Puppy 4.2Beta Seamonkey Bugs & Fixes

[Minnesota]

Let me try to explain what I tried... Beta with six downloaded pets from page five of this post dated 2/16 here. No they do not work on Beta, NO THEY DO NOT WORK ON ALPHA 3

so back to versions I have dated 2/14 Do not work on Beta.
work on Alpha 3.

One change in today's pets for download....
cups-1.4b2-i386-v1.pet 2 /16

version on the 2/14 for alpha is v-2.

so... apples, oranges and possibly pears!

New download did not work on Alpha 3 for me... and neither works for Beta....

In both cases of loading pets for beta.. several do not give end message... so I have no way of knowing if any of the routine was loaded.. I suspect an error in the background that is not showing.

[seeker]

Cups:

Detected my HP D4160 printer but did not not have the model listed so could not finish the setup. Installed hpijs-2.8.2_static-1 and foomatic-filters 3.0 with the installer and finished the setup. Tried to print and got a print job error "stopped /usr/lib/cups/filter/foomatic-rip-hplip failed". Opened foomatic-rip-hplip and set debug=1 and tried to print. Same print job error, no /tmp/tmp/foomatic-rip log file was created. Tried 8 times from different apps & cups 'print test page'. All gave the same error.

I have always had to install the full HPLIP pet to get my HP printers to work and hope 4.2 will eventually have no need for that.

seeker

[trapster]

ttuuxxx wrote:

It burns just about anything, plus If I add gstreamer It can convert movie formats, I'll try that one tomorrow. well have a try with it. I erased another cdrw with it and burned a puppy image just fine on my testbed with nothing else installed other than puppy live 4.2

You say it burns just about anything...but you only say you've burned an image. What else have you burned successfully to test it?

[rerwin]

Observations, testing on a 400 MHz, 160 MB k6-II PC (frugal):

1. While installing my neglected-modem fix (posted above), I found that the Xarchiver was unable to extract files from the tar.gz file; it simply disappeared, as in crashed. I ran it from the console and saw that it ended with a "segmentation fault".

Feb 16 17:38:06 (none) user.info kernel: xarchiver[21192]: segfault at 0 ip b7b9157b sp bf9a3710 error 4
Feb 16 17:38:06 (none) user.info kernel: in libglib-2.0.so.0.1400.2[b7b5d000+8c000]

The file expanded successfully when I used "tar -zxf modem_and_override_fix_to_4.2b.tar.gz". Sounds like a "show stopper" for the archiver.

2. When initializing my Inte536 PCI modem (one of the neglected ones), the modem was detected automatically. But while starting pupdial from the connect icon, it knew it was there but then claimed it could not detect it! The problem is in the modemtest script where it waits 3 seconds for the modemstats command to receive a response from the modem. The time expires before a response ("OK") is received. If I increase the delay to 3.5 seconds, the modem gets detected by modemtest. This problem occurs only on initialization; probing from pupdial later results in detection. So that is the workaround. But, in the interest of a good first impression, the problem can be avoided by editing line 50 of /usr/sbin/modemtest to:

Code: Select all

 sleep 3.5

3. With the striking default background, I found it difficult to read the displays at the right side of the (14") screen - dark letters on a dark background. And, although the 3D effect works well with the icons (white letters, dark sides/shadows), the lettering on the right side (dark letters, dark sides/shadows) just appear smeared. Maybe the dark letters should have light-colored sides. For me, the most comfortable backgrounds are the light grey and the black aurora australis, on which the letters stand out and are easy to read (even with the smearing). It would seem that a background that is dark on the left and lighter in the mid-to-upper right sides would maximize the readability of the desktop.

In general, 4.2b1 feels a little slower than 4.1.2. I have not been following the progress of this project, so most of these issues may have already been reported. Other than these issues, nice work!
Richard

[panzerpuppy]

OK,so what happens if a user (is being told to and) types "rm -rf /" or "rm -rf *" in the commandline?
Every single bit of personal data (s)he has on the disk will go "poof" in 10 seconds.
Is there any way to block this command or display a warning (3 warnings if necessary) if this happens?

