|
Puppy developer news:  Older news |  Later news |
Puppy version 2.13 releasedWe release a new Puppy every 6 - 8 weeks, always with a huge list of new features, too much to summarise in one or two sentences. Basically though, there are major improvements in the user interface and general usability, including upgrading of fonts, ROX-Filer, Video Wizard and bootup and shutdown scripts. New applications developed by Puppy-enthusiasts are making their debut, namely Soxgui, XkbConfigurationManager and PuppyBackup. Other Puppy-applications are significantly upgraded. Get the full story in the release notes: http://www.puppylinux.com/download/release-2.13.htm Download Puppy from here: http://distro.ibiblio.org/pub/linux/dis ... uppylinux/ Lobster
Congrats Barry and everyone :)mirrors and torrents here please http://puppylinux.org/wikka/Puppy213 Raffy
Congrats, Barry and contributors!Are we headed for the PXA soon? Sage
Ah,
Xorg is much better now. However, the <OK> button is missing from
the test box, requiring the 40s wait before it proceeds!BarryK
The
<OK> button is deliberately missing, as on some hardware it
causes the PC to hang. The dialog window now instructs you to press
CTRL-ALT-BACKSPACE to end the test (which, for unknown reason does not
cause a hang on that hardware).Sage
OK,
thanks, Barry. Saw the messages but assumed they were just when things
went pear-shaped! One gets conditioned by expectations.SquiddlyDiddly
Dude,You ROCK! Thanks for the iso but Ibiblio is hopelessly slow any other mirrors? Lobster
mirrorftp://ftp.oss.cc.gatech.edu/pub/linux/distributions/puppylinux/ Nathan
Hey, when di cvs get removed from the devx squashfile? That's aggrevating.Bill St. Clair (billstclair<at>gmail.com)
Mirror:http://s3.amazonaws.com/puppy/index.html BarryK
These guys are also a mirror for ibiblio:http://ftp.nluug.nl/ftp/pub/os/Linux/distr/puppylinux/ John Roberts
Barry, Congratulations!Super release!!!. The UK mirror was pretty fast too (ab. 33kb/s). The new DeJa fonts together with the new keyboard layout configurator trully rock!!! This is a super multi-lang Puppy and one more solid argument for spreading the cult Greatest New Year's gift!!! Many thanks JR Looks fabulous! Xorg works great (at least with my hardware) Looks like the test section for Xorg could use an ok button though. Either way Puppy looks like it is getting better and better! Great Work! Everyone needs a Puppy! MU
there were reported issues with file-associations (to launch a specific
program when you click on a file in the Rox filemanager).I made a hotfix based on what I use in Muppylinux: http://murga-linux.com/puppy/viewtopic.php?t=14125 Might have to be optimized. Mark Pukikuki
Hi, do we get a proper "fullscreen" mode in media player now? ThanksBarryK
Gxine fullscreen? ...yes, works fine now  Goble
Puppy really looks great. Already did in 2.12 version and the latest version still managed to improve it in big leap.Barry. Did you use T2 compiling environment to compile 2.13 Puppy? I guess you did. http://www.puppylinux.com/test/ page has a download link to T2 Puppy 2.10 version. Is there such T2-Puppy-2.13 config package available somewhere? I'd like to try to compile the whole distro using native EPIA C3 processor support. Now when EPIOS distro seems to be lagging behind, I guess it would make Puppy the only "EPIA C3" native distro Never done "full Epia C3 processor type compilation" with any other distro, only with self-compiled applications and Kernel but never for the whole distro. This is yet-another unique feature of Puppy distro which does much more than many other distro but still being compact, easy to use and powerful. Dougal
Goble: 2.13 uses the same basic filesystem as 2.10, so that T2 link is ok. Good luck...BarryK
The base packages for Puppy were only compiled once using T2, and that resulted in Puppy version 2.10.I have not used T2 since then. Packages that have been upgraded plus many extra packages were compiled within Puppy since v2.10. I do plan to go back to T2 and do a complete recompile, probably when we have to do a major upgrade of the GTK libs to 2.10, Xorg to 7.1, etc. Then, I'll create another T2-package for compiling Puppy. You can use the current T2-puppy package, but do note that it will only give you raw binary packages as used in Puppy, you won't be able to create a new Puppy without a lot of work. This is described in the T2-Puppy webpage. PaulBx1
Barry, when you recompile, please give a thought to the encryption
packages (e.g. dm-crypt). I don't know a whole lot about this, but user
John Doe might have some recommendations along that line.white_christmas
the new rox is soooo great!! THANK YOU |
PuppyBackup now reached v0.6.3http://www.murga-linux.com/puppy/viewtopic.php?t=10975 It is now zero minutes to twelve as far as releasing the next Puppy is concerned. I either upload it today while I have access to broadband, or wait another week. And I don't want to wait another week, want to get 2.13 out and start work on 2.14! I don't want to make any further changes, in case there is some unforseen side-effect. Dougal and plinej have improved Pmount, but will hold that one over. Ditto for many other things. Regarding getting my own broadband, I'm just waiting on the installer. He took over a year before visiting my place to measure the wireless signal strength. Now we know I need direct 2-way satellite, and I'm just waiting for him to come by my way again (I'll get on the phone and pester him some more when I get home -- maybe I should post his phone number here ) Chris
(Now
we know I need direct 2-way satellite, and I'm just waiting for him to
come by my way again (I'll get on the phone and pester him some more
when I get home -- maybe I should post his phone number here)Yes a good idea! We can just ring him up with a Happy New Year from Puppy users. Chris Fred Doolie
Yes! We'll lean on him for you.zigbert
I'm
thankful for your upgrade of PuppyBackup to 0.6.3. The version 0.6 had
a nearly total rewrite of code, and was buggy. Sorry for giving you the
extra work with it.Chris
After fixing a newish Compaq Craptop tonight, what a pleasureto run Puppy 2.13 Final now here at home. From when I burnt the ISO in Puppy 1.07, rebooted, setup the Internet, saved back a file, rebooted and now posting from Puppy..under 10 minutes complete...easy peasy And of course I ran the Compaq with 2.13 Beta2, which ran like a champ, unlike the Redmond OS and Mr Nortons Security special...certainly this is a case of where a DOG got eaten by a doggy. Now go back to your kennel for a well earned rest. Distrowatch had you up on top of the pile too. Once again I award the gold star to BK. Thanks Chris |
Xorg Video Wizard improved (again)One problem is that if you choose the <FIX> button and test a video mode, on at least half the PCs I tested on it hung the PC after testing. I have now fixed that, only have access to four PCs right now, but it now works fine on them. I also simplified the dialog window for choosing alternative screen refresh frequencies. Raffy suggested that <FIX> is perhaps not the best name for the button. It is now named <TEST>. Get it from: http://www.puppylinux.com/forum/azbb.php?1167644023 archwndas
Barry,I have tested the beta12, on a Apple-MacBook Pro. It boots but the first problem I have noticed is that it doesn't see the wireless driver. The driver is called Airport Extreme. I have done a little goggle search and I found in gentoo forums something like: a module named bcm43xx.ko. I am not sure if this is the correct one, but even this module is missing from the list of modules suggested by the network set up script. I would like also to report that the network setup script is a bit worse than the older one. What makes it worse is that it doesn't let you set up wireless and ethernet as different choices but it combines them to one. This is not a good idea. The Linux kernel has different choices for wireless and different for ethernet why the network scripts should combine them to one? SquiddlyDiddly
What happened to taking a little break?????BarryK
I did have a break for about half a day. Will take a longer one soon after 2.13 is out. It just isn't possible to get Xorg to automatically come up with a nice display on every possible combination of video hardware. The idea of a single window in which you choose your desired screen resolution then hit <OK> button and up she comes looking real nice ...well, that's still a dream. So, I've made one more change to the Wizard, have restored the first dialog window to a choice between Xorg and Xvesa, as it is in Pup 2.12 and earlier. Feedback is that this is required so you can choose not to go into the Xorg probing at all. |
Full sources for 2.6.18.1 kernel and modules uploadedhttp://www.puptrix.org/sources/kernel-2.6.18.1/ craftybytes
Barry,Will the full 'patched' sources for kernel 2.6.18.1 and modules be included to the v2.13 Unleashed CD ..? Barry Kauler
The Unleashed CD has the kernel patched source (40M) but not the modules.There isn't room for much more on the CD. But, the modules are smaller so can be downloaded more easily individually. craftybytes
