Wary Puppy 5.0-final feedback

Please post any bugs you have found
Message
Author
User avatar
technosaurus
Posts: 4853
Joined: Mon 19 May 2008, 01:24
Location: Blue Springs, MO
Contact:

#151 Post by technosaurus »

I never could figure out why opening urls broke in some programs after gtk-2.14 (as upstream devs moved away from the deprecated code)

gtk+-2.23.90 release indicated this:
622125 Note that gtk_show_uri needs gvfs to spawn URLs
(we don't have gvfs)

which means I now know what to look for in abiword, ayttm etc.. to patch them (any others that you can think of where external URLs don't open?)

or would it just be better in the long run to add gvfs, its not really gnome dependent - just a gio module IIRC?

Edit: on second thought, how does it make sense to require gvfs for this? maybe it would be better to patch gtk using getenv("DEFAULTBROWSER") and fork() exec() {apparently system() is not secure?}
Check out my [url=https://github.com/technosaurus]github repositories[/url]. I may eventually get around to updating my [url=http://bashismal.blogspot.com]blogspot[/url].

User avatar
ttuuxxx
Posts: 11171
Joined: Sat 05 May 2007, 10:00
Location: Ontario Canada,Sydney Australia
Contact:

#152 Post by ttuuxxx »

technosaurus wrote:I never could figure out why opening urls broke in some programs after gtk-2.14 (as upstream devs moved away from the deprecated code)

gtk+-2.23.90 release indicated this:
622125 Note that gtk_show_uri needs gvfs to spawn URLs
(we don't have gvfs)

which means I now know what to look for in abiword, ayttm etc.. to patch them (any others that you can think of where external URLs don't open?)

or would it just be better in the long run to add gvfs, its not really gnome dependent - just a gio module IIRC?

Edit: on second thought, how does it make sense to require gvfs for this? maybe it would be better to patch gtk using getenv("DEFAULTBROWSER") and fork() exec() {apparently system() is not secure?}
Hi technosaurus usually they are looking for either ghelp or yelp, one of these days we\ll going to need two scripts called ghelp and yelp and able to convert the string value for default browsers, since yelp is just a mozilla based browser with a small gui over it.
If you have the Gnumeric sources laying around open as text
/root/gnumeric-1.10.12/src/wbc-gtk-actions.c
then do a text search in Geany for ghelp
then change the 2 lines to
argv[0] = (char *)"defaultbrowser";
argv[1] = (char *)"http://projects.gnome.org/gnumeric/doc/gnumeric.shtml";
save it and compile it
now it will open to the online docs in any puppy version
ttuuxxx
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games :)

dogle
Posts: 409
Joined: Thu 11 Oct 2007, 12:41

#153 Post by dogle »

Despite Wary's great virtues, the unexpected absence of Adblock is a showstopper for me.

I've found Adblock so useful this past year, so I'm back to 4.3.1 now. Even sc0ttman's esteemed blacklisting PET. though perhaps a most valuable complement, is no substitute (and neither that nor Adblock are in the official PPM repository).

Sure, I can download Adblock2 (though the former Adblock floated my boat very nicely), but so for that matter could I download Seamonkey (a la Lupu) or most any net-specific stuff.

So I wonder why Seamonkey is in, yet Adblock out? Space saving? (how much bigger is Seamonkey2? .... and that's a barrel of worms for another thread!).

I see that usr/local/lib/X11/mini-icons/adblock.xpm is still lurking in there, so was the omission an oversight?


I am very impressed by Barry's built-in-installation-help-on-first-shutown, which seems to address the second-biggest-problem reported by our collective newbies very nicely.

User avatar
ttuuxxx
Posts: 11171
Joined: Sat 05 May 2007, 10:00
Location: Ontario Canada,Sydney Australia
Contact:

#154 Post by ttuuxxx »

updated tano same link and name, fixed the lib link
ttuuxxx
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games :)

User avatar
ttuuxxx
Posts: 11171
Joined: Sat 05 May 2007, 10:00
Location: Ontario Canada,Sydney Australia
Contact:

#155 Post by ttuuxxx »

dogle wrote:Despite Wary's great virtues, the unexpected absence of Adblock is a showstopper for me.

