Re: Packages build, news and bugs

[Christopher Rogers <slaxemulator@xxxxxxxxx>]

> Patches ? I thought that as mainstream 2.6.38 kernel has squashfs + xz 
> compression it only remains aufs patches which will be forever out of 
> mainstream. Ubuntu has 2.6.38 kernel with aufs patch.

This is mostly cause of the way slitaz works by loading everything in ram with lzma initrd file. But to make the desktop stable i think it needs some custom code for lzma to work right or something.

Also custom code is needed for gpxe. BTW, there is some drama now with pankso my just be trying to break tazwok for real now. 

zlib: DONT BREAK TAZWOK COOKORDER. Depends and build depends are need for toolchain to be build right.
http://hg.slitaz.org/wok/rev/686e6877305b

zlib: Tazwok must work differentyl all pkgs dep on glibc!
http://hg.slitaz.org/wok/rev/d1ad1340c827

Problem with the cookorder is it only works if you ASSIGN depends and build depends. This is very true with tazwok cookorder. It doesn't matter if glibc-base is installed if its push up the list before gcc or glibc-base. So i know what i'm doing pankso. Don't touch the receipts i'm loading in least it breaks something. These chanages will not break anything.

Now i'm having stress again.

2011/5/24 Philippe Lelédy <slitaz@xxxxxxxxx>

Le 23/05/2011 12:06, Christophe Lincoln a écrit :

Great. When tools are built, I will start thinking about porting
SliTaz to powerpc

What is the point ? Who needs that ?

ME ! I have a bunch of old PowerPC MacBook waiting for re-birth by SliTaz.
I also have enough processing power (Intel 2500K + 8Gb+ SSD) to rebuilt from scratch and anyway I always have to customize SliTaz.
Furthermore, I think SliTaz will be much more robust if it can't be rebuilt for more than one platform !

Packages size
-------------
One new thing I also introduce is a compilation flag for GCC wich
optimize packages for size. The option is -Os and we gain on all
packages wich means a lot of megs in the all distro! I dont have all
core packages builded yet, but I hope we will go back under 30Mb
with this new flag.

This makes me feel very uncomfortable. What is the point ? Who needs
that ? Is it only for entering Guiness book of records ?

Yes, just for the fame :-)

Thks. It's exactly what I was afraid of.

My decision about using or not using SliTaz doesn't depends on a few
less Mb, but on

-1- Boot features: both boot speed and boot flexibility ( network )
-2- Execution speed and RAM used
-3- Ease of customization
-4- Fresh kernel 2.6.39

We need help to prepare all patches for a new kernel. Mabe you can

Patches ? I thought that as mainstream 2.6.38 kernel has squashfs + xz compression it only remains aufs patches which will be forever out of mainstream. Ubuntu has 2.6.38 kernel with aufs patch.

help ?

Furthermore, always refering to KISS each time a decision is to be
made, makes me also uncomfortable. True life isn't that simple.
Simple tools solve only simple problems. Making a computer system
simple for the user is a complex task !

Yes real life isn't that simple, so make computing simple for end user
and us! Often when I refer to KISSS it's about how my mother can use
SliTaz and for our build tools we can have one about 3500 lines or one
with 1500, wich one do you think it is simpler to maintain ?

If they deliver the same service, the answer is in the question, except if on has more clear goals and modular design.

Ph.L.

- Christophe

---
SliTaz GNU/Linux Mailing list - http://www.slitaz.org/

---
SliTaz GNU/Linux Mailing list - http://www.slitaz.org/