Puppy 4.2 RC4 Deep Thought - Bugs & Fixes

Message

alex12 · #381 Post by **alex12** » Thu 26 Mar 2009, 00:35

ttuuxxx: What is dpup?

#382 Post by **ttuuxxx** » Thu 26 Mar 2009, 01:53

alex12 wrote:ttuuxxx: What is dpup?

Dpup is basically Barry's latest idea, he built a application called woof, basically you install it and run it, when starts it ask you what type of puppy would you like to build, Debian,Ubuntu,Slackware, So you choose Debian, then it downloads the main puppy applications and mixes with a Debian base and builds you a custom Debian-Puppy release, Its still in alpha state, but from what I've seen, its really nice. I can't get over how fast FF3 is on it

I'm building another Debian base right, a reduced version.

here's a link
http://www.murga-linux.com/puppy/viewto ... 000#288000
ttuuxxx

trio · #383 Post by **trio** » Thu 26 Mar 2009, 02:36

Flash858 wrote:The RAM/CPU
widget constantly displays a maxed out CPU (this cannot possibly be, as it is an AMD 6000+ Dual Core, and I'm running nothing but the desktop - also - the system is in no way sluggish, but VERY responsive...).

See Screenshot below.

http://hiltonheadphoto.com/screenshot.jpg

I had the same problem when using cable lan and internet is down. When booting with internet on, it's normal. Is that the case? Fyi,That widget is conky based.

Thanks

#384 Post by **ttuuxxx** » Thu 26 Mar 2009, 03:00

trio wrote:
Flash858 wrote:The RAM/CPU
widget constantly displays a maxed out CPU (this cannot possibly be, as it is an AMD 6000+ Dual Core, and I'm running nothing but the desktop - also - the system is in no way sluggish, but VERY responsive...).

See Screenshot below.

http://hiltonheadphoto.com/screenshot.jpg
I had the same problem when using cable lan and internet is down. When booting with internet on, it's normal. Is that the case? Fyi,That widget is conky based.

Thanks

Conky=R.I.P

-ing
ttuuxxx

trio · #385 Post by **trio** » Thu 26 Mar 2009, 03:10

ttuuxxx wrote: Conky=R.I.P -ing
ttuuxxx

So what?...should we prepare a funeral service?

#386 Post by **ttuuxxx** » Thu 26 Mar 2009, 03:46

trio wrote:
ttuuxxx wrote: Conky=R.I.P -ing
ttuuxxx
So what?...should we prepare a funeral service?

For conky yes pleaseeeeeeeeeeee

and use that other base I recommended, now lost in the pwidgets thread

ttuuxxx

#387 Post by **Pizzasgood** » Thu 26 Mar 2009, 04:51

ttuuxxx wrote:
Béèm wrote:If htop is reliable, cupsd is using almost 100% cpu.
Is this healthy?
Is there work on the way to get this down to a normal level?
Beem how about trying a little experiment ?
run top, open Pprocess on the right hand menu it will say with arrows
high, normal, low, click on cupsd and select Low and keep looking at htop and see if your CPU resources lowers, and please report back, if that works, it would be possible to fix it
I don't have that issue so I can't test it.
Thanks
ttuuxxx

That changes the priority, which isn't the same as telling it that it can't use the CPU as hard. What it does is tell it that it has to back off when process with a higher priority want to use the CPU. So the CPU usage should still be at 100%, but when you attempt to start Seamonkey, for example, Seamonkey should then get to use a larger percentage of the CPU cycles than Cups. But anything not being used by higher priority programs would then be snatched up by Cups to do whatever it's doing.

I haven't been keeping track of what's going on with Cups. I am pretty sure that somebody posted a link to a CUPS changelog that mentioned fixing the problem in the development tree though, so I'm not sure why there has been so much confusion and such about it since. Either use what we have, use what they have so far, wait for them to finish, or attempt to isolate the changes that they made to fix it and apply them to the source of the version used in Puppy and recompile it (I assume that Cups is using SVN so that it should be possible to do that).

If the "just live with it" option is taken, then look into adding a 'nice' or 'renice' command into /etc/init.d/cups by changing this:

Code: Select all

${exec_prefix}/sbin/cupsd

to

Code: Select all

nice -n 20 ${exec_prefix}/sbin/cupsd

I don't know if that will actually help, but I think it might.

