Puppy 4.2 RC1 Deep Thought - Bugs & Fixes

[ttuuxxx]

Has anyone besides me had a shut-down problem(black screen0 when shutting down from an ICEwm session. jwm works fine, but I get black screen in ICEwm. My monitor indicator light also changes from green (normal on indication), to orange (no input signal). Is this a known problem with a solution or have I found a new 4.2 bug? Only way to recover is pressing Off and restarting. If it helps I am running live on a Compaq 5002us 1 gig swap 512 memory AMD Duron 800.
Thanks stray

Never had that problem, Could you try it again, but this time run Pprocess and kill all the conky stuff, so all the Pwidgets etc is gone and then try it, I'm wondering if its another timing issue.
ttuuxxx

[ttuuxxx]

I have an acer X193W lcd monitor (1440x900) and an intel 82946 video card. Xorg crashes the computer as soon as the wizard finishes. Xvesa works but is too blurred. How can I get xorg working?

In Unnamed its the opposite except that xorg is usable because it's less blurred.

Maybe try a different setting, I have a 19"LCD square, and usually I use 1024x768, and 24 bit, just the first ones listed on the xorg screen.
Never use the test it always crashes for me also, but when I select 1024x768 it works and then when I test the screen size its registered as 1280x1024 so it doesn't actually load at 1024x768 it loads at 1280x1024 which is perfect, It seams to fool xorg somehow.
ttuuxxx

[Flash]

Has anyone besides me had a shut-down problem(black screen0 when shutting down from an ICEwm session? JWM works fine, but I get black screen in ICEwm. My monitor indicator light also changes from green (normal on indication), to orange (no input signal). Is this a known problem with a solution or have I found a new 4.2 bug? Only way to recover is pressing Off and restarting. If it helps I am running live on a Compaq 5002us 1 gig swap 512 memory AMD Duron 800.
Thanks stray

I had what sounds like a similar problem with alpha 2 (I think it was) when shutting down my multisession DVD. The screen went black and never asked if I wanted to save the session. In my case, the LCD didn't shut down for lack of a signal. The problem went away in the next alpha. I never found out what caused it.

[big_bass]

straypup wrote:
Has anyone besides me had a shut-down problem(black screen0 when shutting down from an ICEwm session. jwm works fine, but I get black screen in ICEwm.

yeah I had that problem too that's why I gave up on icewm
a long time ago *(icewm does look nice though)

I found a fix for it and applied it but never went back to icewm
JWM is fine with me

MU fix
http://www.murga-linux.com/puppy/viewto ... b7d4b53a54


or just add a comment to line # 25

leafpad /usr/X11R7/bin/restartwm

so it looks like this :
#exec killall -9 X #v1.0.7 make sure X is killed.



thank Mark for that one

big_bass

[WhoDo]

the attached wget-1.11.4 is now configured correctly.

Got it! Thanks, aragon. I had a prescott-compiled version from coolpup but the prescott-optimised packages have been throwing bugs so a plain vanilla compile is appreciated.

Cheers

[WhoDo]

I enclose 3 plasma images I created without too much effort that are not even 4k combined

Added to package along with the ttuuxxx-optimised Aurora-Australis.jpg at 40Kb in lieu of 189kb.

[WhoDo]

Boot screen still shows BETA.

An RC is a Beta ... just a Beta with no known showstopper bugs and only a very few niggling issues (unless more are discovered afresh, of course). I have released RC1 because I know that the major showstoppers have been all but resolved - especially CUPS printing.

While I have the opportunity, prompted by your report, I'd like to explain to the community why I released this version as RC1. Because the showstoppers are gone, and other niggles are being worked on, an RC1 release is an opportunity for the rest of the community to join the testing effort. Normally I would discourage the average user from entering the Alpha or Beta testing program, but would encourage them to test at RC1. I hope that makes my reasoning clear.

On tower where printing worked under 3.x and 4.1x, will not print on LPT "Unable to open device file ldevlp0 : Permission Denied". Was able to print using USB.