I've found Adblock so useful this past year, so I'm back to 4.3.1 now. Even sc0ttman's esteemed blacklisting PET. though perhaps a most valuable complement, is no substitute (and neither that nor Adblock are in the official PPM repository).

Sure, I can download Adblock2 (though the former Adblock floated my boat very nicely), but so for that matter could I download Seamonkey (a la Lupu) or most any net-specific stuff.

So I wonder why Seamonkey is in, yet Adblock out? Space saving? (how much bigger is Seamonkey2? .... and that's a barrel of worms for another thread!).

I see that usr/local/lib/X11/mini-icons/adblock.xpm is still lurking in there, so was the omission an oversight?


I am very impressed by Barry's built-in-installation-help-on-first-shutown, which seems to address the second-biggest-problem reported by our collective newbies very nicely.
hi dogle the thing is that extensions in Mozilla based products have two areas where where they can be install,
The first one is usr/lib/seamonkey/extensions/ this one works fine until you start up Seamonkey and it builds a user one for all your custom extensions, located in /root/.mozilla/seamonkey/343e4r34rr.default/extensions now this one is default and you no longer can add extentions to the previous one. since the newer one is default. But the second one has a different 343e4r34rr.default number/name on each pc so it would be impossible to make pets work for adding extentions. because its a different users id on each pc, not a direct location.
ttuuxxx
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games :)

User avatar
ttuuxxx
Posts: 11171
Joined: Sat 05 May 2007, 10:00
Location: Ontario Canada,Sydney Australia
Contact:

#156 Post by ttuuxxx »

If Barry made his Seamonkey/Firefox packages the way I do, then it could be possible.
But first you would have to backup your bookmarks and reinstall any extensions that you've installed.
Barry alters the default font size and other settings tailored for puppy in the root/Seamonkey/2323g232j etc I tailor the settings in /usr/lib/seamonkey/etc etc, That way if you add an extension to /ur/lib/seamonkey/extensions all you would have to do is delete root./.mozilla and restart seamonkey and all the tailored settings would be back in /root/mozilla/seamonkey once again. Where as if you delete root./.mozilla in any of Barry's default installs, you'll lose all the tailored settings that Barry made.
ttuuxxx
http://audio.online-convert.com/ <-- excellent site
http://samples.mplayerhq.hu/A-codecs/ <-- Codec Test Files
http://html5games.com/ <-- excellent HTML5 games :)

User avatar
technosaurus
Posts: 4853
Joined: Mon 19 May 2008, 01:24
Location: Blue Springs, MO
Contact:

#157 Post by technosaurus »

ttuuxxx wrote:Hi technosaurus usually they are looking for either ghelp or yelp, one of these days we\ll going to need two scripts called ghelp and yelp and able to convert the string value for default browsers, since yelp is just a mozilla based browser with a small gui over it.
ttuuxxx
not _usually_ ... that is gnumeric specific, and others may be as well

this is what I am talking about:
(abiword:32138): WARNING **: Operation not supported

may have the same gtk_show_uri issues in recent versions of gimp, thunar, gitg, vitkeyboard, gparted, gedit, shotwell, exo, gnome-mplayer, midori, glade, GNU typing tutor, gobby, frogr, ... a really long list of apps use it now

the question is - do we compile gvfs and recompile gtk, or just patch gtk and recompile - I am game for either way (although PCManFM relies heavily on gvfs, so I lean that way)

this warning appears any time you attempt to do anything involving a url call with gtk_show_uri and possibly others (if you compile abiword in gtk+-2.12 then that whole code block gets skipped and uses an alternate _working_ method) ... but you will get this same warning any time an application tries to open a link using gtk_show_uri - it causes ayttm to lock up

https://bugs.launchpad.net/linuxdcpp/+bug/354148

part of the code is in glib's appinfo.c & it has some gvfs code elsewhere IIRC - so that may mean a whole recompile if it is fixed in glib or gvfs is added

found a possible fix using xdg-open: here
void show_uri(GtkWidget *widget, gpointer _uri)
{
/* Because gtk_show_uri is broken */
gchar *uri = g_shell_quote(_uri);
g_message("Showing %s", uri);
gchar *cmd = g_strdup_printf("xdg-open %s", uri);
g_spawn_command_line_async(cmd, NULL);
g_free(uri);
g_free(cmd);

