libera/#commonlisp - IRC Chatlog
Search
9:35:12
jackdaniel
it is how it shows up in IRC; and yes, I recall the latter nickname but you didn't have it recently
9:49:03
selwynning
if an els submission is accepted, is it possible to make changes to it until the 9th april?
10:25:56
beach
selwynning: There is a period when you must modify your paper to comply with the referee remarks. During that period, you can make other changes as well, but you must obviously make sure that the referee reports are still valid.
10:27:16
beach
selwynning: I just checked the dates, and yes, until April 9, you *must* make changes. Unless of course the referees had no remarks, which would be highly unlikely.
10:29:29
beach
Unfortunately, ELS does not have a second referee round to determine whether the referee remarks were acted upon, either as modifications to the paper or as explanations to the referee(s) indicating why the remark was not acted upon.
10:29:45
bitblit1
Okay, so I basically want to create a wrapper around a condition which has a hard to understand name and error message. I also want to add restarts to it. How do I do this cleanly?
10:30:22
bitblit1
Should I create a condition and then handller-case the code and throw that condition, wrapping the handler-case with a restart-case?
10:30:25
beach
selwynning: So I guess technically you could not act upon the remarks and there would be no documented way of rejecting your paper because of that.
10:44:09
beach
When I see a referee remark that I consider unfounded, I take it as a sign that other readers might have the same remark, so I tend to add a phrase to the papers like "While it might seem appropriate to ..., we think this is not required because..."
11:08:58
beach
Yes, nobody will object, unless you make substantial changes to the core of your paper.
11:59:50
selwynning
whether i am invited to present or not will determine whether i can claim expenses back on my attendance
12:00:37
beach
If your paper is accepted, there will be a presentation programmed for you in the conference program.
12:07:03
beach
A slot for a talk is usually 30 minutes, of which 5-10 minutes are dedicated to questions. So you should plan for a 20 minute talk to make sure you don't run over, and to compensate for those other authors who did run over.
12:09:11
beach
It is infuriating when a speaker does not stick to the time limit. It tells me that that speaker thinks of this presentation as being more important that those of other authors, so that it doesn't matter if those other authors are given less time.
12:16:23
jackdaniel
you may always turn it around and if you want to infuriate beach then plan for 30minutes! ;)
12:34:21
hayley
Given that I hope to present on a stop-the-world (but parallel) garbage collector, it would be in character for me to go over a time limit.
12:41:55
scymtym
maybe you can read the ROOM and schedule pauses when the listener process is sleeping
14:01:48
kami_
hayley: I get 'Made too many states' with one-more-re-nightmare when compiling "(\\s*([0-9a-zA-Z\\+/=]){4}\\s*)+"
14:14:09
hayley
I'll have to poke around later; I have a report to finish for university today, and I had best head to bed now.
16:10:05
Colleen
Unknown command. Possible matches: 8, say, deny, clhs, set, mop, get, grant, time, tell,
16:22:29
JeromeLon
I found several mistakes in the LOOP grammar in CLHS (the text doesn't match the grammar, and implementations seem to agree with the text). Is it well known? Has anyone tried to write a more correct one?
16:24:59
beach
JeromeLon: Also, many Common Lisp implementations use MIT LOOP which does not agree with the standard.
16:25:18
JeromeLon
"Termination-test control constructs can be used anywhere within the loop body", but in the grammar, it's a main clause, so it must come after all variable clauses, so it would forbid (loop repeat 3 for i ...)
16:29:47
beach
I don't recall seeing that in the text, and Khazern respects the grammar. But there is a lot of code out there that don't respect the grammar, so this ambiguity suggests that we should resolve it in favor of the text.
16:34:37
JeromeLon
another case where grammar and text differ is for for-as-arithmetic. "At least one of the prepositions must be used; and at most one from each line may be used in a single subclause." The grammar has a very different set of conditions.
16:36:09
JeromeLon
anyway, it looks like it's not a hot topic, maybe there are no other differences, I'll assume that.
16:44:21
minion
JeromeLon: WSCL: Well-Specified Common Lisp. Pronounce it as "whistle". A project to revise the Common Lisp standard in a non-controversial way. See https://github.com/s-expressionists/wscl
16:48:39
beach
JeromeLon: While the standard is quite good, there are a certain number of bugs still, and lots of undefined or unspecified behavior, even when this is not really justified.