#388 Post by **Pizzasgood** » Thu 26 Mar 2009, 05:44

ttuuxxx wrote:
WhoDo wrote:
Béèm wrote:I applied the pet in the link I referenced and I don't have the problem anymore. Using SeaMonkey 1.1.15
Ok, I'll have another look at it. I thought the pet was for touchpad issues only? I didn't realise it also resolved the INS key bug, which was supposed to require a recompile with --enable-debug option. I'll give it a try though. Glad your problem is solved.
Hey whodo that fix is gtk, It also comes with a bunch of devs that can be removed, I'm not sure if the .so,exec files are stripped,Theres a demo that can be removed, hmmm here use this one for your test, I removed most of the extras also If I was you just rip apart the pet, take the libs, and try each one out, and test each one, Then once it stops crashing, update only the 1 lib thats actually fixing it, or it might add a lot to the iso. 2MB compressed 4MB extracted.
ttuuxxx

http://www.murga-linux.com/puppy/viewto ... 219#288219

disciple wrote:If you don't want to fill up your save file by installing the whole .pet you can extract it manually and just copy libgtk-x11-2.0.so.0.1200.1 and libgdk-x11-2.0.so.0.1200.1 over your existing copies.

trio · #389 Post by **trio** » Thu 26 Mar 2009, 08:13

Whodo

This maybe sound stupid, I have no idea how it works, if you just simply try to use initird.gz from kirk's pupplet (fatdog III?) rather than wait for his response..or it can't be done that way? Just an idea...thanks

capoverde · #390 Post by **capoverde** » Thu 26 Mar 2009, 09:23

Don't know if this has been posted already: setting the upper tray to "Hide" in Ptray it still pops up -- hiding works only for the side trays (XOrg running; will test with XVesa on another box).
Also, with XOrg the keyboard shorcut to open a terminal inside a directory (<'>) doesn't work -- not that it matters much, but this is an old unsolved issue, most probably a tricky one since all other shortcuts seem to work.

* I'd personally set Geany and Leafpad with line or word wrap active by default, though it's easy to switch it on. Best for newbies, too. Minor issues already...

01micko · #391 Post by **01micko** » Thu 26 Mar 2009, 09:26

capoverde wrote: ......... * I'd personally set Geany and Leafpad with line or word wrap active by default, though it's easy to switch it on. Best for newbies, too.

Yes

pa_mcclamrock · #392 Post by **pa_mcclamrock** » Thu 26 Mar 2009, 09:28

Update on debugging Remaster Express: I figured the problem probably had something to do with using up too much RAM, so I looked at the "Memory" section of System --> HardInfo while running Remaster Express. Sure enough, for some reason I don't yet know, the program was using up an incredibly large amount of free RAM (over 500 MB on my system!!) while running mksquashfs and showing the progressbar in Step Three. I then ran the previous remaster script, and it didn't use up nearly that much RAM.

I'm going to have to hack out the code that makes the progressbar work, starting almost at once after I post this message. (Fortunately I've got an alternate, simpler way of keeping the GUI responsive while running mksquashfs, and checking to see when mksquashfs gets done; I already use the alternate way, though not with mksquashfs, in another program I wrote.) This, I earnestly hope, will solve the problem--and I'll test the program without the progressbar, with HardInfo running, to make sure it does! Then I'll submit the no-progressbar version really fast for inclusion in Puppy 4.2 Final.

P.S. I just read that Gnocl 0.9.94 has been released--is it too late to get it into 4.2 as a version upgrade?

WhoDo · #393 Post by **WhoDo** » Thu 26 Mar 2009, 11:19

pa_mcclamrock wrote:P.S. I just read that Gnocl 0.9.94 has been released--is it too late to get it into 4.2 as a version upgrade?

Yes.

WhoDo · #394 Post by **WhoDo** » Thu 26 Mar 2009, 11:28

trio wrote:This maybe sound stupid, I have no idea how it works, if you just simply try to use initird.gz from kirk's pupplet (fatdog III?) rather than wait for his response..or it can't be done that way? Just an idea...thanks

Can't be done that way, trio ... at least not without risking a break or two here and there. We'll see; kirk did say only "two days".

pa_mcclamrock · #395 Post by **pa_mcclamrock** » Thu 26 Mar 2009, 11:37

Further update on Remaster Express: I rebooted, checked memory, and ran Remaster Express; rebooted again, checked memory, and ran remasterpup2. Both of them reduced free memory by approximately the same very large amount. Out of a total 2 GB, I started with about 1.73 GB free each time, and ended up with about 1.15 GB free each time when mksquashfs got done running (that's at the end of Step Three in Remaster Express, and the point where /tmp/root has been created in remasterpup2). I'm thinking the amount of free memory was not further reduced by much when I previously ran remasterpup2 without rebooting after I ran Remaster Express, and that misled me into thinking remasterpup2 didn't reduce the free memory by as much.