Patriot has posted a patch that fixes the issue. ttuxxx used some of Patriot's earlier code, with due attribution, to post a partial fix too but I believe Patriot's UDEV solution is the more complete one and that will therefore go into the next release.

[WhoDo]

Does this use Xorg 7.4? Is there a devx for this Puppy?

"No" to the first question and "Not yet" to the second. I will not be upgrading to Xorg 7.4 in this release. Devx_420 will not be created until I release the Final.

[WhoDo]

1. Trash icon located about mid screen.

Explained elsewhere (see Beta2 thread). I will be revisiting the pinboard positioning and inclusions before I get to a Final release stage.

2. Pwidgets 1.7-1 has introduced some problems that were not present with 1.6 in the last 4.2 beta ie. weather widgets don't work, also after adding rss feeds, slideshow, & todays date, widgets will not respond to any changes. Additional clock skins are also missing. I'll post details to pwidgits thread. EDIT: Switched to icewm and weather widget does work and am able to change widgets again.

I'll leave that to Zigbert and his team to investigate. The weather widget was enabled by default accidentally. Zigbert or I will fix for the next release.

3. Printing works, somewhat, CUPS recognized my HP5280 printer and did provide a driver selection that allows me to print a test page. However, when I tried to update the printer drivers with later hpijs .pets, CUPS would lock up after selecting modify printer.

I note that ttuuxxx is already assisting HP users with their CUPS problems in the latest release.

[WhoDo]

rename the enclosed file pwget.zip to just pwget and replace the file /usr/local/Pwget/pwget

Done for next release.

[WhoDo]

Intel(R) PRO/Wireless 2915ABG Net is not being recognized on this IBM Thinkpad R52... Pre 4.2 it is recognized. ....

I have not changed the wireless driver sets from 4.1.2 so I have no idea why this should be the case. Maybe tempestuous or one of the more wireless-savvy guys can provide an answer?

[WhoDo]

"the way it sits.", ESP Ghostscript 8.15.2 is being __misrepresented__ as "GPL Ghostscript 8.64".

I have no concern whatsoever about space issues, but lying, in the forthcoming release notes, PET package names, whathaveyou, is __wrong__.

Proper, correct, accurate communication/documentation is the __right__ thing to do, Always.

So, please, tell the world, and WhoDo, the truth about which Ghostscript is Really going to be in Puppy Linux 4.2. Ok?

Ok, this really has gone far enough! I really hate the way some guys bandy around the "L" word when there is an inaccuracy or a mistake or some other equally trivial explanation for something other than a LIE. Making that accusation in such a cavalier fashion is entirely counterproductive IMHO.

In some respects you are BOTH right and in others you are BOTH wrong.

The Ghostscript version that appears in RC1 is 8.15.2 - I unilaterally (i.e. without anyone else knowing beforehand) went back to that in an attempt to help ttuuxxx resolve the CUPS printing issues. That version of ghostscript was the ONLY significant difference between the working CUPS and a boat anchor!

I think ttuuxxx was probably under the impression that I included BOTH the earlier 8-15.2 version AND the updated libraries of the 8-64 version. I did not. I was trying to get us back to where the newer version of CUPS had worked in testing but without the added complication of two versions of ghostscript overlaying one another.

Does that make ME a "liar" GG? Or simply guilty of an oversight in failing to let people know what I'd done before I "rushed away" for a brief but overdue holiday weekend? IMHO, the term "liar" ought to be kept locked in a steamer trunk with 4 or 5 different padlocks and not opened until 4 or 5 different key holders agree that it should be let out. What do you say, eh?

OTOH, ttuuxxx is correct when he guestimated that CUPS in RC1 was about 96% new stuff and 4% old stuff. That it is. Does it really matter if it works? I'll get the release notes up to date when I'm ready to release, not before. In the meantime I don't think any of the package authors will be beating on my door demanding recognition for their work.

I understand your pedantic insistance on accuracy in reporting GG, but please save it until we are ready to launch the product on the rest of the Linux world. Until then, it's just a minor glitch in an otherwise very complex and taxing program of development and testing.

