* update copyright year
* update list of ports XXX portmasters please check!
* fix a few paths
* note that you can also get into OFW via a BREAK if on cereal console
* pkgsrc is at something like 5000 pkgs
* remove a few numbers that would need maintenance otherwise
(and that aren't easy to determine - ports (not) released in binary)
* Yank section on "upgrade path to netbsd 1.6", same for 1.5
* There are no more README.export-control files (it seems?)
* Update compressed+unpacked src set sizes (estimate xsrc & pkgsrc)
* mention updating via CVS besides SUP
* Update list of obsolete /etc/rc.d files, new and removed users & groups
* Update list of sites that host/run our hardware (-redback, +isc, +ssh.fi)
* Create list of portmeisters from scratch, with list-portmeisters.pl
Todo:
* what's the exact number of ports, archs and cpus supported?
* the "new in 2.0" list needs ... um, creation. Badly!
* "The Future of NetBSD" may need updating
* where is xsrc.tar.gz?
* Fix formatting bug in automatically created portmeister-list
in INSTALL.txt at "UCHIYAMA Yasushi" (XXX needs a *roff guru?)
* the "releng" crew disagrees between this file and
http://www.netbsd.org/People/groups/#releng
* Add vipw(8) and pwd_mkdb(8) as alternatives for editing /etc/passwd
* give a few more pkg_add examples (kde, mozilla, bash)
* Recommend Packages.txt and README as pkgsrc documentation
* Mention postfix and /etc/mailer.conf
Todo:
* what's the place for a pkgsrc-stable.tar.gz snapshot for 2.0?
sys/kern/exec_aout.c back in *1995*, apparently the line from my
license notice:
* must display the following acknowledgement:
was accidentally dropped. This mistake was propagated into
hpux_exec_aout.c when it was split out of hpux_exec.c.
(Thanks to hubertf for noticing!)
preprocessor hackery. If supporting Thumb add an 18th undefined
handler for undefined instruction traps taken in Thumb state. Kill
MAX_COPROCS and use NUM_UNKNOWN_HANDLERS instead.
Add support for breakpoints set in Thumb code and hand them off to
GDB as required.
Note: this may need further work. For the momement we unconditionally
convert a Thumb SWI insn into an ARM SWI with the operation code
zero-extended to 24 bits. However, this should really be a property
of the emulation (how do we know that all emulated systems would want
to work this way?), but that would mean pushing all the handling down
into each of the emulators.
