log☇︎
700+ entries in 0.019s
erlehmann: yes, but this cost has already been incurred. unicode actually tried to prevent too many new characters in the past, leading to misfeature han unification.
erlehmann: that turns out to be a problem with the abysmally bad TTF file format.
erlehmann: so if code does not care if there exist 250 or 2500 or 250000 characters, why should i?
erlehmann: one point is that incorporating something in unicode does not increase processing complexity by any means. i actually installed a new version of unifont on an old phone and instantly it could display all the new characters.
erlehmann: asciilifeform i strongly suspect you do not distinguish between complexity and sheer numbers. i saw bloat always as referring to needless complexity, size of something only as evidence, not as bloat itself.
erlehmann: no, link?
erlehmann: some writing systems contain non-abstract symbols. all the same to me, as long as everything fits in 16×16 bits.
erlehmann: can it be that your complaint is more with the egyptian glyphs than with unicode itself? ☟︎
erlehmann: also, for everyday communication i find E1, E2, E3, E4 etc. much better than trying to pass around bitmaps
erlehmann: okay, but most of it. a special interest group (egyptologican geeks) used a glyph standard. unicode swallowed that standard. this is how it usually goes. ☟︎
erlehmann: that's all
erlehmann: what actually happened is that people who wanted to use hieroglyphs passed around codes such as A13 – a standard reference exists (gardiner's sign list) that was incorporated into unicode
erlehmann: switching encodings needs out of band signaling which IMO is worse
erlehmann: what do you suggest as an alternative? ☟︎
erlehmann: asciilifeform the sole purpose of unicode is to replace all other glyph standards. from that follows all else, i.e. incorporation of wingdings, emoji, etc.
erlehmann: titcoin
erlehmann: also, any individual car using systemd and literally crashing because of it is unlikely to contain lennart.
erlehmann: consider PHP still exists
erlehmann: problem is timespan. systemd already could not handle an empty string and executed tasks for users named “0day” as root because names starting with numbers are hard, mmmkay?! it might take a long time before it collapses.
erlehmann: so the cool guy must not have reached maximum snowflake level (where no one else could even think themselves snowflake)
erlehmann: sleepy
erlehmann: sorry
erlehmann: i must so the cool guy has not reached maximum snowflake
erlehmann: i must admit that cheered me up a bit, even though i knew how it would turn out halfway through :)
erlehmann: probably not
erlehmann: mircea_popescu in the past i wondered how people who ask that question never consider someone else might think themselves such an exceptional snowflake.
erlehmann: https://blogs.msdn.microsoft.com/oldnewthing/20110310-00/?p=11253
erlehmann: oh, but then i looked for it and there exist several
erlehmann: i remember a post from “the old new thing” about “why can't we have a method to have our window on top of all others?”
erlehmann: it surely has a very low bus factor.
erlehmann: so much about userland, eh?
erlehmann: madness
erlehmann: > One generic approach is to limit the privileges of font processing code in their respective environments, such as enforcing sandboxing of the FreeType library, or moving the font engine out of the kernel in Windows (which Microsoft has done starting with Windows 10). However, that is mostly beyond our reach.
erlehmann: https://googleprojectzero.blogspot.de/2016/06/a-year-of-windows-kernel-font-fuzzing-1_27.html
erlehmann: if i recall correctly, windows has font handling at an entirely wrong abstraction level (think TTF kernel exploit)
erlehmann: too real
erlehmann: http://btcbase.org/log/2016-11-18#1570036 << i almost smiled ☝︎
erlehmann: (conceptually, at least. non-scalable 8x16 / 16x16 glyphs all the way down.)
erlehmann: i use GNU unifont wherever i can. fits in head. ☟︎
erlehmann: seems both like jobs where being sober helps
erlehmann: > drunk waitress / whore
erlehmann: i might want to figure out the UTF-16 LE issue before that. of course, i can do my injection in UTF-16 LE, but i want to know where it comes from.
erlehmann: don't worry, i'll do the “if you use chrome, wine and gnome, visiting random websites might yield vbscript” writeup sometime soon. it's too hilarious.