[WhoDo]

This is really dismaying, in contrast to CUPS 1.1.23 on Puppy 4.1.1 which works perfectly.

Just picking a nit here, David, but CUPS 1.1.23 did NOT work "perfectly", or at all, for me. I kept getting 404 not found errors when attempting to load the Admin so I could even begin to add printers. Even when I got past that, and added my printer, it wouldn't work because of poor support for USB printers. My guess is that 90%+ of modern printers are USB, so that's a showstopper IMHO. That's why I asked ttuuxxx to look at CUPS in the first place. Like all good debuggers, the first thing he did was look for the latest official release, to eliminate all the bugs that had already been fixed. It has been a long and painful process but going back to CUPS 1.1.23 was never an option in my mind.

I'm not picking on you, David. You've done too much good here to deserve that. I'm just gently (I hope) correcting the record.

Cheers

[01micko]

G'day WhoDo

Yeah there are still some wireless bugs, I think HairyWill's was similar to minesota's problem. Mine. as you've read by now is rtl8187. Something got corrupted along the line somewhere.

Cheers

[WhoDo]

And here's the first comparatively little bug I've found: when I closed the ROX-Filer window in which I was examining the permissions of /dev/lp0, I found that the Pwidgets clock now had a picturesque but incongruous background consisting of hard-disk icons and device names from ROX-Filer!

I think this is a problem related to pseudo-transparency with the Xonclock and the background setter. I don't know if Zigbert and the guys have resolved it already as I note that Pwidgets is now up to 1.19-1 when RC1 had 1.17-1, so we'll see, but I suggest it has to do with /root/.config/wallpaper/bg_img not being set to the default for ROX, which is /root/Choices/ROX-background.jpg

Hope that helps

[Lobster]

. . . very complex and taxing program of development and testing.

Have a cute animal break
http://www.murga-linux.com/puppy/viewto ... 175#279175

repeat as required.

I can confirm that RC1 is certainly usable - because I am using it
(I may test for a few hours) but as soon as something crashes, I have to revert to a working system.
Been using RC1 since release. It is going Well


We are Puppy.
A cuter Linux

[WhoDo]

- /etc/ld.so.conf is still missing an entry for /usr/local/lib

Not according to my unleashed tree. If /etc/ld.so.conf is being overwritten by something I'll have to go chasing the culprit.

- /usr/local/sbin is missing in PATH

Again, not according to /etc/rc.d/rc.sysinit in my unleashed tree. Could this be getting overwritten too? I will investigate.

- gadmin-rsync is the old version new: http://www.murga-linux.com/puppy/viewto ... 518#271518 Note: Old one does not work!

Yep. Added it to the tree but must have overlooked adding it to the list of "on" packages. I will fix for the next release.

- xarchive and xarchiver are still both in. is there anything left to do, that xarchiver becomes default and xarchive leaves?

Yep. Find out everything that calls on xarchive and repoint it to xarchiver instead, then hope it still works. The perfect example is Pupzip, which calls xarchive. I haven't tried to repoint it to xarchiver, but that would be a good place to start.

- i don't use neither xchat or meebo, but do we need xchat if we have the online-version?

Yes. The online version is only a pointer so doesn't take any space and gives a nice, user-friendly interface for MSN, etc. Xchat is the main access to our #IRC channel and is preconfigured to jump there because we plan to make it our free online help system for newbs.

- pplog didn't run. i also could not open 'localhost' or '127.0.0.1'. Everything is "403" if i look at pprocess, hiawatha is running as 'nobody'. might that be the problem?

This bugs the life out of me (pun intended) because I can start PPLOG, follow the prompt to fire up Hiawatha and then all is well. Why should it work on my machine and not on yours? I have no idea, since I can't duplicate the problem in the first place. I'm sorry aragon, but that needs to be debugged by someone who actually has the problem and can then pass on the solution for everyone else.

Cheers, and keep on testing and reporting ... it is much appreciated.

[WhoDo]

There are still a lot of bugs for a release candidate. Whats all the rush to release? Lets make sure we get it right first.

Hear, hear!

