Tiny Core 5.0 Alpha 2 Testing

[coreplayer2]

Thanks curaga

[aus9]

Is this what the gurus have been waiting for?

[ANNOUNCE] xorg-server 1.14.2
Inbox
x

Matt Dew <marcoz@osource.org>
00:18 (6 hours ago)

to xorg, xorg-announce
All,
  My apologies for the couple day delay.   The release was scheduled
for last Thursday but a regression was found.

Emmanuel Benisty reported that one of the commits caused a regression
problem.  I reverted that commit and allowed a couple days for some
additional testing.

Chris Clayton has reported that this release doesn't work with gcc 4.8-snapshot.

Matt


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Alan Coopersmith (1):
       Only call xf86platformVTProbe() when it's defined

Matt Dew (2):
       Revert "dix: pre-scale x by the screen:device:resolution ratio"
       Bump rev from  1.14.1.902  to 1.14.2

git tag: xorg-server-1.14.2

http://xorg.freedesktop.org/archive/individual/xserver/xorg-server-1.14.2.tar.bz2
MD5:  5d36a6483e8e301875131e8302c67727  xorg-server-1.14.2.tar.bz2
SHA1: accece109925f46ac92618850f53c7459261113c  xorg-server-1.14.2.tar.bz2
SHA256: f6ddfb3f1cfff2d69d45657aec626ebe8b5847810154ff7df7ed2072b076c86b
  xorg-server-1.14.2.tar.bz2

http://xorg.freedesktop.org/archive/individual/xserver/xorg-server-1.14.2.tar.gz
MD5:  b70d12b001037839ce2c2e69bc5e7c30  xorg-server-1.14.2.tar.gz
SHA1: 9435fd3c55cb561faac365bd47b66ff5f98092d2  xorg-server-1.14.2.tar.gz
SHA256: 01478a74aaf938ddc22ce3495031c610a6fe5e3dcce16da58b6c5f141256984f
  xorg-server-1.14.2.tar.gz

[Juanito]

I assume here that we compile with libpng.tcz loaded  for applications previously requiring libpng12.so.0 ?

libpng-dev, yes

[aus9]

I hope this does not upset anyone. Having looked at all my build depends for all printer packages and the inter-dependency I have made a decision I hope I don't regret.

In the spirit of epdfview, which on 4x is now just a viewer, I plan to re-compile cups on 4x to be just a cable cups
and introduce a new package ....cups-wl  for those who need wireless. IMHO this is a better service to all users of TC in the long run.
In the short term, it means a delay in my 5x building, but I hope means I can test eventually a cups.tcz (usb cable only) on 5x
--then do the same when avahi-dev is rebuilt or dependencies changed etc for cups-wl.tcz

But Juanito may like to have some input on this, and it is likely to delay his hplip rebuild etc.
Sorry about that. I am guessing I will then have
cups-dev (meaning cable cups)
cups-wl-dev (meaning cable and/or wireless cups)
and am guessing will have a split for cups-filters into
cups-filters-dev.tcz (for usb or cable)
cups-filters-wl-dev.tcz (for cable or wireless)

It will take me probably a few days to build and test. On completion and submission I will announce it in the cups upgrade warning post.

As I have not yet built it, feel free to make comments or suggestions. I am also aware of the bloat of new cups and promise I try to see if I can do something about that as well....but no promises of less bloat...unless you count fewer dependencies?

Gordon

[coreplayer2]

cool thanks

Meanwhile, I ran into some issues with the Realtek firmware in firmware.tcz from tc-4.x     One issue which may be effecting the Realtek's function is that both kernel mode drivers and corresponding firmware have been updated.  These updated drivers from the new kernel do not appear to function with the old firmware files, imagine..  Some cleansing is needed.

So I built the 3.8.10 kernel with all it's modules using the tinycore-config and created an extension of the /lib/firmware files.  This extension is missing a few firmware files found in tc4.x firmware.tcz (realtek wlan firmware for example) of which some may have been added later since the tcz's original build.   

Ok, I have this updated firmware-3.8.10.tcz ext. in use and is working as expected with it's corresponding kernel modules (as far as I can test with the limited appropriate hardware).  At least there are no outdated firmware files in this extension.   I'm stepping lightly in curaga's territory here, so curaga I've pm'd you.

[coreplayer2]

More on kernel mode drivers

I needed support for Intel's latest gigabit Ethernet adapters, so I built the latest e1000e driver.  Thanks Intel !!!   

This driver needs no separate firmware and is a direct update replacement for the e1000e kernel mode driver installed in 3.8.10 kernel.


Now supporting the all Intel controllers made within the last few years incl. Intel's latest Gigabit Ethernet adapters  (the I217 adapter is the controller included with Intel's latest Haswell chip on many new boards).