//GError *err = NULL;
//if (!gtk_show_uri(NULL, uri, GDK_CURRENT_TIME, &err)) {
// g_warning("error showing `%s' %s", uri, err->message);
// g_error_free(err);
//}
}

I had written a small script to replace xdg-open before we started including it (for some app I forget which) based loosely on defaulthandler but more simplified

basically it did this:
file:// gets stripped and opened in rox
ftp:// and http:// opens in defaultbrowser
all else are rox $@ #uses rox's mime handling
(unresolved - links that don't have http:// and end in other than htm or html? ... possibly [ ! `rox $@` ] && defaultbrowser $@ ... can't remember exactly)
Check out my [url=https://github.com/technosaurus]github repositories[/url]. I may eventually get around to updating my [url=http://bashismal.blogspot.com]blogspot[/url].

User avatar
jemimah
Posts: 4307
Joined: Wed 26 Aug 2009, 19:56
Location: Tampa, FL
Contact:

#158 Post by jemimah »

Rox has URI handling too - you just need to setup the symlinks in /root/Choices/URI

Code: Select all

# ls -al /root/Choices/URI/
total 4
drwxr-xr-x  2 root root  110 2011-01-06 16:43 .
drwxr-xr-x 11 root root 4096 2010-06-14 18:34 ..
lrwxrwxrwx  1 root root   33 2011-01-10 16:40 cdda -> /usr/local/bin/defaultmediaplayer
lrwxrwxrwx  1 root root   33 2011-01-10 16:40 dvb -> /usr/local/bin/defaultmediaplayer
lrwxrwxrwx  1 root root   33 2011-01-10 16:40 dvd -> /usr/local/bin/defaultmediaplayer
lrwxrwxrwx  1 root root   29 2011-01-10 16:40 file -> /usr/local/bin/defaultbrowser
lrwxrwxrwx  1 root root   13 2011-01-10 16:40 ftp -> /usr/bin/gftp
lrwxrwxrwx  1 root root   29 2011-01-10 16:40 http -> /usr/local/bin/defaultbrowser
lrwxrwxrwx  1 root root   27 2011-01-10 16:40 mailto -> /usr/local/bin/defaultemail
lrwxrwxrwx  1 root root   33 2011-01-10 16:40 rtsp -> /usr/local/bin/defaultmediaplayer
xdg-open:

Code: Select all

#!/bin/bash
case "$1" in  
        '') exit
                ;;
        *://*) exec rox -U "$1"
                ;;
        *@*.*) exec rox -U "mailto:${1}"
                ;;
        *) exec rox "$1"
                ;;
esac

User avatar
technosaurus
Posts: 4853
Joined: Mon 19 May 2008, 01:24
Location: Blue Springs, MO
Contact:

#159 Post by technosaurus »

jemimah wrote:xdg-open:

Code: Select all

#!/bin/bash
case "$1" in  
        '') exit
                ;;
        *://*) exec rox -U "$1"
                ;;
        *@*.*) exec rox -U "mailto:${1}"
                ;;
        *) exec rox "$1"
                ;;
esac
That should work - but could it be trying to open mailto:mailto: in that part (I'm not sure of the mechanics)

should I patch gtk only to use xdg-open then and just comment the glib code or is it too late in the game and we need to patch all the individual apps

btw the actual xdg-open relies on gvfs now too
Check out my [url=https://github.com/technosaurus]github repositories[/url]. I may eventually get around to updating my [url=http://bashismal.blogspot.com]blogspot[/url].

User avatar
jemimah
Posts: 4307
Joined: Wed 26 Aug 2009, 19:56
Location: Tampa, FL
Contact:

#160 Post by jemimah »

Nah, if you give it a URI it breaks out before it gets to that line. That bit is not required anyway - that's just so I can open an email address using Rover.

I think patching glib is the best solution. It needs the patch to look for mounted drives under /mnt rather than /media anyway.