That's one thing I'd really like to see in the final release,if you really care about new users,user experience and security.

'Puppy Data Safeguard' - a proactive measure so new users won't fall for this cheap and dirty trick.

[jrb]

Hi All,
Here is the latest load_sfs. This one restores link to ROX rightclicks, uses "jwm -restart" and "refresh menus" (icewm) instead of "restartwm" (less disruptive) and is bundled with a /root/.xinitrc which contains "fixmenus" just before it starts jwm to clear menus on boot.

You should install zigbert's "fixmenus.tar.gz" from page 5 of this thread so it doesn't fill up your pup_save.

If you don't want the new .xinitrc just rename .pet to .tgz, extract contents, delete .xinitrc and copy to /

Typing this in Opera10.sfs.

[ttuuxxx]

ttuuxxx wrote:

It burns just about anything, plus If I add gstreamer It can convert movie formats, I'll try that one tomorrow. well have a try with it. I erased another cdrw with it and burned a puppy image just fine on my testbed with nothing else installed other than puppy live 4.2

You say it burns just about anything...but you only say you've burned an image. What else have you burned successfully to test it?

When I made that statement, I burnt a puppy iso image booted up on it.To me thats a good test. I ran just fine. I've burnt a DVD+RW since then but I had Gstreamer installed, So I'm going to retry it after without Gstreamer and make sure it does DVDR+W also.
Plus if anyone else would like to test it feel free, I'm working on a lot right now, and can't spend over amounts of time on things. Trying to focus more on 4.2, but I did notice something Xfmedia audio player, uses Xine backend, But it uses way less system resources to play CD's, mp3's, around 1% memory and less than that for CPU.
ttuuxxx

[panzerpuppy]

@ttuxxx: CPU usage can depend on whether audio resampling is used or not.

Before you compare the CPU usage of gxine to other audio players,you should change the the audio resampling setting in gxine.It's set to 'auto' by default.
One more thing to know: ALSA also has its own internal resampling engine.

I use OSS4 instead of ALSA,so I have total control of every feature (disable resampling,disable software mixing,etc.)

On my test box (which uses a much slower CPU than yours),gxine doesn't use more CPU than the other audio players. Actually,it uses less. Other players,like alsaplayer just add a lot more overhead to the 'X' process.

[ttuuxxx]

@ttuxxx: CPU usage can depend on whether audio resampling is used or not.

Before you compare the CPU usage of gxine to other audio players,you should change the the audio resampling setting in gxine.It's set to 'auto' by default.
One more thing to know: ALSA also has its own internal resampling engine.

I use OSS4 instead of ALSA,so I have total control of every feature (disable resampling,disable software mixing,etc.)

On my test box (which uses a much slower CPU than yours),gxine doesn't use more CPU than the other audio players. Actually,it uses less. Other players,like alsaplayer just add a lot more overhead to the 'X' process.

Yes your right about that, but your not the average user, Most users I would say around 90+% would just click a mp3 and play it. I have many friends and if you mentioned the word 'resampling' to them, they would give you a confused dog look. The reason why this uneducated problem exist in the first place is home multimedia players, you pop a music cd in and hit play, Thats it. you have no control. So people get a pc pop a music cd or mp3's in and hit play. What they hear is what they think is prebuilt and the best settings. I've always been an audiophile ever since my first tube amp, to my Mobile DJ business I had in high school to my current computer/bigscreen TV/Home stereo puppy mix. I was the Guy who all my friends went to when they wanted to buy a home/car stereo because i could read spec and understand them very well and also give then a great deal to install them usually a beer or 2 lol. Can't charge friends money for services:)
ttuuxxx

[Pizzasgood]

The icons will now show the ugly symlink arrow, so it is not 100%.......

