From: Mark Oteiza <mvoteiza@udel.edu>
To: Michael Heerdegen <michael_heerdegen@web.de>
Cc: 28254@debbugs.gnu.org, Noam Postavsky <npostavs@users.sourceforge.net>
Subject: bug#28254: 26.0.50; SRFI-2 and-let*
Date: Sun, 3 Sep 2017 18:39:59 -0400 [thread overview]
Message-ID: <20170903223959.GA15782@holos.localdomain> (raw)
In-Reply-To: <87vakzwu5e.fsf@drachen>
On 03/09/17 at 07:48pm, Michael Heerdegen wrote:
>Noam Postavsky <npostavs@users.sourceforge.net> writes:
>
>> I think I'd be okay with dropping support for the S = (S nil) thing in
>> foo-let macros, so that all of the above would give (void-variable x).
>> Although perhaps the incompatibility with plain let would be annoying?
>> To be honest I hardly ever make use of S = (S nil) in plain let either
>> so it wouldn't hit me at all.
>
>I think the main use case is to declare a local variable when you don't
>care about the init value. In the case of if-let, S = (S nil) is not
>useful, since you can't use that binding neither in the "then" clause
>(because it won't be executed) nor in the "else" clauses (which ignore
>all bindings).
>
>Even if an `if-let' form is the result of a macro expansion, the S = (S
>nil) case isn't of any value. So I see no reasons to not drop support
>for it.
If I'm understanding correctly, it is being agreed that
(let ((x 1)) (and-let* (x) x)) ;; => 1
because the macro expands to
(let* ((x (and t x)))
(if x x))
The following patch achieves this, though it breaks some existing subr-x
tests which I haven't yet looked at carefully.
diff --git a/lisp/emacs-lisp/subr-x.el b/lisp/emacs-lisp/subr-x.el
index 849ac19d6a..ec1990110a 100644
--- a/lisp/emacs-lisp/subr-x.el
+++ b/lisp/emacs-lisp/subr-x.el
@@ -83,10 +83,22 @@ thread-last
`(internal--thread-argument nil ,@forms))
(defsubst internal--listify (elt)
- "Wrap ELT in a list if it is not one."
- (if (not (listp elt))
- (list elt)
- elt))
+ "Wrap ELT in a list if it is not one.
+If ELT is of the form ((EXPR)), listify (EXPR) with a dummy symbol."
+ (message "%S" elt)
+ (cond
+ ;; could this be cleaner?
+ ((null elt) (list elt))
+ ((symbolp elt) (list elt elt))
+ ((nlistp elt) (list elt))
+ ((and (null (cdr elt))
+ (atom (car elt)))
+ (list (cl-gensym) (car elt)))
+ ((and (null (cdr elt))
+ (let ((form (car elt)))
+ (or (listp form) (atom form))))
+ (list (cl-gensym) (car elt)))
+ (t elt)))
(defsubst internal--check-binding (binding)
"Check BINDING is properly formed."
@@ -98,7 +110,10 @@ internal--check-binding
(defsubst internal--build-binding-value-form (binding prev-var)
"Build the conditional value form for BINDING using PREV-VAR."
- `(,(car binding) (and ,prev-var ,(cadr binding))))
+ (let ((var (car binding)))
+ (if (and (null (cdr binding)) (atom (car binding)) (not (symbolp (car binding))))
+ `(,var (and ,prev-var ,var))
+ `(,var (and ,prev-var ,(cadr binding))))))
(defun internal--build-binding (binding prev-var)
"Check and build a single BINDING with PREV-VAR."
@@ -122,8 +137,11 @@ if-let*
Each binding is evaluated in turn with `let*', and evaluation
stops if a binding value is nil. If all are non-nil, the value
of THEN is returned, or the last form in ELSE is returned.
+
Each element of VARLIST is a symbol (which is bound to nil)
or a list (SYMBOL VALUEFORM) (which binds SYMBOL to the value of VALUEFORM).
+An element can additionally be of the form (EXPR), which is
+evaluated and checked for nil.
In the special case you only want to bind a single value,
VARLIST can just be a plain tuple.
\n(fn VARLIST THEN ELSE...)"
