<Ober>
who do you distribute/deploy stuff in picolisp? is there any sort of fasl delivery?
<Ober>
or just a static interpreter and a script?
<beneroth>
basically static interpreter and a script (or multiple scripts)
<beneroth>
usually tgz is used :)
<beneroth>
unix standard way
<beneroth>
what is "fasl" ? can't find anything. package format?
<beneroth>
viaken, right
<beneroth>
with picolisp it is always soo obvious.. in hindsight.
<beneroth>
I used (any) for months before I found out about (format) xD
<viaken>
I'm doing the Advent of Code to get more aquainted.
<beneroth>
cool, great idea
<Ober>
fast lisp
<beneroth>
quicklisp?
<beneroth>
Ober, there exists no library/ package management for picolisp (as of yet)
<beneroth>
so far also not really required, as 1) picolisp libraries tend to be pretty small, usually a single source file suffices 2) much stuff that would be in language-specific libraries in other languages is intentionally not re-implemented in picolisp, existing libraries/tools in C/C++ can be used via (native) and everything with a cli interface can very easily be used/wrapped 3) most picolisp libraries in that sense are part of the picolisp standard distributi
<beneroth>
on (see directories picolisp/lib and picolisp/misc etc). there are some (and growing!) number of libraries/repositories from developers other than Regenaxer but it's so far still a rather short list, mostly linked to from somewhere on picolisp.com, and generally easy to find on github or bitbucket
<beneroth>
I think having such tooling in the longterm would be good & useful. I'm working on some stuff into that direction but only as a side-project, nothing expected to come out this anytime soon. I don't know from anyone else working on something in that area.
<beneroth>
my view is that package management independent from picolisp is a pretty common use case, but all solutions appear to be pretty OS/language/stack-specific.. I would like to find an reusable universal solution.
<beneroth>
I will look into quicklist, didn't stumble on it yet
<beneroth>
thanks Ober !
<beneroth>
quicklisp
<beneroth>
argh
<beneroth>
or if you indeed didn't mean quicklisp, please provide me with a link :)
beneroth is now known as away
away is now known as bene|off
<viaken>
There's got to be a better way of turning a TSV into a list of numbers? (mapcar format (mapcar pack (split (line) "^I"]
<viaken>
TSV == tab-separated values
<viaken>
Eh, I've gotten used to it.
<Ober>
fasl. fast loading
<Regenaxer>
bene|off: No, I did not know that SHA-512 is easy to attack for hashed passwords
<tankfeeder>
thats why special functions created like bcrypt, pbkdf, argon
<tankfeeder>
harder to brute force
<tankfeeder>
two function implemented on picolisp
<tankfeeder>
argon binded against monocypher and libsodium to compare
miskatonic has joined #picolisp
<Regenaxer>
Why is not a stronger algo then in crypt(2)? Doesn't Unix security depend on it?
Regenaxer has quit [Read error: Connection reset by peer]
Regenaxer has joined #picolisp
<tankfeeder>
openbsd has bcrypt behind crypt function
<Regenaxer>
not portable
<Regenaxer>
Why not in glibc?
<Regenaxer>
if it is really soo insecure
<tankfeeder>
glibc doest care
<tankfeeder>
we should ask aw
<Regenaxer>
Why not?
<tankfeeder>
he was first who told here about it
<Regenaxer>
I see
<miskatonic>
the number between parentheses refers to sections in the manpages, not to versions of the algorithm, doesn't it?
<bene|off>
yes it still needs a bit of a dedicated hardware setup, but the we're talking about "enthusiastic single person"-level not "NSA building bleeding edge special purpose hardware"-level
<bene|off>
afaik
<bene|off>
inherently because the hashing algorithms used in standard crypt are and always were meant for comparisons, not to store passwords
<bene|off>
why crypt et al use such algorithms I don't know, interesting question.
<bene|off>
or I might be completely wrong, but I don't think so
<bene|off>
bbl
miskatonic has quit [Remote host closed the connection]
miskatonic has joined #picolisp
<bene|off>
(apart from that, I would be interested if Regenaxer can put a rough number (weeks, months) for how long the initial form.l development took - I find it an extraordinary piece of craftmanship)
<Regenaxer>
Still this is not the same at all as plaintext passwords
<Regenaxer>
even rot13 would be better than plain text, protecting from short glimpses into the DB
<Regenaxer>
Indeed form.l took a long time. I started using it in a project imediately, after a week or so, but I remember that after 5 years I thought "now it is round" and thought it finished (it never is finished though ;)
<Regenaxer>
Yes, all my concept, but it is not the first version
<Regenaxer>
First was in the 90s with a Java Form GUI
<Regenaxer>
no, first was plain X11
<Regenaxer>
then Java Form, as a stand-alone application
<Regenaxer>
which was then ported to Java Applets in the browser
<Regenaxer>
The current form.l was started in 2005 or early 2006
<Regenaxer>
tankfeeder, this looks a little like Knight's tour task, not?
