Bug 26798 - Should keep a visible default cursor, even without -retro
Summary: Should keep a visible default cursor, even without -retro
Status: RESOLVED MOVED
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: git
Hardware: All All
: medium normal
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-02-28 07:45 UTC by Samuel Thibault
Modified: 2018-12-13 18:35 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Enable visible cursor on start without -retro (1.64 KB, patch)
2010-08-17 10:04 UTC, Samuel Thibault
no flags Details | Splinter Review
Enable visible cursor on start without -retro (1.79 KB, patch)
2010-08-19 03:43 UTC, Samuel Thibault
no flags Details | Splinter Review

Description Samuel Thibault 2010-02-28 07:45:01 UTC
Nowadays, X, without the -retro option, uses an invisible cursor.

This poses problems with a lot of users: when they are faced with a
completely dark screen without even a moving mouse, they think their
machine is completely hung, while it could just be that X clients can't
connect or are not starting for some reason.

Please put back a cursor, event if it is just a small square instead of
the old-style cross.
Comment 1 Michel Dänzer 2010-03-01 04:37:12 UTC
I always thought -retro should be the default. Distros / admins who can't handle the root weave and default cursor can disable them in the display manager configuration or wherever.
Comment 2 Samuel Thibault 2010-03-01 04:49:55 UTC
The two should maybe not be bound.

I can understand very well that showing a nice black screen is
better-looking for customers etc, so that part as a default doesn't
disturb me so much.  It is not a problem for me to have to give -retro
to get the root weave for the few testing purposes.

However, while e.g. Ubuntu would happily use an option to avoid the
-retro style, I believe it would be better for them to still have some
cursor (not necessarily the old-style cross).
Comment 3 James Cloos 2010-03-02 13:51:07 UTC
>>>>> bugzilla-daemon  <bugzilla-daemon@freedesktop.org> writes:

> I can understand very well that showing a nice black screen is
> better-looking for customers etc,

Then you are more imaginitive that I.  Black is far from a nice screen.

> I believe it would be better for them to still have some
> cursor (not necessarily the old-style cross).

An X, of course.  Maybe even a U+2717 ✗ or U+2718 ✘, but not a cross.

☺ ☺ ☺

Anyway one looks at it, -retro is the better user experience.
Comment 4 Samuel Thibault 2010-03-02 14:33:31 UTC
bugzilla-daemon@freedesktop.org, le Tue 02 Mar 2010 13:51:07 -0800, a écrit :
> --- Comment #3 from James Cloos <cloos@jhcloos.com>  2010-03-02 13:51:07 PST ---
> >>>>> bugzilla-daemon  <bugzilla-daemon@freedesktop.org> writes:
> 
> > I can understand very well that showing a nice black screen is
> > better-looking for customers etc,
> 
> Then you are more imaginitive that I.  Black is far from a nice screen.

From a "pretty" point of view, just ask any artist, between pure black
and some funky grey that looks awful on beamers that have numeric
trapezoid compensation, there's no chance.

> > I believe it would be better for them to still have some
> > cursor (not necessarily the old-style cross).
> 
> An X, of course.  Maybe even a U+2717 ✗ or U+2718 ✘, but not a cross.

Sorry that poor english skills don't allow to understand any difference.
Comment 5 James Cloos 2010-03-03 12:15:29 UTC
> From a "pretty" point of view, just ask any artist, between pure black
> and some funky grey that looks awful on beamers that have numeric
> trapezoid compensation, there's no chance.

I didn't say it *had* to be weave; just that black — which looks like
the display is off — cannot be a good choice.

Anyone who hates weave should choose another tiled image, or a nice
gradient, or a splash image.  Anything but black or white.

>> An X, of course.  Maybe even a U+2717 ✗ or U+2718 ✘, but not a cross.

> Sorry that poor english skills don't allow to understand any difference.

Sorry, it wasn't meant to sound like a flame.  But the cursor is an X, as
in The »X« Windows System.  Calling it a cross looses the link.

(And I did thrown in some ☺; I hope that they showed up as smiling faces!)
Comment 6 Samuel Thibault 2010-04-13 11:43:14 UTC
Once more, one of our users was looking for errors in Xorg.0.log just
because he was getting a black.  If he had at least a mouse cursor,
he would have understood that the X server was properly starting, but
clients weren't able to connect due to mere authentication issues.
Comment 7 Olaf Buddenhagen 2010-06-02 23:23:05 UTC
Proposing to show only the cursor sounds like an attempt at some kind of compromise... Detestable!

The thing which brought this whole mess about, is that apparently (that is, according to hearsay), some people feel deeply offended by seeing a flash of either the root weave *or* the default cursor during server startup, before the display manager is loaded. As a result, certain distributions (notably a rather popular one starting with 'F') have had them disabled through the display manager configuration for some time. On one memorable day, a certain developer from said distribution went ahead and decided that as "everyone" (which apparently excludes another popular distribution starting with 'D') sets this in the display manager configuration anyways, this means it has to be made default, so nobody will be exposed ever again to the stabbing pain of seeing the root weave -- including those who never felt tempted to change the setting before...

Now obviously this is a model example of flawed logic: if everyone overrides it anyways, this means the default only ever shows when someone starts a bare X server for testing purposes -- which is indeed precisely the case where it is *useful* to have something visible.

But then, who said defaults should be governed by what is useful? Statistics can't be wrong after all -- especially those with one data point and arbitrary interpretation...
Comment 8 Samuel Thibault 2010-07-01 04:15:44 UTC
Yet one more user got a black screen without any idea why it's that way
(driver issue, stuck machine, clients not able to connect?)

Please really enable at least a mouse cursor.
Comment 9 Samuel Thibault 2010-07-28 03:19:06 UTC
Still debugging pain with yet another user, with no cursor without -retro...
Comment 10 Samuel Thibault 2010-07-31 16:22:03 UTC
Yet another user is thinking blank == X server got wrong, and not
looking further than server log messages, taking warnings there as
fatal errors...
Comment 11 Adam Jackson 2010-08-09 13:42:33 UTC
A patch to set the initial root cursor would certainly be considered.
Comment 12 Samuel Thibault 2010-08-17 10:04:45 UTC
Created attachment 37922 [details] [review]
Enable visible cursor on start without -retro

Here it is.
Comment 13 Samuel Thibault 2010-08-19 03:43:35 UTC
Created attachment 37972 [details] [review]
Enable visible cursor on start without -retro

I forgot to upgrade my checkout before diffing, here is an updated
patch.
Comment 14 Samuel Thibault 2010-09-29 14:47:51 UTC
Yet another got it wrong.  Can this patch be applied please?
Comment 15 Samuel Thibault 2010-10-17 12:40:15 UTC
Yet another user erroneously thought that completely black screen means
X server not working at all, and didn't investigate the client part at
all.

Can you please apply the proposed patch?
Comment 16 Alan Coopersmith 2010-10-18 20:48:49 UTC
(In reply to comment #15)
> Can you please apply the proposed patch?

No.   Only the X server maintainer can apply patches to the git master repo,
and he requires them to be posted for review first on the xorg-devel mailing
list.   See http://www.x.org/wiki/XServer
Comment 17 Samuel Thibault 2010-10-19 16:24:07 UTC
Ah, at last some news...  Done so.
Comment 18 Samuel Thibault 2014-09-12 16:23:02 UTC
Just for the record, the submission happened on:

http://lists.x.org/archives/xorg-devel/2010-October/014291.html
Comment 19 GitLab Migration User 2018-12-13 18:35:17 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/xserver/issues/252.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.