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: PPM issue

#161 Post by BarryK »

gcmartin wrote:Running LIveDVD Wary 5. SM is steady and I'm not getting any issues with anything else thus far....except for 1. PPM.
Seems after installing a couple of packages, I get a PPM Windowbox with message as follows:

Code: Select all

Puppy has searched for any missing 
shared libraries of the packages 
xorg_upgrade_intel-2.10.0-w5
No missing shared libraries

Puppy has examined all user-installed 
packages and found these missing 
babl gegl
what is the recommended course of action when I see this?
Thanks in advance
babl and gegl are dependencies of the gimp package. If you had installed gimp via PPM, normally the deps would also have got installed.

Solution, run PPM and install babl and gegl. Actually, babl is a dependency of gegl, so if you choose gegl in PPM then babl will also get installed.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

#162 Post by BarryK »

gerry wrote:Just upgraded Wary 104 to 500. EDIT: Booted using Wary-500 cd, pressed "Enter" whenever prompted.

Somewhere in the process, I got a screen that said (more or less):

Either uninstall Puppy2 files, or install the following deps: Click "Install" icon when after Puppy is running and has internet access.

{long list}

Well I copied the list onto a piece of paper as instructed, and after reboot clicked on install, and started the package manager, but so far as I can see, it knows nothing about all these dependencies.

Questions:

What Puppy2 files? How would I delete them? Would I miss them if I did?

If I do need these deps, how do I find them?

Could it be the pets that I have installed? The "Check dependencies of installed packages" shows nothing missing.

gerry
The upgrading code in /etc/rc.d/rc.update has been changing.

There is no support for upgrading from any of the Wary betas or RC to the 5.0 final. Try it and you will get strange results.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

#163 Post by BarryK »

DaveS wrote:
Béèm wrote:
DaveS wrote:LOL.. skiboardsonline.com still scrolls awful.
Everything else looks great so far.
Still not convinced that FireFox is bad and that SeaMonkey is ok? 8)
To the Dark Side I have gone.... Opera! Closed source... Heretic...
Beware the Inquisition :)
Yes, I tried it last week, it runs nicely.

There was one problem for me, my blog. If you bring up my blog and look at the code blocks, long lines are wrapped, overlapping and unreadable. On SM, Firefox and Internet Explorer, the long code lines do not wrap.

Code blocks should display as-is, long lines should not wrap, and even if they do they should not overlap.

Annoying, because that one bug makes Opera 11 unusable for me.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: Wireless-Backlight

#164 Post by BarryK »

Frank Cox wrote:Anyway I reset the router and set the security to WEP and re- created
the pass-phrase. No luck.

Reset the router again to disabled-no security and it connected fine. Cab the driver have anything to do with this problem? Is there something I could be doing? i doubled checked the paraphrase in Lubuntu again just to be sure and it worked fine.
It might be due to the older ipw2200 driver in the "old" 2.6.31.14 kernel.

I will be bringing out a variant of Wary soon built with a more recent kernel, that might work.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: compiling ffmpeg

#165 Post by BarryK »

pskin wrote:BK wrote:
As far as I can recall, the only pkgs not compiled for a 486 CPU are ffmpeg and libvpx ...but I am not sure about mplayer. These were configured for a 686 CPU as otherwise the 'configure' script wouldn't use the MMX instructions -- even though MMX was introduced with the 486. Which was annoying, maybe I could have hacked the 'configure' script.
I think that I posted some notes on how to compile ffmpeg, it can be configured for a 486 CPU and recompiled.
Barry, Thanks, I found your blog entry about configuring ffmpeg.
Am I right in thinking that I should configure it for 486 and that
I do not want to use MMX since the Vortex86MX does not support it?
Or am I mistaken and MMX is supported?

I'll have a go at remastering wary for edubook if I can get these few things compiled.

-pskin
The Vortex86mx does have the MMX instructions, but they don't use that name for legal reasons.

The problem is in the configure script of ffmpeg, it disables mmx if you choose i486. I don't know their rationale.

Anyway, choosing i486 and not using mmx is ok, probably slows video rendering a bit.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: Too Much

#166 Post by BarryK »

Frank Cox wrote:I have know spent nearly a whole day trying to get the wireless to work in Wary and have come to the conclusion it is a kernel problem and there is no way to overcome it.
Hope 431 is supported a while longer.
Oh, you had the ipw2200 driver working with encryption in puppy 431 and not Wary? That is odd. I thought that it was just a case of Wary having an old kernel, but 431 has an even older kernel.