<miskatonic>
i've even seen uuencode used as pseudo-encryption
<Regenaxer>
:)
mario-goulart has quit [Read error: Connection reset by peer]
mario-goulart has joined #picolisp
miskatonic has quit [Quit: ERC Version 5.3 (IRC client for Emacs)]
<tankfeeder>
i've solve it.
orivej has joined #picolisp
orivej_ has joined #picolisp
orivej has quit [Ping timeout: 268 seconds]
orivej has joined #picolisp
orivej_ has quit [Ping timeout: 276 seconds]
mtsd has joined #picolisp
mtsd has quit [Remote host closed the connection]
mtsd has joined #picolisp
mtsd has quit [Client Quit]
<m_mans>
Hi all!
<m_mans>
what can be wrong with this: (gui 3 '(+ObjVal +DbHint +TextField) '(nm +GoSe) '(nm +GoSe) 45)
mtsd has joined #picolisp
<m_mans>
it's in chart (+Set +E/R +Chart) - pretty similar to demo application ord.l
<cess11_>
m_mans: maybe '+DbHint needs to be first in the list
<cess11_>
as prefix on a db related class rather than the html component
<Regenaxer>
m_mans, the 'gui' call looks fine
<Regenaxer>
cess11_
<Regenaxer>
I think the order does not matter
<mtsd>
Order does not seem to matter when I have used +DbHint, at least
<Regenaxer>
m_mans, what problems do you get?
<m_mans>
sorry, was afk. Let me check a little more, and I'll tell about the problem
<m_mans>
The problem is: when I clear position deleting item name in the chart row, item object is somehow modified too, it's 'nm attribute has gone
<m_mans>
maybe +ObjVal does so
<Regenaxer>
yes, I would say so
<m_mans>
ok, I see in form.l.
<Regenaxer>
+ObjVal is for object created behind the scenes, just by entering a value
<Regenaxer>
So deletion would be a consequence of that
<Regenaxer>
I always made an explicit delete button though, ie. "x" in the chart
<Regenaxer>
+DelRowButton
<m_mans>
the idea was to have catalogue of item names, filled out "on the fly"
<m_mans>
but without deleting on the fly :)
mtsd has quit [Ping timeout: 240 seconds]
<Regenaxer>
oh :)
<Regenaxer>
Perhaps +Dflt to ensure a value?
<m_mans>
I'm switching to just +String, plus (+DbHint +TextField) etc
mtsd has joined #picolisp
<Regenaxer>
+Need won't make sense in the chart I think
<m_mans>
My chief was against of special interface for such +Item's so far, the problem is rooted there :)
<Regenaxer>
What kind of special interface?
<m_mans>
I mean dedicated form
<Regenaxer>
ok
<Regenaxer>
yes, might be confusing
<Regenaxer>
I would try (gui 3 '(+ObjVal +DbHint +TextField)
<beneroth>
Regenaxer, thanks for the history talk :)
<beneroth>
initial minimal usable product after 2 weeks is... very quick :D
<beneroth>
impressive
<beneroth>
and yeah, I easily believe it took many years to ripen ("reifen", wie ein guter wein oder käse oder sowas)
<miskatonic>
are html5 forms still compatible with html forms of 1994?
<beneroth>
well I would say html forms from 1994 are still compatible to html5 forms...
<beneroth>
I would guess that usually html5 forms are compatible to html from 1994.
<beneroth>
forms didn't change so much afaik.
<beneroth>
<datalist> and some other stuff are quite newer, but I think that even today most forms don't use much of those newer features
<beneroth>
cess11_, great demo video, excellent. Thanks for creating and sharing! :D
<beneroth>
cess11_, I agree you should spread this on reddit etc :)
<beneroth>
cess11_, also thanks for the interesting botnet takedown link. Also another example that sabotaging crypto is not necessary to get the bad guys.
<miskatonic>
html5 databases do not work with pil databases, do they?
<beneroth>
what do you mean by "html5 databases" ? local storage in browser on client side?
<beneroth>
look at it like javascript.. you can't directly program picolisp in browser (because browsers just don't execute pil), but you can produce browser-compatible javascript with your picolisp program.
<miskatonic>
yeah, it was a stupid question; a browser written in pil could interact with javascript, but not in such a way
inara has quit [Quit: Leaving]
inara has joined #picolisp
<beneroth>
well maybe one could make a plugin to make browsers execute pil
<beneroth>
e.g. like Flash worked
<miskatonic>
that sounds like a security risk
<beneroth>
but than you have to make the users install that plugin somehow
<beneroth>
well browsers are pretty security-critical stuff
<miskatonic>
I thought of the model of luakit, where lua can format strings which get then evaluated and executed as javascript
<beneroth>
could be made secure
<beneroth>
sure, this can be done
gtuser has quit [Quit: gtuser]
<beneroth>
much picolisp stuff (e.g. form.l) already does stuff like this
<miskatonic>
is transpiling pil into js, similar to what happens with anglescript and coffeescript, a possibility for an app server?
<beneroth>
well.. everything is possible as long it is express-able in a turing-machine...
<beneroth>
real transpiling (vs. just generating = writing some javascript code to a file / to a http connection) is not really feasible
rgrau has joined #picolisp
<beneroth>
miskatonic, you mean something like this?