IRC channel logs

2019-10-05.log

back to list of logs

***catonano_ is now known as catonano
<wingo>jcowan: is the intention that (scheme r5rs) also exports "else" (and a "cond" that assumes "else" is bound), even though "else" isn't specified as bound in r5rs?
<wingo>(perhaps weinholt knows also)
<jcowan>wingo: Yes, I'd say so. It is listed among the exports in Appendix A, along with => and ...
<jcowan>I don't know of any Scheme implementation that lets you export unbound identifiers.
<wingo>jcowan: fwiw it is listed in the base library but the r5rs library just specifies "whatever r5rs defines", which is ambiguous in this context i think
<wingo>a strict reading would not include "else"
<jcowan>App A: The (scheme r5rs) library provides the identifiers defined by R5RS, except that transcript-on and transcript-off are not present. Note that the exact and inexact procedures appear under their R5RS names inexact->exact and exact->inexact respectively. However, if an implementation does not provide a particular library such as the complex library, the corresponding identifiers will not appear in this library either.
<wingo>ah sorry you are right!
<wingo>i misread
<wingo>friggin page breaks
<jcowan>The word "defined" was perhaps il-chosen
<jcowan>Note also that it does not say they have only R5RS semantics: `log` should probably accept two arguments.
<wingo>and curiously that list does not include "else"
<wingo>the list on p.76
<wingo>which implies that that "cond" implementation expects "else" to be unbound
<wingo>right?
<jcowan>Okay, my bad. I was looking at R7RS-with-errata. Erratum 22 reads "22. In Appendix A, the (scheme r5rs) library should export syntax-rules, else, ..., => and _."
<wingo>ah, ok
<jcowan>Technically the errata are not _official_, in the sense that the Steering Committee has not met to ratify them, but I certainly recommend their use
<jcowan> https://bitbucket.org/cowan/r7rs-wg1-infra/src/default/R7RSSmallErrata.md
<wingo>jcowan: should there be a similar erratum for symbol=?
<wingo> ?
<jcowan>What's wrong with it?
<wingo>it should error when applied to non-symbols
<wingo>same erratum as boolean=?