Puppy Linux Discussion Forum

[ttuuxxx]

At this point I think I'll be too busy building "Puppy Linux 2.20" to help out with the compiling.
2.20 is basically a woof build of 2.18 with major updates. But I can offer some ideas.
latest Gtk,Glib,Pango,Cairo,Atk. Previous Gtk has issues, the latest cutting edge version is much better.
We've done jwm/icewm maybe its time for xfce with rox? That would be nice, Thunar is just too thunar, lol
If not then at least have 24x24 menu icons all around
ttuuxxx

[dogone]

I'd prefer to see this thread tee off with the following working hypothesis:

"Puppy 4.4CE will be superior to Puppy 4.3 and any other Puppy derivative in the following respects:"

Everyone simply promoting their favorites apps and utilities will not a great Puppy release make. There has to be a central, well defined "reason" for 4.4's existence. What makes 4.4CE imperative? What must the Puppy community have that is not offered in 4.3 or by another existing Puppy derivative?

These questions really need to be addressed BEFORE a 4.4CE project is kicked off.

[Patriot]

Hmmm .....

[ttuuxxx]

I nominate Patriot as the next coordinator sorry Patriot I just couldn't stop myself
ttuuxxx lol

[ttuuxxx]

If you think about if Patriot was the next leader of 4.4 it could make some real good reading entrainment
better yet see below, LOL

[James C]

You misspelled "titans".

Know what you mean though........

[ttuuxxx]

fixed lol thanks, damn typos no wonder why I have so many tt's uu's xxx's in the word tux, lol
ttuuxxx

[01micko]

Well said Patriot! (I know your not joking but it is mildly amusing).

Ecomoney, all these "issues" will be addressed in due course I'm sure.

Dogone, you have some great points there, that is, there must be a point to 440. Let's see how 430 goes first eh?

ttuuxxxxxxx, (er typo?)

[technosaurus]

@ttuuxxx - I have already started recompiling the basic libraries in 4.3b in order to build the next stable releases of gnumeric and abiword (and to test development builds). I may rethink using the latest stable builds and use the ones that you are using in 2.X so that packages can be somewhat interchangeable. I made and installed them without even having to reboot (I even deleted the old and restarted X - working great so far)

On the plus side, I have been able to build most of it with a net decrease in size ~1MB so far. Statically linking libraries that are only used by one binary also saves close to 1MB thus far (this is somewhat difficult to track, so I made a script that uses ldd and sort to make it easier). Also using rdfind on /usr in 4.3b yields another 1/2 MB (mostly images, fonts and timezone info).

Would it be too much to ask, to have packagers move these files (images, fonts, etc...) to a specified directory and make a symlink? This applies to themes as well - it would be faster smaller and easier if a symlink was changed rather than actually copying the entire contents.

@dynamic sfs loader advocates - I agree - Choicepup is pretty sweet

As for browsers, the demo browser (basically arora) in QT-4.5.2 has been the best so far as size vs speed vs functionality (yes it plays flash and media files) For the less than the size of seamonkey we would have the ability to add some excellent QT based apps (same goes for Opera there I guess - it's a bit bigger but you also get mail, torrent and an easy web server)

FYI Here's the list of packages that I am currently running (recompiled) on 4.3b:
aqualung-0.9beta10+svn atk-1.26.0 ayttm-0.6.0 cairo-1.8.8 geany-0.18 glib-2.20.5 gtk+-2.16.6 lame-398-2 libpng-1.2.40 mtpaint-3.31 notecase-1.9.8 osmo-0.2.8 pango-1.25.6 sweep-0.9.1 xz-4.999.9beta zlib-1.2.3 and attempted cups-1.4.1(no printer to test with though prints to pdf ok)

p.s. I am updating my Pcompile gui to automatically set CFLAGS CXXFLAGS and LDFLAGS for smaller faster packages. Look for it shortly after the 4.3 release.

@ecomoney - patches are always welcome
Most of the driver issues can be answered by community released zdrv files

[ttuuxxx]

[technosaurus]

Latest geany is 683k without any drastic code hacking. The size is mostly due to the scintilla improvements. I am still an advocate of putting it into the devx since it is mostly for highlighting code that you would mostly only need if you also had the devx installed. If code browser does C and C++ that is the most important followed by genie and vala

I almost forgot to mention my little experiment with Ermine pro. The developer granted me a full license and it basically allows you put a whole package into a single binary. upx doesn't work with it but you can easily put it in an sfs for a compressed, simple 2 click run. It's not choicepup but it works... results with my abiword tests have been promising thus far. This would be a good option for those programs that don't make the final cut.

[DaveS]

Its a testament to the helpfulness of this forum that I can now comprehend (if not fully understand) most of what Technosaurus wrote. Cool.........

[Lobster]

I have written to Technosaurus and Gray about coordinating the project.
I am pleased to say Technosaurus has replied:

[DaveS]

If the last episode is anything to go by, whoever takes on this mantle is going to need a thick hide.

[ttuuxxx]