@@ -134,27 +152,43 @@ if-let*
(not (listp (car bindings))))
;; Adjust the single binding case
(setq bindings (list bindings)))
- `(let* ,(internal--build-bindings bindings)
- (if ,(car (internal--listify (car (last bindings))))
- ,then
- ,@else)))
+ (if bindings
+ `(let* ,(setq bindings (internal--build-bindings bindings))
+ (if ,(caar (last bindings))
+ ,then
+ ,@else))
+ `(let* () ,then)))
(defmacro when-let* (bindings &rest body)
"Bind variables according to VARLIST and conditionally eval BODY.
Each binding is evaluated in turn with `let*', and evaluation
stops if a binding value is nil. If all are non-nil, the value
of the last form in BODY is returned.
+
Each element of VARLIST is a symbol (which is bound to nil)
or a list (SYMBOL VALUEFORM) (which binds SYMBOL to the value of VALUEFORM).
+An element can additionally be of the form (EXPR), which is
+evaluated and checked for nil.
In the special case you only want to bind a single value,
VARLIST can just be a plain tuple.
\n(fn VARLIST BODY...)"
- (declare (indent 1) (debug if-let))
+ (declare (indent 1) (debug if-let*))
(list 'if-let bindings (macroexp-progn body)))
(defalias 'if-let 'if-let*)
(defalias 'when-let 'when-let*)
-(defalias 'and-let* 'when-let*)
+
+(defmacro and-let* (varlist &rest body)
+ "Bind variables according to VARLIST and conditionally eval BODY.
+Like `when-let*', except if BODY is empty and all the bindings
+are non-nil, then the result is non-nil."
+ (declare (indent 1) (debug when-let*))
+ ;; `(when-let* ,varlist ,@(or body '(t)))
+ (if varlist
+ `(let* ,(setq varlist (internal--build-bindings varlist))
+ (if ,(caar (last varlist))
+ ,@(or body `(,(caar (last varlist))))))
+ `(let* () ,@(or body '(t)))))
(defsubst hash-table-empty-p (hash-table)
"Check whether HASH-TABLE is empty (has 0 elements)."
next prev parent reply other threads:[~2017-09-03 22:39 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-08-27 20:11 bug#28254: 26.0.50; SRFI-2 and-let* Mark Oteiza
2017-09-01 2:55 ` npostavs
2017-09-02 2:10 ` Mark Oteiza
2017-09-02 3:05 ` npostavs
2017-09-02 4:14 ` Mark Oteiza
2017-09-02 5:25 ` Michael Heerdegen
2017-09-02 13:36 ` Mark Oteiza
2017-09-02 18:41 ` Noam Postavsky
2017-09-03 17:48 ` Michael Heerdegen
2017-09-03 22:39 ` Mark Oteiza [this message]
2017-09-04 0:48 ` Mark Oteiza
2017-09-04 14:12 ` Michael Heerdegen
2017-09-05 3:47 ` Mark Oteiza
2017-09-05 15:04 ` Michael Heerdegen
2017-09-06 12:12 ` Michael Heerdegen
2017-09-06 13:06 ` Mark Oteiza
2017-09-06 19:04 ` Michael Heerdegen
2017-09-04 1:13 ` Mark Oteiza
2017-09-05 3:55 ` Mark Oteiza
2017-09-09 0:33 ` Mark Oteiza
2017-09-12 12:39 ` Michael Heerdegen
2017-09-12 13:09 ` Mark Oteiza
2017-09-12 18:44 ` Michael Heerdegen
2017-09-12 20:21 ` Mark Oteiza
2017-09-13 10:16 ` Michael Heerdegen
2017-09-13 11:48 ` Mark Oteiza
2017-09-13 16:46 ` Michael Heerdegen
2017-09-13 16:49 ` Mark Oteiza
2017-09-13 17:05 ` Michael Heerdegen
2017-09-13 17:28 ` Mark Oteiza
2017-09-13 17:49 ` Michael Heerdegen
2017-09-12 12:13 ` Michael Heerdegen
2017-09-12 14:29 ` Mark Oteiza
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://www.gnu.org/software/emacs/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170903223959.GA15782@holos.localdomain \
--to=mvoteiza@udel.edu \
--cc=28254@debbugs.gnu.org \
--cc=michael_heerdegen@web.de \
--cc=npostavs@users.sourceforge.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.savannah.gnu.org/cgit/emacs.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).