freenode/#clasp - IRC Chatlog
Search
14:37:04
Bike
no. it doesn't seem to be caused by having inline asts. next i suppose i can disable inlining entirely
14:37:19
Shinmera
drmeister: any idea when you'll have time to look at usocket? Not saying you should get to it now, just wondering.
14:39:20
drmeister
Shinmera: I'll take a quick look at it now and see if there is anything simple that I can fix.
14:39:51
drmeister
You said pull the latest usocket from clasp-developers and then run (asdf:test-system :usocket) . to test it - yes?
14:43:15
drmeister
Shinmera: It's got to be problems in mimicking the few sb-bsd-sockets functions from ecl that clasp mimics.
14:44:53
drmeister
beach: We see eclector default source positions (cons number number) showing up. This should never happen because we provide our own source-position method.
14:46:01
drmeister
Indeed. Bike has been working with cst for months - and got it working and then when I tried to build it - I saw this.
14:48:06
beach
And because DEFMETHOD defines the generic function if it does not exist, the problem can go undetected.
14:53:06
beach
You then have to make sure you define your method on the generic function in the right package.
14:55:07
drmeister
https://github.com/robert-strandh/Eclector/blob/master/code/parse-result/package.lisp#L11
14:56:21
drmeister
Bike: scymtym suggested that we subclass eclector.concrete-syntax-tree:cst-client and specialize on that. We don't do that right now.
14:57:06
drmeister
It's probably not the problem unless eclector redefines the source-position method specialized on eclector.concrete-syntax-tree:cst-client
14:58:04
drmeister
https://github.com/robert-strandh/Eclector/blob/master/code/parse-result/generic-functions.lisp#L5
14:59:27
beach
And are you sure that the client you specialize on is also the client that it gets passed?