It now appears to me that mksquashfs, no matter how it's invoked, is responsible for the very large decrease in free memory. This would explain why one user reported that remasterpup2 also hung for him (the other apparently hadn't tried it). At this point, since I can't fix the apparent mksquashfs-memory-hogging problem, it appears that about all I can do before the release of 4.2 Final is to put in a prominent warning, telling users that Remaster Express may hang and need to be killed (with instructions) if they don't have enough free memory. Would you like me to do that? If so, how much memory should I say they should have?

WhoDo · #396 Post by **WhoDo** » Thu 26 Mar 2009, 12:19

pa_mcclamrock wrote:It now appears to me that mksquashfs, no matter how it's invoked, is responsible for the very large decrease in free memory. This would explain why one user reported that remasterpup2 also hung for him (the other apparently hadn't tried it). At this point, since I can't fix the apparent mksquashfs-memory-hogging problem, it appears that about all I can do before the release of 4.2 Final is to put in a prominent warning, telling users that Remaster Express may hang and need to be killed (with instructions) if they don't have enough free memory. Would you like me to do that? If so, how much memory should I say they should have?

Could it be run on RAM-challenged machines IF they have sufficient SWAP? I don't like telling users of Puppy that they can't remaster with less than 1Gb of physical RAM, which is what it seems we're looking at here.

technosaurus · #397 Post by **technosaurus** » Thu 26 Mar 2009, 13:06

If memory serves me correctly, remastering can take up to about three times the size of the final iso (including swap).

`f00 · #398 Post by **`f00** » Thu 26 Mar 2009, 19:16

{sidebar} a challenge may also be viewed as an invitation by some, lol

pretty unfortunate if it must be fast RAM on the mobo (HDD or thumbdrive is usually pretty doable for almost anyone, tho)

This time on:
machine i686
processor Pentium III @1GHz
RAM 512Mb
vidcard ATI 9200SE (PCI plain, selected in BIOS pre-boot)
display 17" crt (aged EV700)
Pupmedia CD-RW (using 'some' of i586 savedirs..)

Obviously, xorg.conf switched over to new vidcard at boot - same issue as noted with xorg (when dealing with vidcard with more than one vga out, as many aftermarket vidcards have as a standard), namely the line in xorg.conf, section server layout

Code: Select all

#	Screen      1  "Screen1" RightOf "Screen0"

must be commented out to keep the maus-pointer from disappearing 'offscreen' .. even though there is only one monitor on this card, xorg seems perhaps to be interpreting the card's two vga outs as 'screens' even if there is no display connected to the 2nd vga out.

Nice change from the previous hwr to have the soundcard recognized and pretty much ready to go (SB0350) once I run alsamixer to unmute and set levels..) .. CD in .. %#&@%#&@%#&@gxine (note to self, gxine freezing the machine a nasty local problem if it automounts on audioCD and attempts by default to access a db that's not accessible .. likewise for dumb browsers that attempt to access online resources or start on internet search page .. but that's just me and my silly offline unconnectedness in general .. nevermind, that's for the OS of the future that will have an awareness that it's not connected...) .. Enh, sound/players is one of those things that can freeze up the comp, it seems

(even with local sources).

Otherwise, it seems pretty much hunky-dory (may have to fiddle about in /etc/X11 dir -

two card-specific files there now - one for ATI_MACH64 and one for ATI_RADEON_920042001)

Flash858 · #399 Post by **Flash858** » Thu 26 Mar 2009, 19:20

OK then...Now I have lost my Menu bar in JWM (It just went "poof" when I changed my desktop icons) - Works fine in ICE, and I have done the obvious (restart X, reboot, reinstall JWM UNDER ICE, etc...).

It acutally looks kinda cool, but minimized programs are an issue...

Any clues?

pa_mcclamrock · #400 Post by **pa_mcclamrock** » Thu 26 Mar 2009, 21:38

technosaurus wrote:If memory serves me correctly, remastering can take up to about three times the size of the final iso (including swap).

Do you remember why? Like, is it because the uncompressed files are copied in RAM before they're compressed into the squashfile, or what?

(old)Puppy Linux Discussion Forum

(old)Puppy Linux Discussion Forum

Puppy 4.2 RC4 Deep Thought - Bugs & Fixes

Re: cupsd using almost 100% cpu in htop

Re: Insert key in SeaMonkey issue