(loop when ... end)
Unable to parse email body. Email id is 2806
Unable to parse email body. Email id is 2806
> From: owner-lisp-hug@xanalys.com > [mailto:owner-lisp-hug@xanalys.com] On Behalf Of Nick Levine > Sent: 23 September 2004 12:34 > To: lisp-hug@xanalys.com > Subject: (loop when ... end) > > Section 6.1.6 of CLHS states: > > The optional loop keyword "end" marks the end of the clause. If this > keyword is not supplied, the next loop keyword marks the end. The > construct "end" can be used to distinguish the scoping of compound > clauses. > > So I would expect > > (loop for x below 10 when (< x 3) collect x collect x end) > > to work rather than - as in LWW 4.3.7 and for that matter ACL 6.2 - > telling me that end is an unknown loop keyword. > > Have I read the spec wrong? Is this a defect in the implementation(s)? > Or what? > > - nick The second COLLECT has already ended the WHEN, and an END has to match an IF, WHEN or UNLESS. Try this: (loop for x below 10 when (< x 3) collect x end collect x) => (0 0 1 1 2 2 3 4 5 6 7 8 9) I guess the error message is a little confusing though. (And thanks for introducing me to END.)
> From: owner-lisp-hug@xanalys.com > [mailto:owner-lisp-hug@xanalys.com] On Behalf Of Simon Katz > Sent: 23 September 2004 13:42 > To: lisp-hug@xanalys.com > Subject: RE: (loop when ... end) > > > From: owner-lisp-hug@xanalys.com > > [mailto:owner-lisp-hug@xanalys.com] On Behalf Of Nick Levine > > Sent: 23 September 2004 12:34 > > To: lisp-hug@xanalys.com > > Subject: (loop when ... end) > > > > Section 6.1.6 of CLHS states: > > > > The optional loop keyword "end" marks the end of the clause. If this > > keyword is not supplied, the next loop keyword marks the end. The > > construct "end" can be used to distinguish the scoping of compound > > clauses. > > > > So I would expect > > > > (loop for x below 10 when (< x 3) collect x collect x end) > > > > to work rather than - as in LWW 4.3.7 and for that matter ACL 6.2 - > > telling me that end is an unknown loop keyword. > > > > Have I read the spec wrong? Is this a defect in the > implementation(s)? > > Or what? > > > > - nick > > The second COLLECT has already ended the WHEN, and an END has > to match an IF, WHEN or UNLESS. > > Try this: > > (loop for x below 10 when (< x 3) collect x end collect x) > => (0 0 1 1 2 2 3 4 5 6 7 8 9) > > I guess the error message is a little confusing though. > > (And thanks for introducing me to END.) On second thoughts... The spec says "If this keyword [END] is not supplied, the next loop keyword marks the end." So I guess the /first/ COLLECT marks the end of the WHEN (but it is included within the scope of the WHEN). ??
Unable to parse email body. Email id is 2809
> Try this: > > (loop for x below 10 when (< x 3) collect x end collect x) > => (0 0 1 1 2 2 3 4 5 6 7 8 9) > > Interesting. But not what I wanted - my fault for not defining what I > meant by "to work". > > I was hoping it would return (0 0 1 1 2 2). > Do you know about AND? (loop for x below 10 when (< x 3) collect x and collect x) => (0 0 1 1 2 2)
On Thu, Sep 23, 2004, 09:00, Nick Levine <ndl@ravenbrook.com> wrote > > So I would expect > > > > (loop for x below 10 when (< x 3) collect x collect x end) > > > > to work rather than - as in LWW 4.3.7 and for that matter ACL 6.2 > > - telling me that end is an unknown loop keyword. > > ... > > The second COLLECT has already ended the WHEN, and an END has > to match an IF, WHEN or UNLESS. > > Try this: > > (loop for x below 10 when (< x 3) collect x end collect x) > => (0 0 1 1 2 2 3 4 5 6 7 8 9) > >Interesting. But not what I wanted - my fault for not defining what I >meant by "to work". > >I was hoping it would return (0 0 1 1 2 2). In my experience, whenever using any combination of loop clauses that I've not used at least 20 times before, it's best to macroexpand the LOOP expression to see what it produces.
On Thu, 23 Sep 2004 13:42:12 +0100, "Simon Katz" <sk@nomistech.com> wrote: > The second COLLECT has already ended the WHEN If that is the case one wonders what END is good for anyway...
Edi Weitz:
> > The second COLLECT has already ended the WHEN
>
> If that is the case one wonders what END is good
> for anyway...
Hmmm, yes. I think it ties in with using AND.
Here's an example where END makes a difference:
(loop for x from 1 below 10
when (< x 7)
when (evenp x) collect x
end
and collect (* x 100))
=> (100 2 200 300 4 400 500 6 600)
(loop for x from 1 below 10
when (< x 7)
when (evenp x) collect x
and collect (* x 100))
=> (2 200 4 400 6 600)
So I think:
- WHEN has a subsidiary clause (introduced by the
first COLLECT in Nick's original example).
- When the spec says "If this keyword [END] is not
supplied, the next loop keyword marks the end",
this means the next loop keyword after the
subsidiary clause.
Simon Katz wrote: > So I think: > - WHEN has a subsidiary clause (introduced by the > first COLLECT in Nick's original example). > - When the spec says "If this keyword [END] is not > supplied, the next loop keyword marks the end", > this means the next loop keyword after the > subsidiary clause. Maybe you're right. Then again, maybe you're not. LOOP is nice, but the corner cases are a mess. IMHO the best solution is to just use Jonathan Amsterdam's ITERATE when things get complicated. Marc Battyani's CL-PDF (or was it CL-TYPESETTING?) also recently got bitten by different interpretations of some LOOP subtleties. He decided to switch to ITERATE. Problem solved. Arthur
Arthur Lemmens: > Sent: 23 September 2004 15:04 > To: lisp-hug@xanalys.com > Subject: Re: (loop when ... end) > > Simon Katz wrote: > > > So I think: > > - WHEN has a subsidiary clause (introduced by the > > first COLLECT in Nick's original example). > > - When the spec says "If this keyword [END] is not > > supplied, the next loop keyword marks the end", > > this means the next loop keyword after the > > subsidiary clause. > > Maybe you're right. Then again, maybe you're not. > > LOOP is nice, but the corner cases are a mess. Yes. I wouldn't use those ANDs and ENDs in real code.
Simon Katz wrote: > Arthur Lemmens: >> Maybe you're right. Then again, maybe you're not. >> >> LOOP is nice, but the corner cases are a mess. > > Yes. I wouldn't use those ANDs and ENDs in real code. I use AND. I rarely, if ever, use END. (I tried grepping for END in my code, but that was useless. I use it far too often as a variable name.) Arthur