GNU bug report logs - #18812
24.4; url.el user agent does not include Emacs version

Previous Next

Full log

View this message in rfc822 format

From: Ivan Shmakov <ivan <at> siamics.net>
To: 18812 <at> debbugs.gnu.org
Subject: bug#18812: 24.4; url.el user agent does not include Emacs version
Date: Fri, 24 Oct 2014 19:17:48 +0000

>>>>> "TZ" == Ted Zlatanov <tzz <at> lifelogs.com> writes:
>>>>> On Thu, 23 Oct 2014 21:19:21 -0400 Stefan Monnier wrote:

 PH> The url.el functions for making HTTP requests set the User-Agent
 PH> header to "URL/Emacs" which is not very descriptive.  It would be
 PH> helpful to include the Emacs version, at least.

 SM> Other people feel like it's already too descriptive.  Why should
 SM> all Emacs users constantly tell the whole world exactly which Emacs
 SM> version they're running?

 TZ> Web browsers do.  It helps web developers and is not particularly
 TZ> sensitive if only the major.minor version is used (without too much
 TZ> detail).

	Web browsers (and HTTP/HTTPS clients in general) also provide a
	way for the user to override the default User-Agent: value, and
	I believe that Emacs should allow for that, too.

	While somewhat tangential to the problem as reported, some of
	the issues I’ve seen with the default User-Agent: settings are:

	• Lynx’ default User-Agent: includes ‘libwww’ as a substring;
	  apparently, some Web sites mistake that for the well-known
	  Perl library of the same name, and thus reject the requests;
	  the solution is to either manually remove the respective
	  stanza, or to disable User-Agent: altogether;

	• other sites are known to disallow HTTP requests /without/
	  User-Agent:, however;

	• GNU Wget default identification is also not unknown to be
	  rejected; per my experience, using something like “tegW/1.15”
	  instead (via --user-agent=) tends to resolve the issue;

	• moreover, some sites actually serve different content
	  depending on User-Agent:, – apparently for “SEO” purposes; for
	  instance, a PDF file is served for a request from a party
	  identifying itself as a search engine, while regular browsers
	  get a Web page with a (presumably non-free) JavaScript-based
	  PDF reader set up for that file instead.

-- 
FSF associate member #7257  http://boycottsystemd.org/  … 3013 B6A0 230E 334A

Previous Next

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