freenode/#clasp - IRC Chatlog
Search
16:58:59
Bike
call-method is the longest part in this one. for clasp there's a bunch of other junk for the cases where we don't use effective method functions.
17:04:23
beach
I think I will need to attack that in the morning when my brain is still working. This evening, after having had my favorite coauthor over for lunch, and having debugged my method-combination code, I think I am unable to attack some code of this complexity.
17:05:52
beach
But the number of readers may be fairly limited, so perhaps it is better done as a community project with free access to source and PDF.
17:06:21
beach
I had imagined using CreateSpace for on-demand printing as I do for my other book projects.
17:14:30
beach
makomo: It looks like you are interested in SICL/Cleavir. If that is true, do you have something you would like to work on?
17:18:19
beach
OK, today I think I debugged some problems in my implementation of method combinations. Didier Verna told me the design is sane, so that's what I went with.
17:20:00
beach
Given the distractions today (favorite coauthor for lunch, gardener coming over) I think that's a pretty good day after all.
17:22:54
beach
Tomorrow, I'll rewrite COMPUTE-APPLICABLE-METHOD to do what it is supposed to and include code for the consequences of that rewrite.
18:23:33
makomo
beach: i don't have anything particular in mind currently but of what i heard so far -- bootstrapping, FCGEs and Cleavir sound very interesting
18:24:06
makomo
i don't think i'm up to speed with the whole design/goal of the project yet, i still have reading to do
18:26:36
makomo
also, both a new school year and my first job are about to start, so i won't take anything on currently
20:23:51
drmeister
How does Python work? If you have a backtrace and that includes a source file - if you nuke that source file - the code is supposed to stop working right? GARGH - apparently not.
23:29:11
drmeister
I'm finalizing the buildbot stuff. I can launch EC2 instances now to build things.
0:00:01
makomo
Bike: a quick question since you're half familiar with my DSL. one component can have multiple :process blocks. since the order of expanding macros is unspecified, i'm probably going to have a bad time if i'm relying on the fact that (<- ...) will do work at macroexpansion time and somehow associate the usage of <- with a specific process, right?
0:00:50
makomo
Bike: because, if there are multiple :processes, i'm not really going to know which <- is being expanded (and which process to associate it with)
0:01:30
makomo
well there goes the macroexpansion solution. can't run away from the code walker i suppose