Thanks Barry.. Just checked and yes the modules are much smaller..Are they all required if I use the kernel source for compiling purposes or are they 'selective' based on what I need to compile..? Hope you had a very enjoyable Xmas..? Also - have you received my package to you as yet..? And finally - happy New Year and the best of health and family for 2007.. kirk
Is this the same source as for 2.12? Also, if you post the source as a .sfs file, it makes it very convenient for the liveCD/frugal user. But it's a bit bigger, the one I'm using from 2.12 is 53MB. BarryK
Yes, 2.12 and 2.13 use the same kernel.Only the kernel source package is needed for compiling the kernel -- the other packages are just extra modules. pablo
Are the wifi drivers patched to support packet injection with Aircrack-ng?.Thanks. |
Puppy 2.13BETA2 post-release notesIf you tested BETA1 and then do "puppy pfix=clean" to force a simulated version update for BETA2, one thing that slips through the net is 'rxvt'. This is because its modify date is very old, compared with the rxvt script that is part of the Mrxvt package. You will need to manually copy /initrd/pup_ro2/usr/bin/rxvt to /usr/bin/rxvt. 2. Puppy will load a native wireless network driver at bootup, if autodetected. This is a problem if you do not want to use it and want to use Ndiswrapper instead. The Network Wizard will show the wireless interface is already claimed by the native driver. The script /etc/rc.d/rc.modules has a variable SKIPLIST, which is a list of drivers NOT to load at bootup -- you can add to that manually. For example, my PC loads the 'acx' driver module, so I added that to the list -- make sure it is inserted with a space either side. Probably what we need is for the Network Wizard to be able to do that automatically. 3. I didn't announce it before, but the Universal Installer had a bug when install to HD and create a USB 'boot disk'. The USB Flash 'boot disk' did not boot, but that is fixed in 2.13BETA2. rarsa
Do you mean having a button on the wizard to unload the module and "black list" it?The idea of having a button to unload the module had already been floated around but I though of it as low priority as there is a workaround for that (reload the module and then click "don't save"). Under what circumnstances would the user prefer ndiswrapper to the native driver automatically loaded? Barry Kauler
The native driver doesn't always work.Yes, in the Wizard you need to be able to choose to unload a native driver and use Ndiswrapper. Then the native driver would be blacklisted, which means it gets added to SKIPLIST so it will never even load at bootup. |
Puppy version 2.13BETA2 availableDownload: http://www.puppylinux.com/test/ Updated release notes: http://www.puppylinux.com/download/release-2.13.htm Right now it's 5.30pm Sunday 31st December here in Western Australia (GMT+8.00), and I thought I would upload BETA2 as a New Year pressie for our hard-core Puppy addicts! This is looking pretty good. If anyone wants to give it a quick spin, please do and hopefully the final can come out soon. Please note, if you have tested BETA1 and will be using the same pup_save.2fs/3fs file, be sure to enter boot parameter "puppy pfix=clean" to force a simulated version upgrade. Eddie Maddox (greatnessguru<at>gmail.com)
Release notes for Puppy Linux version 2.13http://www.puppylinux.com/download/release-2.13.htm SeaMonkey upgraded to v1.0.6. v1.0.6, or v1.0.7? I thought it was v1.0.7. Thanks, Eddie pakt
Thanks for the 'New Year pressie'Writing from the eBox right now and running great: video (new xorgwizard) and audio working fine. Only installed Opera 9.02 to make browsing bearable Thanks for a great year and enjoy your time off - you really deserve it Paul Eddie Maddox (greatnessguru<at>gmail.com)
Ahh, yes. Found it...Feature freeze for 2.13BETA http://puppylinux.comfirms.com/news/index. Wednesday, December 20, 2006, 07:33 PM I'll compile SeaMonkey 1.0.7 tonight, ... SM 1.0.7 it is then. Right? Thank you, Eddie BarryK
No, 2.13 still has SeaMonkey 1.0.6.GrumpyWolfe (grumpywolfe<at>gmail.com)
Have it running on my box by cd. Running fine have to ajust desktop to the left. But have internet.Thanks a hole lot Berry Sage
Nice one, Barry. Everything - even Seamonkey - seems a lot faster. Would be even better with Opera instead.Xorg setup is rougher than previously - ended up in 640x480 a couple of times, and the menu reset didn't work on one of those occasions. Probably nothing wrong, as such, just different and less smooth. Knoppix 5.1 with aufs also seems to run nicely. Eddie Maddox (greatnessguru<at>gmail.com)
BarryK Sunday, December 31, 2006, 10:05 AM > No, 2.13 still has SeaMonkey 1.0.6. Then, perhaps there will be an 'official' SeaMonkey 1.0.7 PET package? Yes? Thanks much, Eddie vern72023 (gdcrane<at>bellsouth.net)
removed sm and xorg added opera and wine and at 68MB it runs very well
on the neoware terminal and on the emachine using a usb drive.soxgui is a great new addition george Eddie Maddox (greatnessguru<at>gmail.com)
PuppyLinuxWiki : Puppy213http://puppylinux.org/wikka/Puppy213 Final release coming soon . . . * Upgraded SeaMonkey 1.0.7 Barry, Does this need correcting to say: "1.0.6"? Also, this may need updating now: PuppyLinuxWiki : SeaMonkey http://puppylinux.org/wikka/SeaMonkey Thanks again, Eddie Barry Kauler
Eddie, I do not maintain the wiki.BlackAdder
Have checked out the upgraded ndiswrapper. The bug related to adapters needing firmware downloaded is fixedThis is being posted via a DWL-650+ that did not work with ndiswrapper on 2.12. Everything else that I have checked so far is working very nicely. Happy New Year, Barry. Enjoy your break. And Happy New Year to all Puppy fans too. Eddie Maddox (greatnessguru<at>gmail.com)
Thanks, Barry.And, great news! I just burned and booted Puppy 2.13 on an old HP Vectra VE, 64MB with a Linux swap partition and some Linux ext3 partitions. I absolutely Love this Puppy! Much better than the earlier Puppies I tried the last 2-3 years. I'm writing this from SeaMonkey: SeaMonkey 1.0.6 Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.8) Gecko/20061220 SeaMonkey/1.0.6 Great job, Barry! Thanks so much, Eddie |
Xorg Video Wizard improvedI finally managed to get the ModeLine thingy working. If the Xorg Video Wizard is run from within a running X, via the Setup menu, one of the modules is Xvidtune, which enables real-time tweaking of the displacement and proportions of the display. You can finetune the display to look exactly right, then save the setting and write it to the /etc/X11/xorg.conf file (the Xorg configuration file). The problem up until now is that it didn't work, but I suddenly realised what was wrong and now it works. I can think of a situation in which it might not work, so after 2.13 is released and you try it and it doesn't work, look at /var/log/xorg.0.log and see what error message it reports, then tell me. Another thing, Xvidtune does not work with all video chips, but you will know immediately that it doesn't work as when you use it any real-time changes that you try to make to the display will have no effect. The main thing that I've been trying to do all day though, is get the Wizard to run X with a nice display by default, without having to do any tweaking. One thing it has made me realise is how buggy the 'ddcprobe' program is, but I built workarounds into the Wizard. Tomorrow is New Years Eve, and I'll be taking some time off, so cheers everybody! RR
I'm sure thousands of Puppy fans wish you a very Happy New Year.archwndas
Happy new year Barry. Thank you very much for all that you areoffering to us for free. May everything you wish comes true, in 2007. Craig
BarryOn the Xorg track I did a hard disk install of 2.12 yesterday and encountered one small issue. On boot it detected my mouse as a ps2 I changed it to serial and finished the boot up. I then used the universal installer to do a hard disk install. Reboot off harddrive reset mouse to serial again, finish boot verify all is well and reboot. Still detects a serial mouse and will not hold the setting. Not sure why it keeps overwriting my xorg.conf file. Craig Have a Happy New Year and keep up the good work! zigbert
If
you want, there is a upgrade of PuppyBackup with fixes for Xdialog/GTK2
issues. http://www.murga-linux.com/puppy/viewtopic.php?t=10975The very best wishes for the new year to Barry and his Puppy friend. Sigmund Chris
Tomorrow is New Years Eve, and I'll be taking some time off, so cheers everybody!And about time too. For the "The hardest-working man in the OS business". Thanks for all the enjoyment you have given in the past to all the Puppy community users. Have a great 2007. Chris debernardis
Puppy New Year to all, Puppy New Year to Barry! This is going to be a great year for all you people of good will, hanging around free, libre, open source knowledge and creativity. Please accept my greetings and my thankfulness for so many hours of interest and fun! Raffy
All the hard work of 2006 will turn in big results by 2007 - Puppy's popularity has a lag! Happy New Year!capoverde
Let
me join into the Puppy-happy crowd's gratitude, wishing all the best to
you and all the unbelievably helpful Puppy team. Woof!
