GNU bug report logs - #55727
29.0.50; Incorrect zero symbol transcoding between applications

Previous Next

Full log

View this message in rfc822 format

From: Po Lu <luangruo <at> yahoo.com>
To: Петров Андрей <an.petrov <at> pay-lab.ru>
Cc: 55727 <at> debbugs.gnu.org
Subject: bug#55727: 29.0.50; Incorrect zero symbol transcoding between applications
Date: Mon, 30 May 2022 21:42:51 +0800

Петров Андрей <an.petrov <at> pay-lab.ru> writes:

> Hello, team!
>
> Please check the following behavior:
>
> 1. Start two instances of emacs -Q.
> 2. Switch to the "*scratch*" buffer of the first one and type the following:
>    M-: (insert (format "%S"  "000\000000")) [RET]
>    You should see "000^@000" string.
> 3. Copy the string into clipboard and paste into the other instance of
>    Emacs (which you have started before).
>    I get "000\0000" instead of "000\000000"
> 4. Try evaluating inserted text ("000\0000")
>    C-c C-e
>    I get "000^@0" instead of "000^@000".
>
> It seems that "^@" is transcoded between applications into "\0"
>  instead of "\000". If I do the same test inside single instance
>  of Emacs I get the correct result.

That's not a bug.  Most programs cannot tolerate NUL bytes inside
selection strings, so the Emacs selection converter deliberately
replaces those inside a string with "\0" before sending the string to a
requestor.

When one Emacs instance owns the selection, the contents of the kill
ring are used directly when yanking text, so that conversion doesn't
take place.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.