I agree, it is ugly. This doesn't help remove the arrow, but maybe to fix the ugliness, here is the path to the image responsible:
/usr/local/apps/ROX-Filer/images/symlink.png
(For those who don't know, that is inside a ROX-App - actually, inside the ROX-App, the one for ROX-Filer itself. To get in there just hold shift when clicking on the ROX-Filer directory. Otherwise if you click it you'll just start ROX-Filer).

[ttuuxxx]

WhoDo umumumu <---- caughs On the six printer packages I post a wrong link. The ghostscript package was a previous one. This is the newer, one I also changed the link, on the list as of now,
http://www.puppylinux.asia/tpp/ttuuxxx/ ... 4-i386.pet

Plus I just packaged up the latest X-Chat on page 1
ttuuxxx

[Billwho?]

OK,so what happens if a user (is being told to and) types "rm -rf /" or "rm -rf *" in the commandline?
Every single bit of personal data (s)he has on the disk will go "poof" in 10 seconds.
Is there any way to block this command or display a warning (3 warnings if necessary) if this happens?

That's one thing I'd really like to see in the final release,if you really care about new users,user experience and security.

'Puppy Data Safeguard' - a proactive measure so new users won't fall for this cheap and dirty trick.

Yes a warning for new users is a very good idea but please Do Not!! block the command. I actually use it from time to time ; not a lot but I do use it.

[aragon]

@whodo

some more bugs and updates:
- /etc/ld.so.conf is still missing an entry for /usr/local/lib
- /usr/local/sbin is missing in PATH
- wallpapersetter does not work. to fix delete /root/.config/wallpaper
- mime-type audio_mpeg should be opened with defaultmusicplayer
- on bootup the line where you add your boot-parameters (live-cd) is hard to read (dark blue)
- gadmin-rsync is the old version new: http://www.murga-linux.com/puppy/viewto ... 518#271518
- hdparm is the old version new: http://www.murga-linux.com/puppy/viewto ... 518#271518
- poppler is the old version new: http://www.murga-linux.com/puppy/viewto ... 518#271518
- gadmin-rsync is the old version new: http://www.murga-linux.com/puppy/viewto ... 518#271518
- new version for mtr is here: http://www.murga-linux.com/puppy/viewtopic.php?p=273925
- mtpaint-snapshot: better use dejans advanced version: SEE NEXT POST BY ttuuxxx
- geany 0.16 is here http://www.murga-linux.com/puppy/viewto ... 397#275397

some might allready be reported, but it's hard to remember all posts

aragon

[ttuuxxx]

@whodo

- mtpaint-snapshot: better use dejans advanced version: http://www.murga-linux.com/puppy/viewtopic.php?t=38165

aragon

That snapshot advanced version is already included in the latest Mtpaint I submitted to WhoDo 2 or 3 days ago, so maybe take it off your list??
ttuuxxx

[ttuuxxx]

@ttuuxxx

Hey aragon since it looks like were the only package providers of current software, I was thinking, if you go to sourceforge where puppy is listed we only have 1 current release as of 4.12 listed which is mplayer and thats doesn't even work in the repo how sad ,lol

http://distrowatch.com/table.php?distribution=puppy
sure something we can't update like Glibc or probably xinelib, well we could but I would have to reboot and start fresh and it could take a bit of time and testing to get it right. Wish Barry had configure notes on it
amyways of all the packages we've provided only cups will be the latest. Some of those packages are in the forum but not on the repo ,uuuummmm WhoDo are you going to compare and add some? It would look better for us to have a bit of green and not all black like DSL, lol
anyways I'll do a few.
ttuuxxx

[ttuuxxx]

added latest openssh-client plus made a dev file that includes opensshd which is a terminal connect part and the rest of the package so its complete , Barry never included and its real handy to have so I stuck it with the dev files because it was a 328kb bin file.
so that should be 7 green latest, cups,mplayer, openssh, openssl, httpd, php, mysql

-httpd = The Apache HTTP Server Project is pleased to announce the release of Apache HTTP Server, version 2.2.11. This release represents ten years of innovation by the project, and is recommended over all previous releases!