Please read my explanation for the release of RC1 instead of Beta3. Release candidates are not expected to be bug free, just showstopper bug free, with fixes in train for the others. They are an opportunity to expose the new product to a wider testing audience to see if new problems are exposed in the process. Having squashed the major bugs from the Alpha and Beta series I felt the move was entirely justified.

@ecomoney et al - This release of Puppy has already had more testing releases than any previous official Puppy release except 2.15CE - which had a greater need of testing due to my own inexperience with buidling Puppy from the Unleashed tree. Progress has been steady but inexorable. The Deep Thought product is now ready for wider testing IMHO. I do not expect this version will be followed with a dotpoint-1 update as have some previous official versions either. Barry had to occasionally resort to actual releases in order to get things tested - witness 2.16-1 and 2.17-1 for example. This time around we have a committed and knowledgable pool of community Alpha and Beta testers who have brought us this far without the need to "rush" an actual release out into the wider community to find these bugs.

Please, enough of the recriminations over the choices I have made. They were ALL MY CHOICES and I will live or die by the end result. We don't have that "end result" yet, so less rushing to judgement would certainly be helpful.

PETget still takes far too long on the "updating menus" stage when installing or uninstalling. (EDIT: For some reason, this appears to be true only before you create a pup_save file and reboot--which I just did. Then I installed a couple of .PETs, and "updating menus" only took a few seconds, as it should!)

Let's see if things are better in the next release, since Zigbert has made further modifications to the fixmenus script at the heart of the menu update process.

PETget's instruction to "restart the window manager" (which there isn't any obvious way to do without restarting the X server) is still there, and it appears that you don't even really need to click "Refresh Menus" any more.

Not true. The need varies between JWM and Icewm. In JWM you always have to refresh the window manager, but in Icewm you don't. Taking account of both systems is a bit more tricky than just allowing for one process, but worth it in the overall scheme of things. Maybe we'll move to a single WM again down the track - either Icewm or the newly updated JWM with Patriot's valuable input - but for the moment I'm happy to have both so our users have a real choice of environment and can make an informed decision for later on when Woof builds are the norm.

Restore "Restart IceWM" to the IceWM menu!

That's what Refresh menus does anyway, as ttuuxxx has pointed out.

In SeaMonkey, the View --> Show/Hide menu has "Tab Bar" grayed out, which made me think maybe the tabbed browsing function was hacked out of the new SeaMonkey in a misguided effort to save space--until I saw that the tab bar will appear if you click File --> New --> Navigator Tab.

As ttuuxxx has suggested, this one is probably a Seamonkey issue rather than a Puppy issue. Nevertheless, Seamonkey 1.1.14 is the most stable version of the Seamonkey line to have surfaced so far, and there is little to choose between it and Firepup for performance alone. With ttuuxxx adopting Seamonkey for Puppy, we may well get a Seapuppy version down the track, but not if people keep throwing cold water on his enthusiasm! (Not you David, but an allusion to earlier posts in this thread).

Hope that helps.

[jabu2]

Minnesota wrote:
Intel(R) PRO/Wireless 2915ABG Net is not being recognized on this IBM Thinkpad R52... Pre 4.2 it is recognized. .... Sad
-------------------------------------

Same problem on Fujitsu lifebook 2200 .


412 setup wizard recognises eth0 and eth1 interfaces

4.2rc1 recognises only eth0 (my blue ethernet cable connection/trip hazard)

in 412 eth1 = IPW 2200 intel wireless 2915ABG and both ethernet and wireless interfaces/connections work perfectly

In 42rc1 this wireless 2915ABG driver does show in the drop down list in Puppy's internet LAN wizard but can't be loaded (by me anyway, as I am not a coder or linux expert).

Help please - no wireless, no real go! I am back on 412

Otherwise all seems OK, and thanks to all the team!

[WhoDo]

rc1 full install:

Thanks, Dave. I don't propose to offer solutions to these here as I'm sure Zigbert and his team at the Pwidgets thread are already dealing with them. Just wanted to acknowledge your testing inpu.

Cheers