|
Ndiswrapper now workingAll you guys who have reported on the forum that ndiswrapper doesn't work in Puppy 2.12 and 2.13BETA ...well, you're right, it doesn't. I googled around and found this error is very common. Fortunately the ndiswrapper developers have just fixed it, in v1.33 released yesterday. Their changelog makes a note that there may be another problem when we upgrade to kernels of 2.6.20 and greater, but let's not anticipate trouble! I compiled and installed the new ndiswrapper utilities and driver, and just did the manual steps "modprobe ndiswrapper" and "iwconfig", and there was "wlan0". Running "dmesg" shows the Windows driver loaded okay, and it seems that stuff about shmmax and shmall was a red herring. I haven't tested ndiswrapper in the Network Wizard -- Rarsa, BlackAdder and other wireless networking enthusiasts can test that. I'm on track to get Pup 2.13BETA2 out by New Years Day. Fingers crossed the BETA2 is okay, and we move swiftly to a final release. I'm postponing any further work on the PET package concept, as I want to have a group chat with the main people involved in the evolution of packages in Puppy, including MU, G2 and NathanF, and figure out the way ahead that is agreeable to everybody and in particular to improve the package management experience for users. |
MUT fixed, Geany updated, Gcrontab Help button fixedWhile I was at it, I also improved the mounting of NTFS partitions. Pmount was one step ahead here -- if an NTFS partition was determined to be 'dirty' Pmount still mounted it, with a warning, whereas MUT just failed to mount it. The problem is that NTFS partitions are often 'dirty' (for example, you failed to shutdown Windows properly the last time you used it), so it is an issue of some importance (particularly if you need access to the files in a broken Windows installation). I modified MUT to mount NTFS partitions the same way as Pmount. Geany text editor has been updated to the latest, v0.10. I was informed that the Help button in Gcrontab, the GUI frontend for 'crontab' and 'crond' utilities, was broken. Now fixed. Barry Kauler
Hmmm, rendering of this blog page has gone a bit peculiar. This
happened once before, and I deleted recent posts and it came good.The blog script is a bit odd, doesn't like some character combinations -- the one that has been most annoying is the dollar character followed by numeric digits, which gets changed to something else. I looked through today's posts, can't see anything unusual -- but then, maybe I wouldn't, as the offending character combination would have got interpreted and not rendered as-is. I need to find out if there's an upgrade available. Another annoying thing -- the timeout duration for comments is far too short. Dan Bachmann (nospam<at>danbachmann.com)
This
is great news! One of the great things about Puppy is the use as a
rescue tool. Using MUT, ROX and GParted important files can be salvaged
from an unbootable or expired demo version of Windows.This improvement (able to mount dirty NTFS partitions) will be most welcome from me when that desperate moment when Windows or a hard drive starts to fail. |
New Soxgui, updated Pupdvdtool and Xwget, Rox tweakedThis is so useful, it will be in Puppy 2.13. But puppy does not have the 'soxmix' utility so I took the liberty of removing the "Mix two audio files" button from Soxgui. Jason has updated Pupdvdtool DVD ripper to version 0.5B. Ian has modified Xwget v0.5 to look better in Pup 2.13. I changed ROX-Filer 2.5 to thumbnails disabled by default. This avoids a couple of bugs reported earlier in this News blog, also is better on older slower PCs. ROX-Filer 2.5 is based on GTK2 libraries whereas our old ROX-Filer 1.2.2 (as used in all previous Puppies) is based on GTK1 -- so the new Rox is a tad slower, but disabling thumbnails counterbalances that. Rox now recognises '.log' files. Click on a '.log' file and it will open in a text editor. Jason Pline
Feel free to modify any of my scripts as you see fit Barry. I'm just happy to contribute any way I can. Jason Pline
I
already noticed in Soxgui I forgot to add "ogg" to the drop down box
and I included "svx" which should really be "8svx". I'll probably do
some more work on it tonight and tomorrow. Since Puppy has ffmpeg I'll
probably add some other audio formats and use ffmpeg for the conversion
of the unsupported sox formats. For example wma, mp2, ac3, and probably
some others. Barry Kauler
Well, as long as you don't mind about the button being removed. I would
rather the buttons are all functional without having to download
anything else.But, does 'soxmix' do anything else that might justify putting it in Puppy? It's a bit large, 270K, which was why I left it out -- at the time I thought, another mixer we don't need -- but I never examined its docs. Yes, bringing in the other conversion utilities would be very good. I was aiming to release 2.13BETA2 on New Years Day. So if you have anything by the 31st Dec I'll put it in. Dougal
Barry, it would be good to mention that right-clicking the "show hidden" button in the new Rox enables/disables thumbnails...Jason Pline
Barry, I'm not sure if soxmix does anything else. I'd just leave it out.Jason Pline
I've
got soxgui re-written with ffmpeg in there too. The version in Puppy
supports the following audio formats that aren't in sox.shn (shorten) (decode only) wma (decode only) (newer ffmpeg versions can encode as well) m4a (apple lossless) (decode only) mp2 ac3 ra mp3 The newest version of ffmpeg has several other audio formats as well. I'll repost it once I do a little more testing. |
Gutenprint bug fixed, Puppy 2.13BETAThe way I solved it is as follows. I used the Printer Wizard and setup my Epson Stylus CX5100 all-in-one printer. From a running XPDQ print manager, I chose to print a Postscript file, /usr/share/examples/ps-pdf/colorcir.ps. It seemed to work, but nothing printed. The PDQ system keeps all the files for current print jobs in /root/.printjobs/ and I open file /root/.printjobs/001.log. This reported that Ghostscript had aborted with the error message "cannot find dither matrix file". After sometime I found the file that Ghostscript wanted was /usr/share/gutenprint/5.0.0/xml/dither-matrix-2x1.xml, but it was missing from the Unleashed Gutenprint package. I supplied the missing file, and printing worked. The reason that I had not included it originally was that it is 375KB, and there are also two other files, dither-matrix-1x1.xml and dither-matrix-4x1.xml, that I had also left out as they are also about 370KB each. So, I am forced to put one of them in, but the problem is I don't know what circumstances will require the other two. I cut corners wherever possible and do not want to put in 700KB of files that aren't going to be needed. If you use Gutenprint and it doesn't work, do read the PDQ log file -- if it complains of a missing log file, you will know the solution. I only plan to put 'dither-matrix-2x1.xml' in Pup 2.13-final. I tested printing from XPDQ and from SeaMonkey, both now okay. Kirk tested with a real Postscript printer. One difference from earlier Puppy's is that the Printer Wizard now allows selection of generic Postscript printer, Level-1 or Level-2, and this is the recommended procedure for setting up a Postscript printer, though I cannot confirm it works as I don't have one. The file dither-matrix-2x1.xml is posted here: http://www.puppylinux.com/forum/azbb.php?1167307626 archwndas
Barry,sorry for the post, but I have a friend here who has an Intel Core 2 Duo Laptop and he wants puppy. Unfortunately puppy's kernel doesn't support mutliple cores. Could you compile a puppy kernel with SMP support since all the new laptops are multicore laptops? We do not want any other Linux Distro we want puppy because we just like it more. If you can't do that is there any way to install a custom kernel in puppy? Please don't let us down. Barry Kauler
No, not for Pup 2.13, as that is basically frozen.The problem with recompiling the kernel is that it can potentially break all the existing kernel modules, which means the whole lot would have to be recompiled -- and that's a lot of work. Small configuration changes should be okay though -- after 2.13 is released, I can take a look at it. There have been a few other requests for kernel config changes and we would have to gather them together and make the changes, then see if it breaks the current modules. |
