From movitz-devel at common-lisp.net Wed Apr 4 22:22:20 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 04 Apr 2007 22:22:20 -0000 Subject: [movitz-ticket] #3: add trig functions Message-ID: <078.ccffe6c88d30ad66c0bf32d2a88fb3a4@common-lisp.net> #3: add trig functions -----------------------------+---------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: enhancement | Status: new Priority: major | Milestone: Component: compiler | Version: Keywords: ansi compliance | -----------------------------+---------------------------------------------- specifically sin and cos -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 4 22:23:33 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 04 Apr 2007 22:23:33 -0000 Subject: [movitz-ticket] #4: add floating point type Message-ID: <078.276546e633f1c763c33517da107b1b03@common-lisp.net> #4: add floating point type ----------------------+----------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: defect | Status: new Priority: major | Milestone: Component: compiler | Version: Keywords: | ----------------------+----------------------------------------------------- currently floating point constants are converted to ratios. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 4 22:25:11 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 04 Apr 2007 22:25:11 -0000 Subject: [movitz-ticket] #5: some warnings stick around in movitz:dump-image Message-ID: <078.cf834dea09837c96ee981eae9de0cb8f@common-lisp.net> #5: some warnings stick around in movitz:dump-image ----------------------+----------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: defect | Status: new Priority: major | Milestone: Component: compiler | Version: Keywords: | ----------------------+----------------------------------------------------- dump-image prints warnings about undefined functions. After correcting the undefined function reference, the warning shows up again in dump-image, despite no reference in the new code. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 4 22:49:01 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 04 Apr 2007 22:49:01 -0000 Subject: [movitz-ticket] Re: #5: some warnings stick around in movitz:dump-image In-Reply-To: <078.cf834dea09837c96ee981eae9de0cb8f@common-lisp.net> References: <078.cf834dea09837c96ee981eae9de0cb8f@common-lisp.net> Message-ID: <087.4725da7f6f389ef92154bdb45ba3a064@common-lisp.net> #5: some warnings stick around in movitz:dump-image -----------------------+---------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: defect | Status: new Priority: minor | Milestone: Component: compiler | Version: Resolution: | Keywords: -----------------------+---------------------------------------------------- Changes (by ffjeld): * priority: major => minor Comment: The problem is that it is not recorded which particular function(s) it is that calls some other function. The only thing that gets recorded is "this function has been called". That means it's impossible to undo this even when the calling function is re-compiled, since there may be other functions that have made the same call. The solution is to record more information, namely, for each function which other functions it calls. Upon create-image, all such information is re-generated, and so also these lingering warnings will go away. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 4 22:54:43 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 04 Apr 2007 22:54:43 -0000 Subject: [movitz-ticket] Re: #4: add floating point type In-Reply-To: <078.276546e633f1c763c33517da107b1b03@common-lisp.net> References: <078.276546e633f1c763c33517da107b1b03@common-lisp.net> Message-ID: <087.7ab9c5b3daebcf43e849ad75cc1b1a5a@common-lisp.net> #4: add floating point type -----------------------+---------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: defect | Status: new Priority: minor | Milestone: Component: compiler | Version: Resolution: | Keywords: numbers -----------------------+---------------------------------------------------- Changes (by ffjeld): * priority: major => minor * keywords: => numbers Comment: Personally I am unsure about the value of floats, particluarly so long as they would have to be boxed (i.e. not be immediates). I won't stop anyone else from implementing floats, but as of now I am unlikely to do it . However I see this as another good reason for going 64-bit, where immediate floats would be feasible. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Mon Apr 9 22:40:42 2007 From: movitz-devel at common-lisp.net (movitz) Date: Mon, 09 Apr 2007 22:40:42 -0000 Subject: [movitz-ticket] #6: shallow-binding doesn't work Message-ID: <078.d38933a290138628460e295372f0924d@common-lisp.net> #6: shallow-binding doesn't work ----------------------+----------------------------------------------------- Reporter: ffjeld | Owner: ffjeld Type: defect | Status: new Priority: major | Milestone: Component: run-time | Version: Keywords: | ----------------------+----------------------------------------------------- With shallow-binding, sometimes (don't know when), the bindings seems to change to "older" values, or at least completely wrong values. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Mon Apr 9 22:44:50 2007 From: movitz-devel at common-lisp.net (movitz) Date: Mon, 09 Apr 2007 22:44:50 -0000 Subject: [movitz-ticket] #7: No macros in run-time Message-ID: <078.d2a536161e5a312869a6f66ec6772ad0@common-lisp.net> #7: No macros in run-time ----------------------+----------------------------------------------------- Reporter: ffjeld | Owner: ffjeld Type: defect | Status: new Priority: major | Milestone: Component: compiler | Version: Keywords: | ----------------------+----------------------------------------------------- Macros need to be dumped into images also, so they are available to eval etc. Currently, no macros work at the REPL. -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Mon Apr 9 22:47:05 2007 From: movitz-devel at common-lisp.net (movitz) Date: Mon, 09 Apr 2007 22:47:05 -0000 Subject: [movitz-ticket] Re: #3: add trig functions In-Reply-To: <078.ccffe6c88d30ad66c0bf32d2a88fb3a4@common-lisp.net> References: <078.ccffe6c88d30ad66c0bf32d2a88fb3a4@common-lisp.net> Message-ID: <087.97fcd6ec9c715386c373ecfddb5fd7ce@common-lisp.net> #3: add trig functions --------------------------+------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: enhancement | Status: closed Priority: major | Milestone: Component: run-time | Version: Resolution: fixed | Keywords: ansi compliance --------------------------+------------------------------------------------- Changes (by ffjeld): * resolution: => fixed * component: compiler => run-time * status: new => closed Comment: Some rather slow (approximations to) sin and cos have been added (in ratios.lisp). -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 11 22:09:58 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 11 Apr 2007 22:09:58 -0000 Subject: [movitz-ticket] Re: #6: shallow-binding doesn't work In-Reply-To: <078.d38933a290138628460e295372f0924d@common-lisp.net> References: <078.d38933a290138628460e295372f0924d@common-lisp.net> Message-ID: <087.8b65e201d99e9bbc592bfb5dc5a333a5@common-lisp.net> #6: shallow-binding doesn't work -----------------------+---------------------------------------------------- Reporter: ffjeld | Owner: ffjeld Type: defect | Status: closed Priority: major | Milestone: Component: run-time | Version: Resolution: fixed | Keywords: -----------------------+---------------------------------------------------- Changes (by ffjeld): * resolution: => fixed * status: new => closed Comment: The problem was that the unwind-protect compiler mistakenly called dynamic-unwind-next before the cleanup-forms were executed. Therefore, the dynamic environment in the cleanup-forms was erroneous. This would go undetected with deep-binding because here dynamic-unwind-next has no side- effects (i.e. the symbols' value-cells are not touched). -- Ticket URL: movitz movitz From movitz-devel at common-lisp.net Wed Apr 11 22:33:21 2007 From: movitz-devel at common-lisp.net (movitz) Date: Wed, 11 Apr 2007 22:33:21 -0000 Subject: [movitz-ticket] #8: Add weak pointers Message-ID: <078.90f46721feeb23f6764ff055235dc6b9@common-lisp.net> #8: Add weak pointers -------------------------+-------------------------------------------------- Reporter: sabetts | Owner: ffjeld Type: enhancement | Status: new Priority: major | Milestone: Component: compiler | Version: Keywords: | -------------------------+-------------------------------------------------- LiCE requires weak pointers for buffer markers -- Ticket URL: movitz movitz