bug#36219: GNU Emacs 26.1 --eval rejects trailing whitespace

[Noam Postavsky]

severity 36219 minor
tags 36219 + patch
quit

<andpuke@foxmail.com> writes:

> Since commit a90d5e6309c0306d931d398506b242c3eb4f40d7 which fixes bug #23159
> the `--eval' command line option signals (error "Trailing garbage following
> expression: %s") on more than one expression recevied.

> This causes some old shell script snippets which has trailing newline for
> formatting no longer works, such as the follow script from
> https://orgmode.org/manual/Batch-Execution.html#Batch-Execution


> Since `eval-expression' accepts additional white space and newlines after a
> single expression, this mismatch behavior of `--eval' should be considered as 
> a
> bug.
>
> My proposed fix is apply `string-trim-right` to received string of expression.
>

I think I'd prefer checking the trailing garbage for whitespace, which
is what eval-expression does.  See following patch:

>From c067e710eb7d6e1606f0cdd76e37fe84104d33d8 Mon Sep 17 00:00:00 2001
From: Noam Postavsky <npostavs@gmail.com>
Date: Sat, 15 Jun 2019 08:40:23 -0400
Subject: [PATCH] Allow trailing whitespace in --eval argument (Bug#36219)

* lisp/startup.el (command-line-1): Don't complain about trailing
garbage if it's only space, tab, or newline characters.
---
 lisp/startup.el | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lisp/startup.el b/lisp/startup.el
index 32051c232c..f5463f2c93 100644
--- a/lisp/startup.el
+++ b/lisp/startup.el
@@ -2376,7 +2376,9 @@ (defun command-line-1 (args-left)
                             (read-data (read-from-string str-expr))
                             (expr (car read-data))
                             (end (cdr read-data)))
-                       (unless (= end (length str-expr))
+                       ;; Allow same trailing chars as minibuf.c's
+                       ;; `string_to_object'.
+                       (unless (string-match-p "[\s\t\n]*\\'" str-expr end)
                          (error "Trailing garbage following expression: %s"
                                 (substring str-expr end)))
                        (eval expr)))
-- 
2.11.0

This is a regression from Emacs 25, and the fix looks pretty safe to me;
although the bug isn't especially severe.  Should it go to emacs-26?