libera/#sicl - IRC Chatlog
Search
8:26:40
heisig
beach: Yes, I am back home. I've been back home for quite some time now, actually. But I wasn't really productive for the last few days.
8:27:28
heisig
I'm not allowed exercise, and I can only eat liquid food. That takes a toll on productivity. But things are getting better each day.
8:29:04
heisig
The good thing is I already feel fully recovered. But I am not allowed to strain my jaw after the surgery. Hence the restrictions.
8:30:02
heisig
Did I read that correctly in the logs that the custom declaration in the sequence functions is giving you trouble?
8:31:32
heisig
Because I think with the call-site optimization planned for SICL, neither the declaration, nor the sealable metaobject machinery is actually needed.
8:32:10
beach
Well, that's what I am hoping, but I think you should leave it in there in case some other implementation wants to use your sequence functions.
8:33:29
heisig
I understand. Have I missed any other developments in the last few weeks? I think I have seen some progress on register allocation.
8:35:47
hayley
I think I am just about done with register allocation. The last part is to introduce register locations and stack operations to have LIR, which is fairly boring except for adapting arrangements, which I have written code for already.
8:41:11
beach
But I think I am the only one who can do that right now, so it is good that hayley and Gnuxie are working on the more independent modules.
14:09:10
beach
Given how easy it is in Common Lisp to ruin the global environment during development, I can already see how important first-class global environments will be for the use of an IDE.
14:12:05
beach
With an implementation such as SBCL, I find myself having to restart the system fairly often, and even more so because of the way SICL bootstrapping abuses the SBCL global environment. It would seem reasonable that this action would then be replaced by ditching the current first-class global environment and create a fresh one.
16:07:29
beach
Enough work for today. I'll hang around a bit and then I need to go fix dinner for my (admittedly small) family. I think all the phase-specific imports from the host now create instances of SICL SIMPLE-FUNCTION in E5, but there is still a function that is called for every environment that still imports "raw" host functions. I need to break that up so that it is different in different phases.