19:10:56
NotThatRPG
JeromeLon: clisp enforces the loop clause order constraint that you cite. (it's one of the many reasons that clisp is a drag to use! ;-) )
19:30:25
jackdaniel
if a loop honors all requirements and then some more (i.e it allows less strict order), does that constitue a non-conformance (as opposted to an extension)?
20:09:46
yitzi
jackdaniel: there aren't any exceptional situations listed for LOOP so out of order clauses are non conformant, but what to do about them seems unclear.
20:13:41
jackdaniel
but implementation handling it is not much different from defpackage :local-nicknames
20:14:51
yitzi
Correct, I probably should have said that in the absence of "Exceptional Situations" implementations are free to ignore the out of order clauses.
20:16:34
NotThatRPG
yitzi: Is that true? Or should the implementation reject them as ill formed? I'm not sure that implementations are free to simply ignore code they consider ill-formed unless the spec explicitly defines it as an Exceptional Situation
20:23:06
ebrasca
After udpating packages I can't open my webpages. I get "Not Found The requested URL / was not found on this server."
20:25:38
ebrasca
I also can't slime-connect to my server, I get "Can't locate module: SWANK-IO-PACKAGE::SWANK-QUICKLISP"
20:35:22
yitzi
NotThatRPG: What would "reject them as ill formed" mean? I haven't seen anywhere in the spec where it states what will happen. Granted sometimes this info is hiding somewhere.
20:36:17
NotThatRPG
Well, if I stick random gunk into a DEFPACKAGE form, I expect the implementation to error, rather than saying "well, PART of this form is fine, so I'll go with that."
20:48:09
NotThatRPG
Right, so I'm saying that you can assume that a CL implementation will reject ill-formed code without any explicit notification that it will do so.
20:58:22
NotThatRPG
I don't know off hand if the spec permits extensions. I assume so, but I haven't read carefully enough. I would have thought they would have said that LOOP *cannot* be extended -- there is NO mechanism for extension, instead of that there is no *standardized* (my emphasis) mechanism for extension
21:03:40
jackdaniel
conforming implementation n. an implementation, used to emphasize complete and correct adherance to all conformance criteria. A conforming implementation is capable of accepting a conforming program as input, preparing that program for execution, and executing the prepared program in accordance with this specification. An implementation which has been extended may still be a conforming
21:03:46
jackdaniel
implementation provided that no extension interferes with the correct function of any conforming program.
21:04:23
jackdaniel
if the standard does not specify, that "out of order clauses" should signal an error, then a conforming implementation may freely extend loop (or provide means of extending it)
21:05:24
jackdaniel
on contrary, if a conforming program does not work on the implementation (because some function has more required arguments than in the standard), then the implementation is not conforming
21:08:59
NotThatRPG
@jackdaniel: I think the question here was more whether an implementation is permitted to "extend" by simply reading arbitrary extra stuff and ignoring it. So would it be ok for an implementation to accept (LOOP FOR X a Ham Sandwich from 0 below 10 ...) without having "ham sandwich" be defined as having some meaning. I'd say no.
21:10:06
jackdaniel
well, it is me who whave asked the question so I have some hypothesis what it meant :)
21:10:55
NotThatRPG
I guess it's a question between allowing someone to add a new construct, or use an existing construct differently
21:11:36
NotThatRPG
I was responding to yitzi who seemed to be claiming that it would be fine for a conforming implementation to simply throw extraneous stuff on the floor. I was claiming not
21:12:21
jackdaniel
I personally don't see a difference (given that we ignore for a second sanity) - if the implementation decides to assume that all programs are correct and segfault otherwise then it is still conforming
21:12:29
ebrasca
After updating libraries I can't open my pages. I get "Not Found The requested URL / was not found on this server."
21:13:38
jackdaniel
(of course we need to keep in mind that signaling an error is often required by the standard)
21:15:10
jackdaniel
sane implementation will make sensible choices for parts that are not specified by the standard
21:16:07
jackdaniel
for example standard does /say/ what happens when clauses are out of order. one sane solution would be to try to deal with that (MIT), another is to signal an error. insane choice would be to iterate 4 times and print "HELLO"
21:19:06
jackdaniel
there's this program called "sl" that punishes user for mistyping "ls" - it animates an ascii locomotive for 5s
21:19:49
jackdaniel
not that important to the point, don't you think?:) but you are most likely right
21:20:41
jackdaniel
either way, my point is that it is one choice or the another, so saying that mit loop has a defect (compared to sicl loop) on that area is not "correct correct"