Gxine, JWM, PureFTPd updated, xorgwizard improvedI also upgraded JWM window manager from v1.8rc4 to v1.8. There is feedback from the forum that the final release works okay, so I decided to go for it. Kirk made a small modification to his PureFTPd FTP server package, so I upgraded that. The Unleashed package is now version 1.0.22-1. Forum link: http://www.puppylinux.com/forum/?1163553598 There are a couple of bug reports for 2.13BETA that I cannot reproduce. Kirk reported that if the pup_213.sfs file is copied from the CD to the same place as the pup_save.2fs file, Puppy still used the one on the CD. However, I tried it and puppy used the pup_213.sfs that was copied to the hard drive. So, I cannot flag this as a bug. Debernardis reported that "screen lock does not work". I presume that is referring to the "lock" icon on top-right of screen. I tested it, it works fine. Forum link: http://www.murga-linux.com/puppy/viewtopic.php?t=13883 Thanks to Raffy and other for providing feedback on setting up Xorg. I have modified the xorgwizard script and will post it on the forum. Forum link: http://www.murga-linux.com/puppy/viewto ... p;start=15 bodigi
I had the same problem (screen lock does not work).I solved it running mrxvt and typing xlock.bodigi ferikenagy (ferikenagy<at>yahoo.com)
about ROX 2.5 bug: freeze on /proc, with thumbnails on, after some
research it appears to me it is not an ROX bug! in /proc directory only
"kmsg" file cannot be open (no cat,more...) and cannot be copy part of
it with "dd if=/proc/kmsg... so it is obvious it cannot be accessed
(opened) in a normal way. There was a discussion in forum how is
recognized the file type in windows (after the name of file extension )
and in linux (after rhe first bytes in file ,an sort of magic number,
use of 'file' command ) and the old ROX used the windows style after
the file extension name .It seems that new ROX 2.5 is upgrade to recognize the file type in linux manner by opening and reading the beginning of file , and because '/proc/kmsg' cannot be opened it is waiting (like dd if=/proc/kmsg..)in an endless loop which freeze ROX....So until now evrything is ok with ROX (not with /proc/kmsg), there is another little bug in it (no related with kmsg): if you try to 'kill' any rox windows (from right klick on menubar) it will kill also the background ROX the pinboard with descktop background image and descktop icons too... Somebody should verify my theory...who knew ROX how works internally.... "cat /proc/kmsg" and "dd if=/proc/kmsg..." are stucked but "file /proc/kmsg" reports "/proc/kmsg: empty" interesting! Dougal
Good work, ferikenagy! Barry, it now makes sense why they re-wrote the thumbnail code -- using magic numbers is more accurate. Maybe something can be done about that problematic file? BTW, I highly recommend upgrading Geany to 0.10 -- it has many improvements over 0.9. kirk
The problem with the pup_213.sfs not being loaded from the hard drive
is a SATA drive problem. If I boot with "puppy PMEDIA=satahd" boot
option, then the pup_213.sfs file IS loaded from the hard drive. If I
don't pass that option then it loads from the CD and sets PUPMODE=13.Just tried booting 2.12 from the CD, same problem. I got this laptop a couple weeks ago and went straight to a frugal install of 2.12. Had the same kind of problem with PUPMODE=13, so I added PMEDIA=satahd to grub. Here's my PUPSTATE file when booting 213B from CD: PUPMODE=13 PDEV1='hdc' DEV1FS='iso9660' PUPSFS='pup_213.sfs' PUPSAVE='ext3,sda4,/pup_save.2fs' PMEDIA='idecd' SATADRIVES='sda ' #these directories are unionfs layers in /initrd... SAVE_LAYER='/pup_ro1' PUP_LAYER='/pup_ro2' ZDRV='ext3,sda4,/zdrv_213.sfs' kirk
If
you left click on Xlock nothing happens. If you right click, you get a
list of options, but not the option to set the password. As bodigi said
you can type xlock from a terminal window to set the password.PaulBx1
"if
you try to 'kill' any rox windows (from right klick on menubar) it will
kill also the background ROX the pinboard with descktop background
image and descktop icons too..."Hmmm, sounds vaguely like this problem: http://www.murga-linux.com/puppy/viewtopic.php?t=13790 pakt
kirk, the Xlock bug (as several other 2.13b bugs) is caused by the use
of Mrxvt masquerading as Rxvt. If you replace the Rxvt wrapper script
with the real Rxvt binary, Xlock will work as before ;)Paul |
Various small fixesNote that you can upgrade a pup_save.2fs file to pup_save.3fs with internal ext3 f.s., but there are extremely limited situations in which the ext3 journal will actually bring any advantage. I have modified the /initrd/sbin/init script so that a filesystem check of the pup_save.2fs/3fs file is done at every boot. This only takes a couple of seconds on fast hard drive media, a bit longer on USB media. Puppy 2.12 loads the framebuffer module for the video hardware, even though it is not used. On all the PCs I tested, this seemed harmless, however it appears in some cases the framebuffer driver is causing trouble. So, I have modified /etc/rc.d/rc.modules to skip over the framebuffer modules and not load them. Mrxvt terminal emulator is causing problems with some scripts. Although Urxvt looks good, I decided to fall back to Rxvt. However, Mrxvt and Urxvt are now PET packages so can be installed and used by anyone. ROX-Filer 2.5 has a curious feature called "iconified windows", in which you minimise a window to the taskbar and it also becomes an icon on the desktop. I suppose this could be useful for systems that do not have a taskbar, but in Puppy's case it is more like eye-candy gone mad. I have turned it off by default. |
ROX-Filer bugsUsing Filer 2.5. I already posted a few weeks ago about a scaling bug -- for large-thumbnails selection, small images are scaled up -- Puppy Linux has a heap of 16x16 xpm's and they all get scaled up and look really crappy. Filer 2.4.1 was released in December 2005, so that's how long those thumbnail bugs have been there. I even went back to ROX-Filer 1.3.10, but it has another serious bug with the scrollbar that makes it unusable -- that also is a GTK version problem I think. It doesn't help that the GTK developers keep moving the goal posts. JB
Barry,Thank you for the remarkable work you and the other contributors have done to bring Puppy Linux to the masses. Puppy Linux is truly a one of a kind distribution. Looking forward to release 2.13. Happy New Year. JB PaulBx1
Go
back to 1.2.2? I am pretty used to it, and it does what I want it to
do. Maybe others have need of the new features and pretty looks, I
don't know...Dougal
Barry, maybe the bug with /proc has something to do with Puppy?In the past I mentioned a bug where PupRox crashed when entering a certain directory (in the FLTK source) -- it ended being the result if a badly written xpm... Maybe someone could check Rox 2.5 with another distro and see if the same thing hapnens? Eyes-Only (eyesonly<at>gwi.net)
Good suggestion Dougal! And since I'm in VL5.8 Standard Gold at the
moment I opened my Rox 2.5 which I have installed in there. I went to
"/proc" and...I'm just returning from a reboot of X. LOL! I hope this helped? ;) Eyes-Only "L'Peau-Rouge" vern72023 (gdcrane<at>bellsouth.net)
Personally I do not like "thumbnails" and have always turned them off in Rox.I think that by default they shd be turned off , just as "iconified" windows shd be turned off Anyway I have been using Rox 2.5 since I started with Puppy and I thought its adooption in 2.13 was a good step forward George Barry Kauler
ROX-Filer 2.2.0 thumbnails work perfectly, and the /proc bug is not
there. A Rox developer rewrote the thumbnails rendering for ROX-Filer
2.4.1 and that is what has stuffed things up.I don't know why they did that, as the old thumbnail rendering works well. ...oh well, I guess they had their reasons. I guess I'll stay with 2.5, and there should be a fixed release sometime. Unless something else serious turns up... |
Rxvt-Unicode terminal emulator looks goodI'm testing Urxvt right now, it works well. It doesn't have tabs, but that is low on my priority list. More important is the support for mixed fonts, including X11-core and Xft fonts -- and what is particularly interesting is that Urxvt can mix them on the same line. You specify a base font, and if a character is needed that is not in that font, Urxvt can look in other fonts for it. Urxvt is developed by Marc Lehmann and the project home page is: http://software.schmorp.de/pkg/rxvt-unicode.html pakt
Is
there a demand for rxvt with tabs? Personally, I prefer to have
separate terminal windows open making comparison and copying easier.I have always added the '-p' switch to the script calling Geany to disable tabs for the same reason. Btw, the new Geany version uses the '-i' switch instead to disable tabs. Paul Redrum
Yes
! Tabs in terminal is very important for some people -including me !!
Blame firefox for spoiling us all with anything that can be tab-able !!Jason Pline
Tabs are nice but I agree mrxvt is too buggy right now and I don't know
of any other terminals that wouldn't require a massive amount of libs.