Intel® 82563 Gigabit Ethernet PHY
Intel® 82566 Gigabit Ethernet PHY
Intel® 82567 Gigabit Ethernet Controller
Intel® 82571EB Gigabit Ethernet Controller
Intel® 82572EI Gigabit Ethernet Controller
Intel® 82573E Gigabit Ethernet Controller
Intel® 82573L Gigabit Ethernet Controller
Intel® 82573V Gigabit Ethernet Controller
Intel® 82574 Gigabit Ethernet Controller
Intel® 82577 Gigabit Ethernet PHY
Intel® 82578 Gigabit Ethernet PHY
Intel® 82579 Gigabit Ethernet Controller
Intel® 82583V Gigabit Ethernet Controller
Intel® Ethernet Connection I217-LM
Intel® Ethernet Connection I217-V
Intel® Ethernet Connection I218-LM
Intel® Ethernet Connection I218-V
Intel® Gigabit CT Desktop Adapter
Intel® Gigabit PT Quad Port Server ExpressModule
Intel® PRO/1000 PF Dual Port Server Adapter
Intel® PRO/1000 PF Quad Port Server Adapter
Intel® PRO/1000 PF Server Adapter
Intel® PRO/1000 PT Desktop Adapter
Intel® PRO/1000 PT Dual Port Server Adapter
Intel® PRO/1000 PT Quad Port Low Profile Server Adapter
Intel® PRO/1000 PT Quad Port Server Adapter
Intel® PRO/1000 PT Server Adapter

While I can reluctantly wrap the new e1000e driver up in an extension,  really is best to replace the original e1000e kernel module with the new build.

??
 

[coreplayer2]

Keep up the good work there aus9,  am looking forward to an epdfview for tc-5.x 

 

[aus9]

epdfview for tc-5.x will be submitted once  I have no build errors for gtk2-dev
epdfview-printer is much delayed and a low priority until all cups in 4x and re-built then 5x
---if no objections or suggestions are rec'd. But its likely that package will either depend on cups-148 or cups (the bloat editon) to offer wireless support

feedback is welcome.

2) heck, if I go split cups, I could also submit
cups-148

cups.org has some server issues but we have the src.bz2 file, maybe our TCZ checker will accept a build script based on it
That will solve the bloat. Whether builders then choose to use it and keep their own dependency bloat issues low is up to them. 

[aus9]

Juanito

re your suggestion on weechat

I'd suggest that weechat is recompiled to use the newer gnutls, rather than gnutls2

my bloatware cups use

--enable-ssl --disable-gnutls --enable-openssl

for 4x (then assume 5x) would you prefer to see that I enable gnutls?

[bmarkus]

When you have a choice use openssl instead of gnutls

[aus9]

bmarkus

thanks will disable gnutls as before.

[curaga]

Re firmware:

As PM'd to coreplayer2, please don't use the kernel source for firmware. It's outdated and lacks many of the files. Use the git repo given as a source URL in the firmware tczs.

edit: I see the URL has changed and the latest is https://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/

Re e1000e:

I wouldn't really want to replace the upstream module with a third-party one, even an open-source one from Intel. It may have other complications wrt kernel version etc. So please create a separate extension for the third-party module.

[coreplayer2]

Ok, I agree the kernel sources are are missing much firmware.    will investigate creating latest firmware extension from the git source,   thanks



re e1000e driver,   Understood.   will do..

[aus9]

offtopic 4x preparation that will impact on 5x preparation
####
This is going to take me a lot longer to reduce bloat on 4x to make it slimmer on 5x.
I have diverted from making cups148 when I realized my build script for ghostscript lacked stripping commands and could be built with fewer dependencies.

Actually stripping was missing from a lot of my builds 

Its not a lot but fontconfig & libpaper are no longer needed.
printing with gutenprint PPD and hp PPD still good

Code: [Select]

gs test.pdf still works. Printing of a pdf not tested at this stage.

I won't submit, at this stage until I get cups148 to be as small as I can
and still print stuff...not yet built.

So far purely ghostscript files go from 22.37 to 22.3 M
with extra savings of 1.48 for fewer depends

My plan is to rebuild again, so ghostscript will depend on libcups148
but that will impact on hplip-2.7 migration

EDIT
firefox (v 22) on 4x now has a pdf viewer and printed my test.pdf

now to move back to cups148

[aus9]

some ppl may still use commands, so I confirmed ghostscript_common loaded but not the full monty
a quick google showed a nice link
http://www.linuxjournal.com/content/tech-tip-using-ghostscript-convert-and-combine-files

Code: [Select]

tc@box:/tmp$ gs -q -dBATCH -dNOPAUSE -sDEVICE=pdfwrite -sOutputFile=/tmp/out.pdf /tmp/test.ps
To save you clicking the image its a pdf being viewed on 4x in firefox
Pretty colours....won't need to take the red pill tonight 

I forgot to mention that ghostscript will have ghostscript_common removed from depends just like the prev maintainer had it