ELLIOTTCABLE changed the topic of #elliottcable to: a _better_ cult
mcc has joined #elliottcable
<whitequark>
katlogic: pirated copy?
<whitequark>
there's literally no functional difference between licensed and unlicensed st, except for the nag dialog
<whitequark>
and in fact you could get rid of it yourself using gdb, sed and ten minutes of your time
<whitequark>
it's not obfuscated or even stripped. find something like license::check_license() and patch in mov $1, %rax; ret
<whitequark>
retq? idfk
<whitequark>
ELLIOTTCABLE: I'm very much explicitly uninterested in discussing programming languages anymore. At all. Anime however is another question...
<mcc>
i am implementing a language interpreter and this is making less and less sense every day
<whitequark>
mcc: o/
<whitequark>
I remember commenting on your blog. it was odd to realize that there's no way to subscribe to further comments.
<mcc>
there isn't?
<mcc>
i'm using some old wordpress. i don't understand wordpress.
<whitequark>
well, I spent like ten minutes trying to figure it out and eventually gave up -_-
<whitequark>
pretty much no way to have a sustainable discussion that I found. it was sad.
yorick has quit [Remote host closed the connection]
Rusky has quit [Remote host closed the connection]
<katlogic>
You're putting too much faith into how much the binary actually changes between versions :)
<katlogic>
usually only when compiled with new gcc
<katlogic>
which is once every 2 years or so
<whitequark>
I'd still have to update the checksum in the compatibility list or something
<whitequark>
and I hate the very idea of that
<katlogic>
compatibility list of wat?
<katlogic>
nah, its just search and replace
<katlogic>
just like the nag screen crack
<whitequark>
compatibility list of binaries
<katlogic>
Yeah, binary patching is not exactly OCD friendly
<katlogic>
but works most of the time and its super easy for target audience :)
<whitequark>
unfortunately, it's not as simple as flipping a few bits, I'd have to find a free section and add a few jumps
<whitequark>
I'm lost as to what is "super easy" here
<whitequark>
easier than installing a package, even
<katlogic>
whitequark: just sayin, theres empirical evidence to support my claim in comments under that blog :)
<whitequark>
wat?
<whitequark>
the issue described in the comments was related to the fact that the guy didn't select xim as the GTK input method
<whitequark>
it is not in any way whatsoever related to my patching method
<whitequark>
and again, I challenge you to present *any* facet in which binary patching is simpler or easier than LD_PRELOAD
<whitequark>
the LD_PRELOAD hack even works across sublime major versions without changes (the packages are different because of dependencies)
<katlogic>
whitequark: In general, yes, that makes it easier for you, but not for target audience.
<katlogic>
Significant amount of sublimetext users barely know what gcc and gdb is.
<whitequark>
...
<whitequark>
what?
<whitequark>
have you even *read* the *first three paragraphs*?
<katlogic>
nah; its tl;dr
<katlogic>
just read the comments
<whitequark>
oh
<whitequark>
thanks; it's convenient to know that it's pointless to talk to you
<katlogic>
whitequark: Out of curiosity, where you point out to that petrov guy he didn't have input method selected?
<katlogic>
poor soul had to pull your ppa
<whitequark>
"Edit: I've actually make it force GTK_IM_MODULE to xim in version 1.1.1, so it should now work regardless."
<whitequark>
and above: " Try also (together with LD_PRELOAD) exporting GTK_IM_MODULE=xim. From what I've seen in GTK source, this should override any other setting."
<whitequark>
it turns out that LD_PRELOAD thing in the package itself always worked, and he just didn't set GTK_IM_MODULE properly, based on the observed effects.
eligrey has quit [Quit: Leaving]
colbyl has quit [Read error: Connection reset by peer]
<whitequark>
it's bad (doesn't work for my purposes) but not too bad (seems architecturally ok and well-written)
<whitequark>
(bad) the gcode renderer it implements is too primitive for anything moderately advanced (also, javascript). but ok as a first step.
<whitequark>
frankly I still think that, given you want to stay in FOSS world, it's best to translate artwork to gcode with (any) CAM, and then write the rest parametrically, by hand
<whitequark>
pen and paper beats freecad by like a thousand miles, unless you do something really complex, in which case freecad doesn't currently cut it too.
<whitequark>
(e.g. no working assembly workbench...)
<whitequark>
with 3D printers, you have openscad, which works, because transforming STL to 3D printer gcode is fairly trivial
<whitequark>
at worst, you have to add some supports
<whitequark>
but transforming STL to a CNC mill gcode is absolutely *not* trivial, in fact most commercial software does it pretty badly
<whitequark>
starting with the fact that STL is a really shitty format for a model; it can't represent *any* curved elements, only meshes
<whitequark>
with the resolution and principles of work of a 3D printer it's not that bad, but for a mill it is
<whitequark>
e.g. commecial sw does reconstruction of arcs from their straight-line approximation to make the result bearable
<whitequark>
so a proper format to store models would be STEP (or IGES, but IGES is kinda horrible; it dates back to EBCDIC days), which stores the models as NURBs
<whitequark>
but then even transforming such data to trivial 3D-printer gcode is very hard, much less if you add all the insane tricks that proper CAM needs to do, so there's basically zilch of them
<whitequark>
and the opensource foundations for CAD/CAM are truly horrifying. the best thing you can currently find is OpenCASCADE. it's an obscure, overcomplicated, enterprisey API in C++ which makes you wish you'd rather put your hands in a blender rather than try to write code compatible with it
<whitequark>
and there's not much choice otherwise
<whitequark>
</rant>
<mcc>
this sounds awful
alexgordon has joined #elliottcable
<whitequark>
mcc: it's a pain, yes
<whitequark>
I've tried to get started several times, but the math has a nasty habit of defeating me completely.
alexgordon has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<whitequark>
I mean, I sort of suck at trivial analytical geometry, stuff like "determine whether the 2D shape is convex or concave"
<whitequark>
I can do it, but it's extremely cognitively demanding
<whitequark>
and working with NURBs defeats me.
<whitequark>
my dream is to implement a 5-axis CAM, but I'm afraid I'll never actually do that
<glowcoil>
i really want to just actually play with real people now
<glowcoil>
done it *very* little
<alexgordon>
yeah
<glowcoil>
it's cool though, got a lot of things lined up in the near future hopefully
<glowcoil>
hopefully going to be producing/engineering for like a punk-ish girl band my friends are starting (has a good chance of happening but you know how these things are)
<glowcoil>
going to do some cool soundtrackey electronic collabs with a friend hopefully
<glowcoil>
and 2 ensembles with the university hopefully
<alexgordon>
glowcoil: sounds like fun
<glowcoil>
alexgordon: been doing much recently?
<alexgordon>
glowcoil: not music. wwdc tomorrow
<alexgordon>
(not going)
<alexgordon>
but they'll probably announce 10.10 syria
<alexgordon>
so
<alexgordon>
gotta get choc ready
<alexgordon>
glowcoil: then holiday
<glowcoil>
10.10 syria?
<glowcoil>
alexgordon: fun, where?
<alexgordon>
glowcoil: well the rumours were that they were going to call it "syrah", although it might be redwood or yosemite now
<alexgordon>
glowcoil: tropical islands :P
<glowcoil>
cool
<glowcoil>
ohh 10.10
<glowcoil>
ok
<glowcoil>
i always thought they had a lot more cats to get through