Hopefully the bugs get resolved.
|
|
To-do list for 2.13-final
Version 2.13BETA is basically frozen, no upgrades of packages, nothing
major, just basic bug-fixes before the final release. These are some
things that I already know need to be fixed or looked-at:1. Mrxvt problems I may revert from Mrxvt back to Rxvt, as the former is buggy. The problem is pasting long lines, they do not always wrap properly. Also, Mrxvt is incredibly slow to shutdown sometimes. I have also been examining my usage patterns, and I don't use the tabs much anyway, mostly because I use Rox's right-click to open a terminal in the current window. 2. Modem Wizard I have many plans to improve this, just haven't had the time. I will see what I can do before 2.13-final. 3. Serif fonts in welcome window The initial "woof woof" welcome window displays with a Type1 serif font. Ditto for the Help window. The welcome window is using an HTML-viewer executable named 'bareview' and the Help window is using a HTML-viewer executable named 'seaview' and these use the SeaMonkey libraries. I will see if I can change to DejaVu Sans font for consistency with everything else. 4. Complete changeover to ext2 New pup_save.2fs files now have an internal ext2 filesystem, but much of the description in the shutdown dialog windows refers to pup_save.3fs and ext3 -- needs updating. 5. F.s. check at bootup For pup_save.2fs personal storage files, I'm thinking that Puppy will do a f.s. check at every bootup, rather than the default every 25 mounts. I think that is what Puppy did way back when Pup was using ext2, just before we moved to ext3, and it doesn't slow down the bootup by much. I have to find an old Puppy and look at the code -- as I seem to recall it had some special feature to do a f.s. check then if an error do it again with forced repair, or something like that. 6. PET package dependencies If you look in /root/.packages/packages.txt (also livepackages.txt, which is a subset of the first file, a list of PET packages that can be downloaded and installed), you will see that each entry has a dependency list. For example, here is the entry for Pupdvdtool: "pupdvdtool-0.5" "pupdvdtool-0.5: dvd ripper" on "GTK2APPS +vamps,+vobcopy,+ffmpeg,+dvdauthor,+gtkdialog 128K" \You can see the list of dependencies. Versioning is not required when referencing the official packages. Most important, the dependencies are hierarchical, for example '+gtkdialog' will also mean that all the dependencies of 'gtkdialog' will have to be installed, including GTK and all of its dependencies. So, you don't need a huge list of all dependencies in each entry. Note, a minus sign before a package, for example '-somepackage' would mean this package is incompatible and must not be installed. Also, some packages will always be in Puppy, even in a bare-bones Puppy, so will never have to be explicitly named as dependencies. Anyway, many entries in 'packages.txt' are not fully populated with their dependencies. I will do some more work on this, but if you have repsonsibility for, or interest in, a particular package, perhaps you might like to fill-in the dependency-list and send it to me. 7. Turma has rude exit dialog GuestToo used a hex editor to fix this, and I had his 'fixed' Turma in Puppy (I think) but that seems to have gone astray and the rude one is back. Barry Kauler
Lobster has started a Forum thread for reporting 2.13BETA bugs:http://www.murga-linux.com/puppy/viewtopic.php?t=13883 BarryK
Number 3 is fixed. The welcome and help windows now display with DejaVu
Sans-serif font. I had to edit /usr/lib/seamonkey-1.0.6/greprefs/all.js.Kal
Was
playing around with new jwm 1.8. Removed delay on blinky and added
height="24" to freememapplet in the .jwmrc-tray . Everything seemed all
right on the tray.Kal PaulBx1
I
like rxvt, never saw any problem with it. If you decide to keep mrxvt,
would it be possible to keep both? rxvt as a backup in other words?
Until mrxvt gets solid, anyway...There is a bug that some users have turned up with 2.12, I wonder if you've seen it? It would be nice to get that cleaned up in 2.13 too. http://www.murga-linux.com/puppy/viewtopic.php?t=13692 Dougal
Why
did GuestToo have to use a hex editor? It's just a simple matter of
commenting out a little bit of code -- MU did it and made a dotpup a
month or two ago.Both JWM and Geany have new versions out, BTW. Raffy
Re:
Fonts and opening MS documents - Thanks for bringing up the matter of
fonts, Barry. Here are some notes about OO1.1.4 that you or other Puppy
developers may find useful.The smallest OpenOffice build that we're using is 1.1.4 and this is in the PET/pupget repository. This version converts presentation fonts to narrower ones, and naturally the presentation display would be very different. Even the originally prepared fonts in OO1.1.4 tend to be narrow. Maybe just linking to the new fonts in Puppy can solve this problem, or is the solution a lot more complicated? Hacao's friend in Hanoi also suggested that a newer version (is it 1.1.5?) of OO can handle Open Document file types. If there is a way of adding to/replacing the fonts in OO to the new set being used by Puppy, that will be welcome, as it could even reduce the size of the OO PET package. BarryK
Dougal,Maybe I'm getting mixed up with Gkdial -- G2 may have used a hex editor on that, as I think there are problems when we try and recompile it. I think G2 did also modify Turma, but perhaps he modified the source. BarryK
I
might go back to Dillo for the internal HTML viewer. The problem is
that 2.13BETA has "bareview", a viewer that uses the SeaMonkey
libraries, uses a lot of memory. I'm testing on an old 64MB 500MHz PC
and bareview is barely able to run in that space and is very slow to
start -- which is not good for the first welcome window.Raffy
Re: SeaviewBarry, I just had a good experience with Seaview - got it running when I typed "Dillo" in console. I can say that Seaview keeps separate Internet files, as Seamonkey can't use its Forum cookie. I have started a thread about seaview at http://murga-linux.com/puppy/viewtopic.php?t=13953 It has no right-click but is fast (in my trusty old Celeron 400). Re: Font Consistency Based on your (Barry's) inital discussion: "the Help window is using a HTML-viewer executable named 'seaview' and these use the SeaMonkey libraries. I will see if I can change to DejaVu Sans font for consistency with everything else." - yes, this will be a very good technique which hopefully we can also use with OpenOffice. :) |
|
Puppy version 2.13BETA
This is not a general release, it is for our Puppy-testers only.Download: http://www.puppylinux.com/test/ Release notes: http://www.puppylinux.com/download/release-2.13.htm There are some things to do before the final release. I'll itemise them soon. pakt
Thanks Barry - looking forward to exploring itPaul Lobster
This really is a Star release. The new Rox is speedy, very nice to use.
The graphic start up screen is so much simpler and Pet is looking good.Rarsa has some exciting news which I look forward to hearing more about. I got a PM message from someone today, offering to help on the wiki and they do not even have Puppy yet . . . Yes Puppy really is that good. Many thanks Barry, this looks to be a very exciting release :) F M Lynch
GreatJob!!!2.13 found my Prism2_usb module, installed the linux-wlan-ng driver, configured, and activated the prism2_usb module automatically. I was connected to the internet on initial bootup. WOW!! Thanks Barry I am looking forward to exploring the other exciting new features. debernardis
It
is really wonderful and works absolutely allright, quick and slick,
with my preferred apps. I'll test it thoroughly and I'll make the next
rudy on it. The only thing I need to bury my other distro is bluetooth
support.Thanks Barry! vern72023
Very nice BarryI took the pup_213.sfs and only removed SeaMonkey then merged my own apps into it ( Opera and Wine and Win32 ) and saved it back - it came out at 71MB which is really very nice and tight and I may not even remove the other elements I normally do like xorg given the size. Then I copied the puup_213.sfs and vmlinuz and initrd and zdrv_213 onto a Lexmark usb drive I use to test the distros with. It upgraded the existing pup_save very cleanly just the usual gotchas which i fixed from the tmp save. The new distro works well onmy main machines, including the low memory ones; BUT even better it worked without any major driver tweaks on my wife's e-machine - which frankly had not behaved well under any of the 2.x distros - networking has been abysmal with the forcedeth drivers on 2.1 and 2.12 on this Another great job Barry - Raffy
Thanks, Barry, and Happy Holidays!!! (Uh, hope you're having even a half-day of holiday. :) )
|
Notes on PETget, Dillo, HV3In the case of the primary repository, http://distro.ibiblio.org/pub/linux/dis ... uppylinux/, the older .tar.gz pupget packages are in subdirectory 'pupget_packages-1' whereas the new .pet packages will be in 'pet_packages'. This is good, as there is no rush to migrate the packages. Same thing if you download a .pet or .tar.gz to your PC first -- when you run PETget and choose 'local' button for the download repository, PETget will handle the old format as well as the new. I did a quick test of the new dependency mechanism. Package 'cutecom-0.13.1' has 'qt' as a dependency. When I ran PETget and chose to install Cutecom, it did so, then continued and installed 'qt-3.3.6' as well. PETget in Pup 2.13BETA will not have support for the proposed advanced features of .pet packages, as I'm just focussing on it basically working for now. Currently, a .pet package is just a .tar.gz with md5sum stuck on the end. But, we propose the 'specs' file, which among other things will be handy for alien .pet packages (not in the official Unleashed suite) as it will have the dependency list plus all the info needed to create an entry in the 'alienpackages.txt' file. I'll tackle this for the final release of v2.13 (or BETA2 if necessary). Regarding Dillo. Many scripts assume Dillo is present, but for a particular custom build of Puppy it may not be (for example, v2.13BETA). It is not satisfactory to always have dillo as a symbolic link to whatever HTML viewer is present. Therefore, I have created /usr/local/bin/defaulthtmlviewer. For any scripts that have to launch a HTML viewer, please use that instead of Dillo. Regarding HV3, the Tcl/Tk HTML viewer and web browser. I have setup the UNleashed package so that if you create a live-CD with only HV3, no other browser, then HV3 will automatically be setup as the 'defaultbrowser' and 'defaulthtmviewer'. Sage
It
all seems to be getting very complicated, Barry? Simplicity was (and
is) the great unifying principle and one of the major attractions of
your masterpiece. I'm sure you'll be able to reassure us mere users.BarryK
Complicated? Just the reverse. It is a move toward simplicity. There is
a transition period, in which the old pupget packages have to be
recognised. later there will only be PET packages.Perhaps later the dotpup packages will be history also, all converted to PET packages (MU already has a tool for that). Then there will just be one consistent and simple package format for Puppy. Much of what I write in this blog is information for Puppy developers, so there may be detail that is confusing for users. Sage
One tries to follow, in outline at least, what developers are doing so
that there is a minimum understanding of changes when they occur and
what will be required of users. Two old maxims often appear relevant: A little knowledge is dangerous. Oh, ye of little faith.... Anyway, thanks for the explanation, Barry. Merry Christmas; over & out. Dougal
The "defaulthtmviewer" is a good idea. I was just thinking about
something like that after you mentioned replacing Dillo with that
Mozilla viewer. As MU pointed out, that viewer is dependent on
Seamonkey, so now when people have a Seamonkey-less Puppy they can just
change the defaulthtmviewer...BTW, Barry, I'm quite sure the man script uses Dillo explicitly. BarryK
For 2.13BETA I kept dillo as a symlink.
