freenode/#lisp - IRC Chatlog
Search
9:43:17
phoe
FYI, I've asked blix.com to intervene wrt the troll that repeatedly connects to #lisp from their network. If they don't respond, I'll ban the whole ASN.
10:09:29
pve
Hello! Can I expect closer-mop:class-prototype to work on a condition class on most implementations? It work on SBCL.
10:12:27
no-defun-allowed
I'm fairly sure you cannot expect condition classes to be standard-classes, but I don't know of any implementations that don't do that.
10:15:27
pve
no-defun-allowed: hmm, on SBCL they appear to not be standard-classes, but the method is still defined for condition classes
10:20:56
pve
I need to use class-prototype when doing make-method-lambda to get the generic-function-method-class, but I wonder if it's ok to just do (make-condition 'my-condition) and use that as a "prototype"?
10:25:17
phoe
SBCL has its own condition classes, and everywhere else conditions are standard objects with full MOP support.
10:47:37
adlai
is the purpose for which Franz's LLGPL exists primarily to enable the distribution of their compiler as closed-source?
10:47:54
adlai
phoe: I wrote, a few days ago, that I prefer to take the licensing issue elsewhere. it is a long and nasty conversation.
10:48:44
no-defun-allowed
As far as I know, it's to clean up how late binding interacts with component boundaries.
10:49:19
adlai
I also wrote [in #lispcafe] that the conversation on this topic should be in an officially logged channel.
10:50:39
no-defun-allowed
That also includes redefinitions and generic functions, apparently, but it's just clearing up what's a derivative work and what's not.
11:07:43
adlai
well that's just an extraneous landmine: :ASDF has no slot named ASDF/COMPONENT:LICENCE.
14:15:30
semz
What would be a good way to strip down a save-lisp-and-die SBCL image a bit? 30MB is manageable, but still hurts for what is basically a large script with no overly complicated features.
14:15:35
semz
To preempt the inevitable "what about --script": Installing SBCL on the target is no option, sadly.
14:19:35
scymtym
and it didn't reduce the size? for me, it reduces the size to maybe between 25 % and 30 % of the uncompressed binary
14:29:48
scymtym
in any case, if you use core compression, note the following mild disadvantages: 1) the binary will require libz on the target system 2) startup well be slightly slower (by a small fraction of a second) 3) the compressed core cannot be memory-mapped and thus not be shared between multiple processes. other than that core compression is very nice
14:31:46
scymtym
i thought the Windows build used mingw, not wine. building with wine is (was?) also possible, but /that/ gets really unpleasant
14:37:22
scymtym
(not because wine is bad or anything - i'm making good progress implementing a McCLIM Windows backend using wine)
16:16:05
pxpxp
This doesn't look like a lisp question, yet I can't find an answer on the Internet. Maybe because other languages/frameworks hide this possibility from the developers? So here it is: for those of you who have made websites with redirection (e.g. after login), did you use HTTP redirection or did you directly call the target handler? The second way avoids a round-trip time which seems unnecessary at first glance
16:17:09
pxpxp
e.g. client: POST login; server: Moved Permanently to ...; client: GET ...; server: sends the target page
16:20:47
pxpxp
By "target page", I mean the page that will be served on login success, e.g. the main page
16:25:30
beach
There is work on other backends, including a Windows backend and a browser backend. I don't remember the state of those. You can ask in #clim.
16:25:47
scymtym
Psycomic: what is best depends on your requirements. lists with short descriptions can be found at https://www.cliki.net/gui and https://github.com/CodyReichert/awesome-cl#gui and https://lispcookbook.github.io/cl-cookbook/gui.html#introduction
16:27:11
cl-arthur
pxpxp: Both approaches should work in general. If it's for a login or other such rare occasions, it doesn't seem particularly performance-critical, though.
16:36:27
Xach
pxpxp: you can do whatever you like but i would be concerned with someone looking at their profile page after login (for example), bookmarking it, and being brought back to the wrong place (the login handler URL)