Re: dpkg flex-based status file parser, for 35% speedup
> OTOH linking shell like dash statically is another question :)
0. Making conf files "available" and "status" broken down per package basis.
1. Package information -- valid shell scripts (see [G]).
2. Together with "$PKG.*" files from dpkg/info directory it packed
in one file, valid shell script. BTW *list files can be turned into
a shell script easily, to reduce size, for example. Example of
handling current `ls -1`  ones can be found in [G 'cleanup_linux()'].
0. All is one file:
0.1 no waste of (usually 1K-4K) blocks on file system
0.2 checking like md5 summing, tracking one file are easy.
1. Less pressure on directory inode. Time speed-up on every kernel for sure.
2. (Philosophic) Shell is the most basic thing. Thus going back to roots
is another good thing. Forgetting all that dark age, where there was
UNIX zoo with very unreliable implementation of basic tools. This is
where GNU have origin as well as Perl. The former with outrageous
ambitions and NIH affected bash. The latter everything everywhere, non
system, wtf is posix tool (while posix as a standard have ugliness on
itself, it's accepted mostly one).
0. Basic set of utils, like dd, sed, rm, mv, shell may use, is open.
(no, no not "busybox"!!!)
1. "Version:" parsing and handling (i'll try to bring something sh
Shared, since no feedback at all on my crazy propositions :)
-o--=O`C info emacs-manual : not found /. .\ ( is there any reason to live? )
#oo'L O info make : not found o ( yes --- R.I.P. FSF+RMS )
<___=E M man gcc : not found `-- ( viva Debian Operating System )