|
PBrename, Pupdvdtool, Xwget, Geany updatedIan (Ian in Forum!) has upgraded Xwget GUI frontend for wget to v0.5. This also now in Puppy. The very latest Geany text editor, v0.9 is now in Puppy. PaulBx1
For what it's worth, the geany developer mentioned they are just about to release 0.10:http://www.murga-linux.com/puppy/viewtopic.php?t=13790 JJ
Hi, thanks. can you upgrade the games too? and put some solitarie
|
Feature freeze for 2.13BETAI'll update some packages tonight, then do testing for a couple of days. This BETA will of course not be for general release, just a treat for Puppy testers at xmas. I'll compile SeaMonkey 1.0.7 tonight, which also means I have to recompile Gaim to be sure it's ssl plugin works. I'll update the Gimp PET package as the old one is compiled against gimp-print but we are now using Gutenprint. There are a few other pkgs to upgrade, including some of plinej's. I'm not going going to recompile the kernel at this stage. Some kernel config changes were suggested to support more drivers, but will have to hold that one over. J_Rey
I hope you noticed my response about compiling 1.0.7 with profile roaming support....J_Rey
OK nevermind.... I just read the previous posts.S. Joanou
At
nearest convenience, an update of IPW3945 and IEEE80211 and other
peripheral code would be wonderful. Currently, it freezes my SMP Dual
Core Celeron machine and the update may fix the problem.Happy Holidays! PaulBx1
We are still at busybox 1.01, while the latest release is 1.30 (with patches):http://busybox.net/ Normally I wouldn't care, but I found one patch they have in their system that seems to be needed for cryptsetup: http://bugs.busybox.net/view.php?id=838 Unfortunately, John is having problems booting, very possibly because this patch is missing. If this is in fact the sticking point, it makes encryption (beyond the basic cryptoloop) work difficult, if not impossible. I suppose he or I could compile the latest busybox and press on, but it would be even more convenient if Puppy 213 had the latest. If this is not possible, oh, well... BTW I cannot tell if this patch is incorporated! I may have to email the developer to find out. I have found a tutorial where someone showed how to use cryptsetup and busybox together, using the initrd-style boot, so that suggests later busyboxes do have the required patch. I can't tell from the release notes though. PaulBx1
Now I'm thinking this was a patch to add cryptsetup itself to busybox.
I don't think it's been implemented, and I don't think it's for
cryptsetup-LUKS anyway. So never mind...John Doe
I filled PaulBx1 in on the Busybox/init issue in the encryption thread.Also, I'll try not to drop any many more "kernel driver bombs" on you.  Regarding both these topics, maybe you should start up two threads in the developers group and we could get a couple people together who really want to grind out the modules and busybox/init upgrade. We could work it all out for 2.14. I wouldn't mind setting a 6 week goal like that. Anyone else? p.s. The occasional problem people have posting here is most likely the length on the session timeout. I just had it happen (thought too long I guess). Your anti-spam gif sets it's value on the server in a session variable. When the user takes longer than the session timeout to post back, their entry is invalid even if it is correct as the server has forget what the value is to validate against. Although I copied to geany before I submitted (thanks Sage). q9f0nh5pctdvzmw (q9f0nh5pctdvzmw<at>googlemail.com)
SeaMonkey rocks..
|
tic, Isomaster, PuppyBackup, PuppyMirrorForum member mig21 announced Isomaster version 0.6, a tool for opening up an iso file. This has received very positive comments on the Forum. Puppy is now upgraded to this version. Sigmund (zigbert on the Forum) has created a couple of little gems. PuppyBackup version 0.5 is now in Puppy. I have removed my old 'bkup2cd' utility from the menu as it is unmaintained and now there is just PuppyBackup in the menu. Zigbert also created PuppyMirror version 0.1, for syncing directories, and I have also put this into Puppy. (note to zigbert: there are some error messages when run this from a terminal, which might need to be looked at). PuppyBackup and PuppyMirror are in the "Utilities" menu. Zigbert
Errormessages when running from terminal: date: invalid date `+%s' expr: syntax error expr: syntax error It is fixed for the upcoming 0.6. For PuppyMirror I don't get any messages. Please let me know if you do. BarryK
I got those errors with PuppyMirror.I had better check that I got it the right way round! zigbert
The error is not critical. Xdialog is getting wrong dateformat. It is
the reason why Xdialog-calender shows today by default. - Not
backupdate as expected. I'll guess it is nothing to rush about???
|
SeaMonkey, Gaim upgraded, various small issuesGaim is upgraded to version 2.0.0beta5. Testing Gaim brought to light another little problem with the new Xdialog. Gaim is launched via /usr/sbin/gaimshell, which in turn launches /usr/sbin/gaim-autosetup.sh (which was developed by J Marsden) which starts Gaim with automatic login to our #puppylinux channel. The autosetup.sh script runs Xdialog, which displays a dialog to enter a nickname -- no OK button, requires pressing ENTER key to quit the dialog window, except that it doesn't work. I don't know how to tell it that ENTER key is equivalent to clicking OK button, so I modified the script so that it does have an OK button. An interesting thing I have done with the new SeaMonkey. The source package has a demo HTML viewer called TestGtkEmbed, which is only 16KB and starts instantly. It has a very (very) basic menu, but I modified the source to not have any menu, no status bar, no nothing, just a window, and I set the default window size to what I wanted (hardcoded), and called it 'bareview' -- this is really handy to display the welcome message at first bootup. Maybe handy for other purposes, like some of the stuff MU has used Dillo for. TestGtkEmbed has a simple text-only menu with Forward, Back, Stop, Refresh and a URL-entry box, and I tried to replace the buttons with icons ...but my GTK programming skills are not up to it -- I created something which just crashes. Really don't have time to go off at a tangent anyway. I'm trying to get the internal modem on my laptop working. An ALSA driver kind of works, with the Smartlink daemon: # smodemd --country=AUSTRALIA --alsa modem:0...and amazing, I can send "ATZ" to /dev/ttySL0 and the modem responds with "OK". But, it won't dialout. So, a work-in-progress. MU
I was experimenting with gtkmozembed in the past, for example this Flash-based clock uses it:http://murga-linux.com/puppy/viewtopic.php?t=8516 The problem simply is, that it depends on all the mozilla libraries - so it will not work in puplets without seamonkey. (versions based on firefox or opera). For this reason I decided to use dillo for the debianinstaller, as it is very small and so can run in any puplet. Mark alienjeff (alienjeff<at>charter.net)
From today's posting: [i]There was a request on the Forum for the
'sroaming' extension, so I have included that, although I haven't got a
clue what it does.[/i]With all due respect, am I the only reader who found this excerpt, uh ... a tad unsettling? klhrevolutionist
Barry I cannot find a post about this extension. And after
investigating this is closed source extension, I realize that yuor keen
to using the best of both worlds but I will have to protest on this.There are many people who make great suggestions and slowing down the browser with a closed source intrusive extension is not the best idea in the world. I have to disagree very strongly on this BK and though you may not care the user who unwittingly uses this extension by accident or curiosity might not like this either after finding out who owns the extension. And what information is collected by google. Google is a known cia asset and puppylinux should not participate with such agencies. http://www.infowars.com/articles/bb/goo Google does not need to be involved with a community that is aimed towards end-users and allowing such activities would not be in puppylinux best interest. The information is not encrypted before going to google and vice-versa. Barry: Please do not include this in puppylinux alienjeff (alienjeff<at>charter.net)
Before the word "paranoid" starts to be promiscuously thrown about, ask the Mozilla folks what "sroaming" is about:https://addons.mozilla.org/search.php?q=sroaming&type=A&app=mozilla just a thought... rather than subject people to closed source extensions that are hard to even find info on (sketchy...) to a foss browser (bringing you back to square one in a sense) why not include the install file in the distro, so those that WANT it attached to seamonkey can just click "install"? yeah, only useful to people that have pupfiles/remaster then... but easier to install than it is to remove? p.s. from the book i'm currently reading (just fun coincidence?) "deliver a gift that looks attractive, yet steals the very key to your security. sharpened over the milennia, this technique still works against everyone except the truly paranoid." - cliff stoll (i just thought it was well said) SouthPaws (jaguar1<at>netzero.net)
Will this help...?http://www.altavista.com/web/results?it Barry Kauler