Note, there is a log of the connection attempt at:

/tmp/simple_network_setup/rc_network_wireless_connection_log

...take a look at that.
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: ffmpeg and libvpx on vortex86MX (gecko edubook)

#167 Post by BarryK »

pskin wrote:@Barry,
I have managed to compile the latest svn source of ffmpeg
and libvpx. I am not sure how to test these thoroughly.
I tried ffplay on a variety of formats (mpeg, .avi DivX, etc) and managed
to see the movies with sound. I got some alsa pcm underrun messages
on the console, but there seemed to be no problems with the movie.
What other formats should be tested?
How do I test the rest of ffmpeg?
(I am not even sure about what it is supposed to do.)
How do I test libvpx?

I am afraid I hardly know anything about multimedia stuff.
Any pointers from anyone are welcome.

I am not certain about mmx support in vortex86Mx but I went ahead
and configured with --enable-mmx anyway. If problems come up I'll
try changing the config.

For the ffmpeg configuration the --enable-libfaad option apparently
no longer exists since it is not needed, according to some discussion
on the ffmpeg-dev forum.

-pskin
If you test mplayer, that uses ffmpeg.

There are sites that have lots of video test files that you can download. The mplayer site has such a collection -- but I just looked, can't find it. Ah, found it!:

http://samples.mplayerhq.hu/
[url]https://bkhome.org/news/[/url]

User avatar
BarryK
Puppy Master
Posts: 9392
Joined: Mon 09 May 2005, 09:23
Location: Perth, Western Australia
Contact:

Re: wary 5 on gecko (surfboard/edubook)

#168 Post by BarryK »

technosaurus wrote:
BarryK wrote:even though MMX was introduced with the 486. Which was annoying, maybe I could have hacked the 'configure' script.

Adobe Flash -- perhaps the older version 9 player would work?
Re ffmpeg on gecko: mmx was introduced after 486 - the vortexDx is a 586 without mmx or 3dnow (the newer vortexMx does have mmx though) - so it may benefit from -march=i586 -mtune=i586
Thanks for the correction. I had somehow got it stuck in my head that MMX came in with the 486.
[url]https://bkhome.org/news/[/url]

User avatar
Aitch
Posts: 6518
Joined: Wed 04 Apr 2007, 15:57
Location: Chatham, Kent, UK

#169 Post by Aitch »

There seems to be some dispute about MMX, this post suggests 386MMX...

http://www.pcguide.com/ref/cpu/arch/int/instMMX-c.html

...though I, too, thought it was 486 first, as I had a 486/60mhz with MMX, though it may have been a DX2...?

more here, though maybe for a re-introduced 386....

http://www.programmersheaven.com/tags/E ... =Security&

Gosh, I feel old :oops:

Aitch :)

User avatar
technosaurus
Posts: 4853
Joined: Mon 19 May 2008, 01:24
Location: Blue Springs, MO
Contact:

#170 Post by technosaurus »

@Aitch
& I think you misread your source
MMX defines 57 new instructions and the first significant break with forward compatibility since the introduction of the 386
one of the ways Intel saved die space for those 57 new instructions in the Pentium MMX was to share a register with the 387 fpu ... unfortunately they cannot both use it at the same time - causing lower performance due to conflicts... they could have easily added another register later but MS compatibility prevailed ... Instead the "686" or Pentium 3 and beyond have an alternative fpu called "sse" - this can be enabled separately with -mfpmath=sse

... pretty sure you didn't have a 486 with mmx either
cpuinfo would have mmx in flags
>cpu family : 4
>model name : 486 DX/4
>flags: fpu

more details about the pseudo-functions for each instruction set
http://gcc.gnu.org/onlinedocs/gcc-4.0.0 ... -Functions

and how/when/why to enable them when compiling
http://gcc.gnu.org/onlinedocs/gcc-4.0.0 ... tions.html

anyhow - it would be safe to use -march=i586 and -mmmx and even -mtune=i686 (yes you can mismatch) but not necessarily -msse or -m3dnow

gstreamer uses a compatibility layer to interface with different cpu classes differently (orc ... used to be liboil)
Check out my [url=https://github.com/technosaurus]github repositories[/url]. I may eventually get around to updating my [url=http://bashismal.blogspot.com]blogspot[/url].

Post Reply