The logic behind this is simple, the more patches you add which haven't been accepted into the official tree (often not even in the 2.5 development series tree) the more likely something is to break - or a gain in one area could result in a decline in another. I have looked through the patches collection for gentoo-sources-r10 and decided that about 75% of them I don't really want. Furthermore, having scanned through the ChangeLogs of 2.4.20 I have decided that the fixes and improvements contained therein are a lot more important to me - generally.
I definitely want preemptive support and variable-hz support, perhaps also the gcc-3.x support for kernel compilation flags. Also any important fixes (note "fixes", not features) such as Andrew Morton's.
It is difficult to argue the case for better performance or "must-have" features when the fact is probably that few of the people who slavishly use them probably perform any significant benchmarks to qualatively test whether it is, in fact, more suitable - or considers that breaking existing software is not conducive to improved performance or reliability!
The adoption of Bitkeeper to maintain the sprawling kernel tree, and the need for a more formalised approach to kernel development (no doubt, because Linux is turning into big business) is quite interesting. Is the current kernel really so bad that we simply cannot possibly run it without patching it into the middle of next week? On the contrary, it's never been better.
Also, some patches that are really good (such as Robert Love's) implement features that are may not actually used by much of the code we run. Take scheduler-hints - nice, but does anyone know for a fact of any program they're running that hooks into the new syscall? I personally don't know. Any Google search just ends up with the usual plethora of developer-oriented mailing list postings which don't tend to answer such a question or even explain what it's really all about (for example, I had to go to Microsoft's site to find out what E820 support is - can you believe that?!) Until I do know, I may just as well not run it - and I sure as hell can't find any information on whether it will be used. Or take the O(1) scheduler which seems to get rewritten quite infrequently. That concerns me because the rest of the kernel may have changed a fair bit in the intervening time. And it seems that SMP users have the most to gain from O(1) but I'm not one of them.
The long and short of it is that I've decided to be quite conservative with my patches and will be running 2.4.21-pre2 with my own chosen patches from now on (also I'll be moving my server away from the gentoo-sources and to 2.4.20 with a few select patches - this server actually had a few problems with gentoo-sources-r10 - it never got past the stage of initialising my Ultra-SCSI devices). Also, when considering patches I will favour those that have actually been committed into the 2.5 tree, as that's a sure sign of a certain level of code quality for it to be accepted. That's not to say that very good patches don't get accepted because the maintainers are overwhelmed, but it's something I've got to go on. I am going to stop running kernels unless I know what's been patched and why.
It's good to see that gentoo-sources is being emphasised as the cutting-edge kernel, and that the xfs sources will be considered the safer option in future (from Gentoo Weekly News). That move was needed, I think.
I feel the same way about compile options too. People really do get carried away with those! I'm running Red Hat beta next to my Gentoo machine now. The machine on which that's situated is less powerful than my Gentoo machine, but I can't really say that the Red Hat machine (with i386 compiled code - and Red Hat's relatively conservative but well tested and tweaked kernel) is really perceptively slower during average desktop usage - and both machines are running similar software - XFree 4.2.99, Gnome and the like. I just stick to -march and -O2 and leave it at that.
Well, just my two pence (or should I say cents) worth anyway ... YMMV and all of that.
I'm definitely looking forward to the 2.6 kernel too - it sounds like it will be something special. |
|
| Back to top |
|
|
ghetto
Guru
Joined: 10 Jul 2002
Posts: 369
Location: BC, Canada
|
| Posted: Mon Jan 06, 2003 10:16 pm Post subject: |
|
|
Great post kerframil,
I totally agree with your point of view, Ive felt the same way for a while but just never bothered to sit down and think what to do about it.
I used to patch my own kernels but i found it a bit frustraiting at times so for a while now ive been sticking to the popular kernels, but like yourself I find i dont need a lot of whats in there.
What Im curious to know is how did you find out exactly which patches are being used on the gentoo-sources kernel?
Something else id like to see is something like a central patch repository for various kernel developers, I used to hate having to hunt all ove the net to find the latest and greatest patches that i hear everone talking about 
How do you get your patches?
_________________
Blizzard you suck. |
|
| 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
|
|
Portage & Programming