sroaming is not a closed source extension. It is in the extensions directory in the SeaMonkey source, hence open source.extensions/sroaming/ has a README file, but it doesn't make any sense to me. Here it is: [blockquote]This implements 4.x-like roaming. To make the implementation vastly more simple, it has been decided that no syncing during the session happens. The design will not allow that either (at most sync in certain intervalls). A full-blown dynamic implementation that immediately update the server when a data change occured requires the cooperation of the data providers (bookmarks, prefs etc.) and is thus a huge change that I will leave to somebody else to implement independant of this roaming support here. alecf made such proposals a longer time ago on n.p.m.prefs, they sounded very interesting, but unfortunately, nobody implemented them so far. When the users selected a profile, we will check, if it's a roaming profile and where the data lies. If necessary, we will contact the server and download the data as files. We will overwrite local profile files with the downloaded ones. Then, the profile works as if it were fully local. When the user then logs out (shuts down Mozilla or switches to another profile), we upload the local files, overwriting those on the server. Following Conrad Carlen's advise, I do not hook up using nsIProfileChangeStatus, but in nsProfile directly. That just calls |nsISessionRoaming|. Its implementation uses various protocol handlers like |mozSRoamingCopy| to do the upload/download. These in turn may use generic protocol handlers like the netwerk HTTP protocol to do that. Also following Conrad's advise, I do not store the roaming prefs in the prefs system (prefs.js etc.), because that it not yet initialized when I need the data (of course - prefs.js, user.js etc. might get changed by us), but in the Mozilla application registry. For the structure, see the comment at the top of prefs/top.js. Overview of implementation: - transfer.js (the Transfer class and support classes) contains the non-GUI logic to transfer files and track the progress and success. - progressDialog.* shows the progress to the user (and also works as interface to the C++ code) - conflictCheck.js is the "controller", controls the overall execution flow. It determines what has to be done (which files to transfer when etc.), including the conflict resolution logic, and kicks off the transfers. There is a long comment at the top describing the implementation. - conflictResolution.* is a dialog to ask the user when we don't know which version of a file to use. [/blockquote] When someone requested it (on the forum I thought), he gave a brief description that seemed harmless, though I didn't fully understand and have forgotten the bit I did understand. Anyway, if noone comes forward and states that this is good to have, I'll recompile without it. Barry Kauler
Well, Eddie has posted that SeaMonkey 1.0.7 is now out, so I'll recompile tonight. Is sroaming in or out?Eddie's post: http://www.puppylinux.com/forum/azbb.php?1166636607 Barry Kauler
A
note on Gaim. V2.0.0 is about 3.4M (uncompressed, and that's after
being cutdown somewhat) compare with v1.5.0 at about 2.5M.I'm not happy about that and will stay with 1.5.0 for Pup 2.13. V2.0.0beta5 is a PET package though (ie, what we previously called a pupget pkg). PaulBx1
Here is a description what roaming is about:http://gemal.dk/blog/2004/08/31/using_roaming/ Might we already have other ways to accomplish the same thing? While there may be more paranoia than called for (if it is really open source), my personal preference would be to not complicate things this time. Maybe bring it up on the forum, see what people think about it, see if there really are any security holes there. Better to find out before installing it, than after when some (few, apparently) will have come to depend on it. Or, let individuals install it if they want it. klhrevolutionist
It was my understanding that this was a google ext. I misunderstood...Barry do what you always which is what you do always.. I was worried for a second... Thanks for clearing up the matter ! BarryK
PaulBx1, thanks for that link, it clarifies things immensely.I found a note in my Notebook, it was J_Rey who requested sroaming. I might take it out. I'm not sure if I like to idea of a central server that can automatically download to my computer. I'm getting paranoid too! J_Rey
Wow. I just missed this whole conversation and it looks like it may
already be decided. Maybe this deserves more discussion in a forum
topic but for now:I asked to include it since Barry disabled it although it normally would be installed [b]by default[/b]. To set up profile roaming, see Edit -> Preferences... -> Roaming User to share profiles with other SeaMonkey installs (or other Gecko-based browsers with it enabled) which is easiest done as File Copy instead of setting up a server to do it. So enabling this would allow SeaMonkey to share [i]one profile[/i] between multiple OS's on one machine although if a server is used then of course would extend beyond that machine. Now, adding the support for this to be available does not mean that it will automatically use a server, it has to be specified and enabled. I do admit that the documentation for this is sparse, but the SeaMonkey developers have included by default for a time now.... BarryK
Maybe I'll leave it in then! alienjeff (alienjeff<at>charter.net)
From Mozilla/Seamonkey's own documentation (prefs.properties) for [i]sroaming[/i]:"12 ActivationWarning=Be warned that the roaming implementation is still new and experimental. It may accidently overwrite your roaming profile files (both on the server and locally) with a bad or empty version. Please make regular backups of your profile to prevent possible data loss. See the release notes or ask your administrator for where you can find the files. Please report any dataloss bugs you find. This warning won't be shown anymore for this profile." And [i]please[/i] don't default to the lame excuse that it's old documentation that simply overlooked by the developers. |
More on reverting to ext2 for pup_save filePuppy currently creates a personal storage file named pup_save.3fs with a ext3 f.s. inside it. The only exception to that was if created on a floppy disk it will be pup_save.2fs and contain a ext2 f.s. Ext3 is ext2 with an added journal. This journal is useful as it can repair a f.s. automatically, and that is most useful if the power goes off unexpectedly. The problem is that it doesn't work too well when the loop-device is used to mount the save-file, which is how it is done in Puppy (and how it has to be done). It does work if the underlying f.s., that is, the f.s. of the h.d. partition in which the pup_save.3fs file reside, itself has a journal (ext3 or reiserfs). However, the journal inside pup_save.3fs is basically useless in all other situations, including (especially including) USB Flash devices. It has also been pointed out that ext3 has a speed penalty. I will add that there is a considerable space penalty also and this is an important point for Flash devices. So, I have implemented a solution. I have edited /initrd/sbin/init and /etc/rc.d/rc.shutdown. At first shutdown, when you get to choose where to create a pup_save file (or if you booted off a usb flash or internal h.d. partition you don't get any choice, the boot partition is used), if the host partition is ext3 or reiserfs then a pup_save.3fs file with internal ext3 f.s. is created. In all other cases a pup_save.2fs file with internal ext2 f.s. is created. If you already have a pup_save.3fs file, and want to mount it as ext2, just rename the file to pup_save.2fs. There is one thing undecided. For the pup_save.2fs file it would be good to run a periodic f.s. check at bootup. I'm wondering about the best way to do this. It would have to be done in the /initrd/sbin/init script, before the pup_save.2fs file is mounted, but I'm not sure how to trigger it. I don't want to run the check at every boot. What about based on the date, like once a week? Or, if there's an error report when it gets mounted. rrolsbe
Can
someone comment on the flash write wear-out factor as it relates to
using this new ext2 method with puppy. I thought I understood how ext3
with a journal worked, guess not!Thanks for all the hard work! Regards Ron alienjeff (alienjeff<at>charter.net)
My
Gentoo box, which sees very little use as it jealously sits next to my
Puppy machine, fs checks every x-number of boots. This seems to be a
more logical approach than checking every x-number of days, though
coding might be a tad trickier.Another approach would be a pop-up dialog box triggered by date, boot number or whatever, with a reminder message, "do you want to check your file system now?", etc. Sage
As
AJ points out, many distros just announce that fsck will run every so
many boot ups. Especially in older distros, anyone with a quick eye and
a slow PC can read the relevant statement as it scrolls past. Guess
there's a simple decrementer in there somewhere? 30 and 60 seem popular
numbers. Not sure whether there's any statistical significance in those
choices? Perhaps not. Would refer to full HD installations, of course;
not sure about USB fobs, multi-RW and type 1 installs?PS. Folks contributing here would be well advised to copy their copy to a clipboard before 'submitting'. On several occasions, some of my mindless gems have gone missing into cyberspace. No idea why - I don't do SW, but something similar can happen over on John's board if two contributors try to post simultaneously. Pizzasgood
There is a timeout here. If you take too long, the anti-spam code at
