BLFS 6.3: to be or not to be?

[Alexander E. Patrakov]

Hello,

BLFS SVN still targets LFS-6.3 and is in a deep freeze due to this, and I object 
to this fact due to the following:

People don't report incompatibilities, even as bad as "a package fails to build 
from source", as this is the case for GPM 
(http://linuxfromscratch.org/pipermail/blfs-support/2007-October/063848.html - 
the report has been ignored) and such important multimedia package as Xine-lib 
(http://www.nabble.com/--xine-Bugs-1820958---Linux-2.6.23-header-change-breaks-build-for-xine-lib-td13683064.html). 
There are, of course, many more applications that are broken, and, worse, people 
seem to think that "BLFS is broken" is the normal state of affairs. This will 
stay this way if BLFS targets LFS-6.3 only.

This hurts both books. For BLFS, I don't think that it will be able to catch up 
with such things shortly after LFS-7.0 release, when the bug reporters are 
literally taught to be silent. More minor bug reports are also suppressed by the 
simple fact that such grave bugs exist. And "testing" LFS without an agreed-upon 
continuation (i.e., BLFS) is meaningless. I am afraid of a situation similar to 
http://linuxfromscratch.org/pipermail/lfs-dev/2006-December/058601.html, but 
with some other package.

Then, the freeze hurts multimedia applications. Xine-lib (in version 1.1.8) got 
a fix for "cracky" sound when playing low-bitrate internet radios 
(http://bugs.debian.org/396881). FFMpeg in November got support for J-frames in 
WMV videos 
(http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/2007-November/010311.html), 
and today it started decoding interlaced H.264 videos (available from DVB-T 
broadcasts) without visual artifacts. However, the version in the book is old 
and doesn't support these important features. Besides, it has security bugs (it 
crashes on corrupted H.264 streams that one can obtain from DVB-T in 
Yekaterinburg in areas with not-so-strong signal - now fixed). How can one start 
testing the new version when the book is in such deep freeze without even an 
approximate timeline?

Finally, the LFS 6.3 book itself contains these words:

"If this book is more than six months old, a newer and better version is 
probably already available" 
(http://www.linuxfromscratch.org/lfs/view/6.3/chapter01/changelog.html).

I.e., in less than a month, the only non-obsolete LFS book will be the 
development version. So why target a soon-to-be-obsolete version of LFS now? 
Maybe it is a better idea to state that the BLFS 6.3 release is not going to 
happen at all, and retarget to LFS SVN?

-- 
Alexander E. Patrakov