erlehmann: first writeup http://news.dieweltistgarnichtso.net/posts/python-base64-langsec-fail.html
erlehmann: :D
erlehmann: phf he
erlehmann: telnet style
erlehmann: it is very simple to just stay connected. PONG
erlehmann: i understand completely, i wrote an IRC client myself once.
erlehmann: seems sane otherwise
erlehmann: it just returns [SOH]PING[SOH]
erlehmann: let me check
erlehmann: phf hmm, might be.
erlehmann: meanwhile, i'll do a writeup
erlehmann: i don't even know deedbot's given name!
erlehmann: proper OPSEC ☟︎
erlehmann: only client so far that answers to PING with unknown
erlehmann: deedbot is funny btw
erlehmann: real quality software that tries to call getpid(2) >70000 times in <5 seconds
erlehmann: because journald spends like 95% of its time trying to find out its own pid
erlehmann: bonus: only affected program is systemd
erlehmann: it became slower, or so fefe claimed
erlehmann: also have you heard about getpid()
erlehmann: only one of those inputs conforms to the expected grammar
erlehmann: b'Vm'
erlehmann: >>> base64.b64decode(b'V=m=0=w')
erlehmann: b'Vm'
erlehmann: >>> base64.b64decode(b'Vm0=')
erlehmann: in other news, the base64 module in python3 will happily eat garbage and pretend it is filet mignon. ☟︎
erlehmann: the only thing that prevented me from immediately gaining execution is that the created script for some reason is UTF-16 LE – and of course my code-injection-via-filename is UTF-8.
erlehmann: i should probably do a writeup of the vbscript thing, it's too funny
erlehmann: trinque i sometimes write on http://news.dieweltistgarnichtso.net/posts/index.html
erlehmann: but in gnome-exe-thumbnailer, the filename is inserted into ad-hoc vbscript that is then executed using wine. fun stuff!
erlehmann: in one, i found the exact same vulnerability
erlehmann: this made me look at a few thumbnailers available for GNOME https://bugzilla.gnome.org/show_bug.cgi?id=784630 ☟︎
erlehmann: so it turns out that thumbnailers are worse than antivirus, regarding parsing things better left unparsed
erlehmann: dignitas, swiss non-profit for assisting with suicide
erlehmann: asciilifeform i think mircea_popescu may be onto something with awareness of transitions. there was this one organization for sucide help (something swiss) and many people who get greenlighted by them actually do not kill themselves.
erlehmann: but i can't really put it together
erlehmann: for some reason the topic makes me think of the book “the queer art of failure”
erlehmann: mircea_popescu thanks for link to article about social meaning of train suicide
erlehmann: number goes up, mood goes up
erlehmann: upvotes, most likely
erlehmann: > There’s a very scary trend online now where young people kill themselves online. How terrible is that?!
erlehmann: this ends on a weirdo note http://hkhousewife.com/travel/tokyo-earthquakes-and-train-suicide/
erlehmann: like, years ago, of course
erlehmann: guy in my school from parallel class suicided himself by highspeed train. there were no buildings around with more than three floors (excluding ground level).
erlehmann: i regret my remark
erlehmann: mircea_popescu 100 floors is reasonably uncommon
erlehmann: ah, but terminal velocity is <--[drag]-- o|< --[gravity]--> (rotate by 90 deg)
erlehmann: oh wait, you meant something different
erlehmann: mircea_popescu of course i have no idea how to calculate drag from human body and believe skydiver websites
erlehmann: suppose people wanted to splat faster, terminal velocity for human in air is around ~200 km/h and is probably not achieved from standard issue rooftop.
erlehmann: fear of heights? accessible places?
erlehmann: shop around for a cheap death!
erlehmann: suicide tourism, lulz
erlehmann: > JR East, the line we were trying to take, is the most popular for suicides because it charges the bereaved families the least.
erlehmann: source?
erlehmann: i am glad such stuff exists, because association with it shows stupid. if i remember correctly, debian response to keybase fuckery was: we consider stuff showing up on keybase compromised, any keys showing up will be shot on sight.
erlehmann: because it's easier (trust us!!!111)