the bottom expires. At the forums, if somebody posts before you finish,
when you hit submit it will warn you first so you can change the post
to reflect that. If you hit submit again it will go through.PaulBx1
Barry, I don't know if your solution is complete. It's not that the
underlying fs has to be ext3 only, but it also has to be mounted
"data=journal" or "data=ordered", at least according to one of those
things I found and posted in that thread. I don't think reiserfs will
work unless it has those options.Also there is the problem that all the Puppy scripts seem to be expecting .3fs pup_saves, and when they mount they are mounting as ext3 file systems (for example, I'm guessing the resize pup_save script does it, and I know the script to encrypt pup_saves does too although that is not yet released). So the ramifications might be larger than we think. Maybe a global search through the scripts would turn up all the cases... On the fsck, I'm wondering if it makes sense to trigger that on shutdown, so as not to slow down Puppy's fast boot. That is, when you boot you are sitting there waiting for it to come up, while when you shut down you typically walk away and don't care if it takes a minute or 20 minutes. It could be more frequent too, maybe even every time. Or just have a dialog on shutdown so if someone is going through a lot of reboot cycles he can skip the fsck. Or have two shutdown commands in the menu, one with fsck and one without. It might seem a little funny to have it on the shutdown, but then only one disk read at boot is unprotected; if people schedule them every 30 boots, that clearly is a lot worse. PaulBx1
Of
course, if there was a crash, it might be good to be able to trigger an
fsck on boot as well, since you don't have a shutdown in that case. Maybe add a boot parameter?PaulBx1
Now
that I re-read my own post (!) back on that thread in the forum, where
I said an fsck only takes a second on a 512MB file, maybe we should
just do it every time, during the boot. It's not going to slow down
anything much...wmwragg
The way that ext2, and ext3 file systems handle when they are to be checked is through the tune2fs command:tune2fs man you can use the -c switch for max counts or the -i switch for interval between checks (days, weeks etc...) I believe the fsck will use this info and run a check or not depending on what it finds. Pizzasgood
I'm
fine with every so many boots. It's still much faster than the scandisk
on Win98SE, and it never starts over halfway through.Barry Kauler
Question, does tune2fs store the information about periodic checks in the filesystem itself?Puppy first boots in the initial ramdisk, which cannot store anything permanently, it is just the initrd.gz file. So, any instruction re f.s. checking would have to be done at shutdown when the pup_save file is created and stored in the f.s. itself. Barry Kauler
rrolsbe,It makes no difference if ext2 or ext3, Puppy writes to the usb drive every 30 minutes regardless. That was an arbitrary choice, could be made more or less than 30 mins. Dougal
The info about FS checks is definitely in the FS. I remember I used to
see the message about "need to run fsck" in Puppy in the past -- you
get them when the FS gets mounted.I don't really have an opinion in this Ext2/Ext3 business, but I think it is important to try and not complicate the init script too much… so I hope using both won't add too much needless complexity. If Ext2 is used, maybe-- besides the periodic check -- when pup_save is mounted, /etc/.XLOADED can be checked for and, if found, pup_save be unmounted and fsck-ed before remounting? Again, might add complexity. Barry Kauler
The simplest thing is just to go back to ext2 for all pup_save files,
regardless what partition f.s. it's on. As was pointed out, even if the
host f.s. is ext3 or reiserfs it still may not be correct to make a
journal work in the pup_save file.Phil Roberts (opplr<at>hotmail.com)
Hi
Folks, we recently discovered Puppy and are really happy with what it
does for us. We do have a slight problem and it is related to this save
file commented on in this thread.We boot from a USB flash stick ( 1 Gb Cruzer ). We used the HP formatting program to format it with FAT 16 ( the only way this system would boot it to my knowledge ). After the format, we SYSLINUX F: it and copy the 3 puppy files to the stick. Boot up is fine. When it goes to save the xxxx.3fs file, it must corrupt something in FAT 16. The next boot hangs when searching for pupxxx appears. I can copy the contents of the flash drive to a folder on my WIN2k box, do the FAT 16, SYSLINUX and copy everything back and it will boot fine. This doesn't seem too good. Is there a method to save the pup_save as a FAT 16 file? wmwragg
tune2fs saves the info in the file system itself. As the pup_save file
is a file system this should not be a problem, as the info will be
saved within the pup_save filesystem, and this can be checked when the
pup_save file(file system) is mounted on boot, or am I missing
something?PaulBx1
Well, I guess the fsck has to be done on an unmounted filesystem, so if
it were stored inside it would have to be unmounted again if the count
had gone to zero, then fsck'ed, then mounted again. Should work I guess.I still think it makes sense to try fsck-ing these small filesystems every time we boot, at least on the first go around. It might not be time-consuming enough to worry about using tune2fs. Another concern is that tune2fs would need to be added to initrd; though maybe that's not a big deal. John Doe
Correct me if I'm wrong, but regardless of how puppy chooses to save in
the future, he's backed into a corner as far as how he chooses to load.
Ext2 and Ext3 will have to both be supported because some ext3 users
will have to upgrade.
|
PETget package managerWhen Pup v2.13 is released, I will create a new directory on ibiblio.org, 'pet_packages' to replace 'pupget_packages-1', but will leave the old directory there to cater for older versions of Puppy (although I will have to check how much space ibiblio has now allotted to me). The new repository will have .pet packages, replacing the .tar.gz packages. petget accepts stuff on the commandline, same as the old pupget but more: # petget +xbubble-0.2.4 (install official .pet package).pet packages are the same as the old .tar.gz packages except have a 32-byte md5sum appended onto the end of the file. In fact, tar can be used to expand them directly, there will just be a warning about extra bytes on end of file. I decided not to be too ambitious about expanding petget to handle non-official repositories. MU is doing excellent work and is covering that side of things. His PSI could either use petget or Nathan's pkgtool with some enhancements. The next thing I want to do is consider a new specification file to go inside the .pet tarball. Currently there is the 'keyword' file, but I'm thinking of a file called 'specs' which can have any arbitray data in it, but let's make it in bash-compatible format. For a start, it could have: PETKEY=Xbubble #keywordIn the case of an alien .pet package, this info can be used to create an entry for /root/.packages/alienpackages.txt also for xbubble-0.2.4.keyword. The petget script can take care of this. So, if you download an 'unofficial' .pet package, just click on it and it will install and a correct entry will get created in alienpackages.txt. Dependencies will also get installed if missing, as long as they are from the official .pet repository. However, if PSI calls pupget, perhaps there needs to be a special instruction to pupget not to do dependency checking, as PSI does it? This 'specs' file is not yet implemented and if anyone wants to comment please go for it. Lobster, what about PET being an acronym for Puppy's Enhanced Tarballs? Pizzasgood
Just
a reminder that "uninstalling" built-in packages would be nice (maybe
with a separate button though). Also, it might be better if the
dependencies part could include all of them, because a custom Puppy
might not have things that the standard Puppy would. It wouldn't help
with puplets from the remaster script (unless the person doing it fixed
the list by hand), but it would for the ones from Unleashed.MU
>However, if PSI calls pupget, perhaps there needs to be a special
instruction to pupget not to do dependency checking, as PSI does it?I think it would be better to remove the dependency-check from PSI, if an official .pet is installed. PSI cannot yet automatically install missing dependencies, just display them. So your approach is more advanced, and I will prefer to use that :) You still could run the PSI dependency-check manually with a buttonclick, if a program will not run. This might happen with external .pets, where your dependency-resolution might fail. ################### In the next version, I'll try to replace the pkgtool-calls with petget. Mark Dougal
Barry, I just updated the pmount file again:http://dotpups.de/incoming/dougal/tmp/pmount.gz Lobster
"Lobster, what about PET being an acronym for Puppy's Enhanced
Tarballs?" Sounds like the Dogs Boll**** to me (meaning woof woof in
cockney)I am glad that MU is integrating his PSI ideas and I agree that Nathans ideas are very welcome. It would be great if Muppy and Grafpup and Puppy all use the .Pet standard. It would also be a natural development of dotpup, which we can still use and upgrade from. MU
Muppy is simply a blown up Puppy, so it inherits all systemutilities and programs from it.Grafpup was different, becau |