aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--etc/NEWS16
1 files changed, 9 insertions, 7 deletions
diff --git a/etc/NEWS b/etc/NEWS
index 7f89ef92ec4..2b8ef4da1d7 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -1361,6 +1361,15 @@ non-nil, but the code returned the list in the increasing order of
1361priority instead. Now the code does what the documentation says it 1361priority instead. Now the code does what the documentation says it
1362should do. 1362should do.
1363 1363
1364+++
1365** 'format' now avoids allocating a new string in more cases.
1366'format' was previously documented to return a newly-allocated string,
1367but this documentation was not correct, as (eq x (format x)) returned
1368t when x was the empty string. 'format' is no longer documented to
1369return a newly-allocated string, and the implementation now takes
1370advantage of the doc change to avoid making copies of strings in
1371common cases like (format "foo") and (format "%s" "foo").
1372
1364--- 1373---
1365** The function 'eldoc-message' now accepts a single argument. 1374** The function 'eldoc-message' now accepts a single argument.
1366Programs that called it with multiple arguments before should pass 1375Programs that called it with multiple arguments before should pass
@@ -1556,13 +1565,6 @@ Emacs integers with %e, %f, or %g conversions. For example, on these
1556hosts (eql N (string-to-number (format "%.0f" N))) now returns t for 1565hosts (eql N (string-to-number (format "%.0f" N))) now returns t for
1557all Emacs integers N. 1566all Emacs integers N.
1558 1567
1559+++
1560** 'format' is no longer documented to return a newly-allocated string.
1561This documentation was not correct, as (eq x (format x)) returned t
1562when x was the empty string. 'format' now takes advantage of the doc
1563change to avoid making copies of strings in common cases like (format
1564"foo") and (format "%s" "foo").
1565
1566--- 1568---
1567** Calls that accept floating-point integers (for use on hosts with 1569** Calls that accept floating-point integers (for use on hosts with
1568limited integer range) now signal an error if arguments are not 1570limited integer range) now signal an error if arguments are not