Ya but you think they could make a proper manual where you don't have to rename or delete over 1000 pages because they name each page like this mixed up in a folder, It took me about 2hrs to rename and delete all the other languages and rename the extensions to actually work, it went from 11mb to 3.8 extract manual.
index.html.en index.html.jp <---so delete the jp one and rename index.html.en --->index.html Just friggen nuts, well someone better enjoy the manual, I won't be ever doing that again for a very loooooooong time.


ttuuxxx
mysql
php
added
openssh-client
openssl
httpd <--- 3 packages for the repo (Apache HTTP Server)

[ttuuxxx]

I've compiled MySQL PHP and package them up tomorrow when I get. up. Strictly for the repo, I guess when Puppy 4.2 finally gets released we'll have some toys to brag about, maybe bring over some new blood developers and figure out a simple non-root accounts for web servers. We'll there's always a need for home test servers for websites also.
ttuuxxx

[WB2009]

no message about 'success' and 'new grub text' when I did a frugal install. no message 'success is installed' when I installed thunderbird.

uncommunicative but otherwise impressed with responsive and impressive puppy!

[rerwin]

The 4.2b1 /etc/rc.d/NMODULESCONFIG file omits mods made for puppy 4.1.1 and 4.1.2. In addition, it propagates an "overkill" response to Lucent/Agere modem driver problems.

1. The 4.1.1 changes were made only to commentary regarding preference of one module over another and do not affect functionality.

2. The 4.1.2 changes removed PCI_OVERRIDE entries for devices previously not detected by the automatic process. However, that upgraded process no longer supports those entries. My "modem_and_overrides_fix" archive file posted earlier in this thread re-implements the function provided by those entries; some of them are no longer needed because the new detection process does work for them.

3. The one change I have made to the preference list is to remove the entry, "martian_dev:ltserial". I make the case for that as follows:

a. It was added because some Lucent modems were reported to not work with the martian_dev driver. But not all; I have two old such PCI modems (11c1:0441 and 11c1:0442) that work perfectly with martian_dev.

b. Martian_dev is the more recent Lucent/Agere driver and moves a half meg of code to user space, reducing, I assume, that much locked-in-RAM storage. It appears to be the preferred technical solution, over ltserial/ltmodem.

c. In 4.1-4.1.2, the old module-override logic is broken, resulting in no module being loaded at all for the affected device. This left only the "preference" technique, which forces use of the old ltserial driver for all such modems and blocks any exposure to the benefits of the newer driver. The above "modem_and_overrides_fix" corrects the problem with overriding the driver for particular devices, allowing a more "surgical" solution.

d. The "enlightened" override technique would require adding a line in the PCI_OVERRIDES assignment statement to specify that ltserial should be used for the vendor/device IDs of a modem reported to be unusable with the automatically assigned martian_dev driver (e.g., ltserial 0x11c1 0x0448). (Example only; 0448 probably can use martian_dev.)

I attach a merged version of the 4.2 /etc/rc.d/MODULESCONFIG file, which is the 4.1.2 version with the PREFLIST assignment statement taken from the 4.2b1 version. Here is the difference listing between 4.2b1 and the merged version, showing only commentary changes, the deletion of the now-ineffective "override" PCI_OVERRIDES entries, and the updated PREFLIST:

8a9
> #put one entry on each line...
10,17d10
< Intel536 0x00008086 0x00001040 #v2.17 this info not builtin to module
< Intel537 0x0000e159 0x00000001 #v2.17 this info not builtin to module
< mwave 0x00001014 0x0000007d #not in modinfo. see hack below.
< ltserial 0x00008086 0x00007110 #v3.98 PCI-to-ISA bridge to possible built-in Lucent modem. (Bridge not present in id list.) Lucent on Intel 82371AB/EB PCI-to-ISA bridge (EIO mode)
< slamr 0x00008086 0x00002486 #v3.98 Agere Systems AC97 Modem
< agrserial 0x000011c1 0x0000048c #v4.00 nothing in modinfo. Agere chip.
< agrserial 0x000011c1 0x0000048f #v4.00 nothing in modinfo. Agere chip.
< yenta_socket 0x00001217 0x00006933 #O2 Micro, Inc. CardBus/SmartCardBus Controller v4.00
36,37c29,32
< #'modalias'. In such a case, here we can specify a preference. Each entry
< #here is of the form 'module1:module2' where module2 is the preferred choice.
---
> #'modalias' (that is, they are both claiming the same hardware). In such a case,
> #here we can specify a preference. Each entry here is of the form
> #'module1:module2' where module2 is the preferred choice.
> #note, list needs a space char at beginning and at end.
8a9
> #put one entry on each line...
10,17d10
< Intel536 0x00008086 0x00001040 #v2.17 this info not builtin to module
< Intel537 0x0000e159 0x00000001 #v2.17 this info not builtin to module
< mwave 0x00001014 0x0000007d #not in modinfo. see hack below.
< ltserial 0x00008086 0x00007110 #v3.98 PCI-to-ISA bridge to possible built-in Lucent modem. (Bridge not present in id list.) Lucent on Intel 82371AB/EB PCI-to-ISA bridge (EIO mode)
< slamr 0x00008086 0x00002486 #v3.98 Agere Systems AC97 Modem
< agrserial 0x000011c1 0x0000048c #v4.00 nothing in modinfo. Agere chip.
< agrserial 0x000011c1 0x0000048f #v4.00 nothing in modinfo. Agere chip.
< yenta_socket 0x00001217 0x00006933 #O2 Micro, Inc. CardBus/SmartCardBus Controller v4.00
36,38c29,33
< #'modalias'. In such a case, here we can specify a preference. Each entry
< #here is of the form 'module1:module2' where module2 is the preferred choice.
< PREFLIST=' rt2500usb:rt73usb ath5k:ath_pci orinoco_nortel:hostap_plx orinoco_plx:hostap_plx orinoco_tmd:hostap_plx orinoco_pci:hostap_pci martian_dev:ltserial bcm43xx:wl b43:wl b43legacy:wl ssb:wl prism54:p54pci r8169:r8101 rtl8187:r8187 '
---
> #'modalias' (that is, they are both claiming the same hardware). In such a case,
> #here we can specify a preference. Each entry here is of the form
> #'module1:module2' where module2 is the preferred choice.
> #note, list needs a space char at beginning and at end.
> PREFLIST=' rt2500usb:rt73usb ath5k:ath_pci orinoco_nortel:hostap_plx orinoco_plx:hostap_plx orinoco_tmd:hostap_plx orinoco_pci:hostap_pci bcm43xx:wl b43:wl b43legacy:wl ssb:wl prism54:p54pci r8169:r8101 rtl8187:r8187 '

Please consider this update for the next beta. Thanks.
Richard

[ttuuxxx]

added php, mysql, I didn't add phpmyadmin, because when I looked at it, theres nothing to compile, its a prebuilt plugin, plus you don't get any green text at distrowatch for it also you pretty much have to have everything setup correctly first then add it. Also please don't think you can just install these packages and away you go, not even close They need a lot of configuring to get them working, More for the advanced admin. LAMP is a better way of doing thing, kind of straight out of the box, But this gets us a bit of brownie points Later on I might do some more packages we get to closer to the final release.
ttuuxxx


@@@Update@@@ I found a better way of compiling MySQL, So I'll redo it and upload again,
ttuuxxx