We’ll see what we do with that “m”…
I’ve updated src/etc/ssl.certs.shar, read about the procedure. Someone might want to update the nss certdata.txt file (to mark them as untrusted); for OpenSSL it’s enough to have them renamed. Luckily, the CAs agree wanting to switch to SHA-1 (yuck, broken as well). Only, do they get a new root certificate (probably not), and where does an OS vendor get such news?
Update: This appears to be not needed, as per this comment. However, how credible is this person? Where is proof? I’d recommend everyone who has not yet read it to look at On Trusting Trust. Anyway, the CA certificates are now trusted again; let’s just hope the backdated demo intermediate CA was the only one generated in the meantime. We really need my SSL known_certs proposal, I think. What’s with these MD2 certs anyway? (end of update)
I also ported the e3 editor to MirBSD, it’s 17 KiB i386 statically linked (stand-alone) binary. GNU GPL, but good for custom-made (hehe…) install/rescue system bsd.rd kernels. I would like to add TinyIRC-MirOS as well, but need to get an exception licence for crunchgen(1) use.
Both tinyirc and mksh(1) now use TIOCGWINSZ more properly and aggressively. The shell, in particular, now always has COLUMNS and LINES set.
Yesternight, not only my server crashed in the softupdates code again (*sigh* two RAID 1 rebuilds and a lot of fsck(8)), but also one of these energy saving light bulbs (13W). This proves that they do not have a longer lifetime than regular bulbs.
I oppose the planned ban of regular light bulbs, because, although almost(!) all of mine are energy saving ones, some of the propaganda apparently is wrong, plus they are not good in situations where the light switch is used very often, or when the light is only on for very short amounts, such as in the entrance area of an appartement (say 1½ m²).
What did I do? Meh, just some bootloader hacking.
Today, we gained arc4random_buf(3), arc4random_uniform(3) in libc and arc4random_atexit(), which is undocumented, but called from _exit(2) and all exec-style functions via execve(2) and exect(2). Since omalloc(3) needs arc4random(3) anyway, it is no real overhead, but a great way to make modelling the pool stuff even more difficult. Thusly, arc4random_pushk(3) retires in favour of arc4random_pushb(3) but stays as cpp(1) define.
Next plans are: rewrite src/lib/csu for pcc, make a port of uw-imapd with libc-client.so (maybe LTMIRMAKE?), make the pine and alpine ports use it, make the php-imap extension work with it again, further improve src/kern/ stuff and maybe www/mk/
More deep changes are however still postponed until OpenBSD is merged better.
I also fixed makefs(8) again ☺
Word has reached me that Benny and Przemek are both suffering from illness (nothing bad, just a cold, don't worry) and thusly will not be attending the congress either.
To these who do: happy celebrating!
I have just placed a couple of MirOS CDs and flyers (English and French only though ☺) in the hands of a friend while helping him to move to Berlin. Either him or Benny (bsiegert@) or Andreas (gecko2@) will distribute them at 25C3, or we’ll place them somewhere. I (tg@) will not attend, replaced (ahoka@) probably neither. You might be able to track down Benny in between some of the talks if you have any questions regarding MirPorts and MirOS, though.
To these who do, happy celebrating!
svn2cvs.sh is an mksh(1) shell script able to convert from Subversion repositories (remote access via URLs, needs installed svn client) to RCS/CVS repositories (local access via pathnames, needs installed rcs(1) ci(1) co(1) GNU RCS, as well as cvs(1) GNU CVS, for rcsfile(5) handling). For more information, see the included help as well as the commit message.
If you have any improvements or requests regarding this script, please contact me or the miros-discuss@ mailing list. There is a discussion thread for it.
Today I updated a couple of things (Perl, OpenSSL, OpenSSH, libpng, Lynx, Sendmail) in MirOS-current. Should be pretty much all needed, and stuff should still be working.
XFree86® 4.8.0 is out, we’ll update once we have human resources to do it.
The next snapshot may not contain the newer X11 unless stable. If imported and stable before the #11 release (whose criteria include OpenBSD merge), it’ll be used there.
Today I’ve got an interview at a potential new employer. *crosses fingers* Wish me luck. Or employ me. (Donations to be used for travel and accomodation for “events” are still desired for all of us developers.)
mksh R36b is out, grab it while it’s still hot ☺ There’s an impressive change list for a mere bug-fix version, and this is the first one where ahoka@ has contributed directly. Upgrade is recommended for all users. The new memory allocator has been backed out for stability.
MirOS-current has mksh R36b plus some changes from mksh-current, mostly renaming set -o utf8-hack to set -o utf8-mode denoting said technology is solid after ages of testing. MirOS-stable has R36b.
Yay, the IBM X40 is back. No explanatory letter, just an invoice over 0 € (funny), but it works. The BIOS clock was off by an hour, I suppose they have booted a Windows® – maybe its “driver” has fixed the graphics chip. So I can go back trying to build a working snapshot again.
And never boot a GNU/Linux, especially not KNOPPIX!
PACKSTATION rules! Even though I yet have to figure out how to use it for sending out a snail mail package.
Maybe I should just go and get stuff for soldering odem back to life, though. Anyone want to buy an otherwise perfectly fine working IBM X40 with docking station and laptop bag? It’s not used much even by the previous owner (french company) and in good state.
We have Mirzilla Firetapir now, but I need to figure out how to get a cleanly separated xorg-reachover port for MidnightBSD done… unless I’ll be porting XFree86® (client part only) over.
I’ll still be using nwt this evening for recreation. And the win2k box nocd has RANDEX in Cygwin, using tinyirc, irssi, and mirsirc. I should definitively create a webpage on the site here for randex. Maybe make a tinyirc port – once it can deal with SIGWINCH…
Some search engine bait: Tom Kohnen, thanks for your old nwt laptop. The MirOS OPTU-8 encoding does UTF-8 (WTF-8?) and CESU-8, while OPTU-16 can do UTF-16 in pcc. Mirzilla Firetapir is the Vutral name for Mozilla Firefox, ports/www/firesomething. Vutral too helped to invent the MirOS RANDEX protocol (entropy exchange over IRC), which has nothing to do with the Win32 worm, just random numbers. My graphician – smultron – needs to expand mksh and I suggested, besides letter spelling (em ka es ha, or for the English, em kay es aytch), or the “polish/klingon variant” (m’ksh), or the official expansion “MirBSD Korn Shell”, he call it “MidnightBSD Korn Shell”. The invalid ones were “make shell” or em kay shell, this is something else. On a side note, we’ll be having a logo for mksh(1) soonish.
If you look carefully, you might notice small changes to the website, both HTML and RSS versions. Enjoy!
Well, maybe not Netcraft. But I couldn't resist the joke, having been a regular Slashdot reader for years, seeing its quality sink, until I suddenly could not login any more with lynx(1), best browser ever.
This morning, I tried to power it on from cold state, and could not even see that password (machine + IDE) prompt. After I typed blindly, I was able to use Fn-F7 (seems as if it's not enabled that early in the boot process), but only for less than one second each try, despite unplugging the AC power supply again. Well, I made it through the boot and shut down the box cleanly again, but it confirms that it's not a heat issue. But what else could it have been? At first I thought that maybe the PSU sent too much current or something, but that's also not it. Perhaps Tonnerre is right and it's all Knoppix' fault ;)
Looks like I'm going to do some hacking on nwt today...