Latest Cygwin update and Emacs in Mintty

Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

After the latest Cygwin update I'm hitting an interesting problem with
emacs-nox running in a mintty: when Emacs starts, it decides that the
background color is gray instead of the usual white (for all but the
rightmost character in the status line, interestingly enough).  I have
the normal mintty background set slightly off-white (to #F8F8F8) and
mintty reports itself as xterm-256color (the same happens in when TERM
is set to screen-256color).  Emacs starts up with the correct background
color, draws the status bar and the menu bar and only then switches the
background to gray.  The gray it choses is very slightly lighter than
the status bar (so lightly in fact that I can make out the difference on
only one of my three monitors).  Emacs really thinks it is using a white
background, as evidenced by the fact that (set-background-color "white")
will produce exactly the same result.  I get my usual background back
with (set-background-color "#F8F8F8"), but I have no idea where the
wrong setting for the named color "white" comes from.

I'm just trying this at home remotely logging in from a konsole terminal
on my Linux box: Here the background chosen is slightly darker than the
status bar, the status bar seems to be #B8B8B8 and the background for
"white seems to be #B4B4B4.  Curious and curiouser...


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf microQ V2.22R2:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Achim Gratz writes:
> After the latest Cygwin update I'm hitting an interesting problem with
> emacs-nox running in a mintty: when Emacs starts, it decides that the
> background color is gray instead of the usual white (for all but the
> rightmost character in the status line, interestingly enough).  I have
> the normal mintty background set slightly off-white (to #F8F8F8) and
> mintty reports itself as xterm-256color (the same happens in when TERM
> is set to screen-256color).  Emacs starts up with the correct background
> color, draws the status bar and the menu bar and only then switches the
> background to gray.  The gray it choses is very slightly lighter than
> the status bar (so lightly in fact that I can make out the difference on
> only one of my three monitors).  Emacs really thinks it is using a white
> background, as evidenced by the fact that (set-background-color "white")
> will produce exactly the same result.  I get my usual background back
> with (set-background-color "#F8F8F8"), but I have no idea where the
> wrong setting for the named color "white" comes from.
>
> I'm just trying this at home remotely logging in from a konsole terminal
> on my Linux box: Here the background chosen is slightly darker than the
> status bar, the status bar seems to be #B8B8B8 and the background for
> "white seems to be #B4B4B4.  Curious and curiouser...

Ken, did you have a chance to look into this?


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf rackAttack:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/14/2017 3:00 PM, Achim Gratz wrote:
> Achim Gratz writes:
>> After the latest Cygwin update I'm hitting an interesting problem with
>> emacs-nox running in a mintty: when Emacs starts, it decides that the
>> background color is gray instead of the usual white (for all but the
>> rightmost character in the status line, interestingly enough).  I have
>> the normal mintty background set slightly off-white (to #F8F8F8) and
>> mintty reports itself as xterm-256color (the same happens in when TERM
>> is set to screen-256color).  Emacs starts up with the correct background
>> color, draws the status bar and the menu bar and only then switches the
>> background to gray.  The gray it choses is very slightly lighter than
>> the status bar (so lightly in fact that I can make out the difference on
>> only one of my three monitors).  Emacs really thinks it is using a white
>> background, as evidenced by the fact that (set-background-color "white")
>> will produce exactly the same result.  I get my usual background back
>> with (set-background-color "#F8F8F8"), but I have no idea where the
>> wrong setting for the named color "white" comes from.
>>
>> I'm just trying this at home remotely logging in from a konsole terminal
>> on my Linux box: Here the background chosen is slightly darker than the
>> status bar, the status bar seems to be #B8B8B8 and the background for
>> "white seems to be #B4B4B4.  Curious and curiouser...
> 
> Ken, did you have a chance to look into this?

Sorry, I must have missed this when you first sent it.  I'll take a look.

Ken


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/15/2017 6:47 AM, Ken Brown wrote:
> On 6/14/2017 3:00 PM, Achim Gratz wrote:
>> Achim Gratz writes:
>>> After the latest Cygwin update I'm hitting an interesting problem with
>>> emacs-nox running in a mintty: when Emacs starts, it decides that the
>>> background color is gray instead of the usual white (for all but the
>>> rightmost character in the status line, interestingly enough).  I have
>>> the normal mintty background set slightly off-white (to #F8F8F8) and
>>> mintty reports itself as xterm-256color (the same happens in when TERM
>>> is set to screen-256color).  Emacs starts up with the correct background
>>> color, draws the status bar and the menu bar and only then switches the
>>> background to gray.  The gray it choses is very slightly lighter than
>>> the status bar (so lightly in fact that I can make out the difference on
>>> only one of my three monitors).  Emacs really thinks it is using a white
>>> background, as evidenced by the fact that (set-background-color "white")
>>> will produce exactly the same result.  I get my usual background back
>>> with (set-background-color "#F8F8F8"), but I have no idea where the
>>> wrong setting for the named color "white" comes from.
>>>
>>> I'm just trying this at home remotely logging in from a konsole terminal
>>> on my Linux box: Here the background chosen is slightly darker than the
>>> status bar, the status bar seems to be #B8B8B8 and the background for
>>> "white seems to be #B4B4B4.  Curious and curiouser...
>>
>> Ken, did you have a chance to look into this?
> 
> Sorry, I must have missed this when you first sent it.  I'll take a look.

I also see some strange behavior, but I can't reproduce what you're 
seeing.  Here's what I tried:

1. I edited .minttyrc to set BackgroundColour=#F8F8F8. 
(ForegroundColour is 0,0,0, which I guess is the default.)

2. I ran 'emacs-nox -Q'.  The background color didn't visibly change; 
it's still off-white.  The mode line and menu line appear to be reversed 
(black background, off-white foreground, though I can't really be sure 
that the foreground is off-white rather than white).

3. I evaluated (set-background-color "white").  This caused the 
background to turn gray.  The mode line and menu bar didn't visibly change.

4. I evaluated (set-background-color "#F8F8F8").  The background stayed 
gray.

I need to investigate 3 and 4, but first I'd like to understand why you 
and I are seeing different things.  One difference is the value of TERM. 
  You say "mintty reports itself as xterm-256color".  Do you mean it 
does this by default, or do you have something in your startup files 
that causes this?  When I start mintty, I get the following:

$ echo $TERM
xterm

Ken

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Ken Brown writes:
> 1. I edited .minttyrc to set
> BackgroundColour=#F8F8F8. (ForegroundColour is 0,0,0, which I guess is
> the default.)

OK.

> 2. I ran 'emacs-nox -Q'.  The background color didn't visibly change;
> it's still off-white.  The mode line and menu line appear to be
> reversed (black background, off-white foreground, though I can't
> really be sure that the foreground is off-white rather than white).

I think the background gets set to "white" by my .emacs (or rather the
part that customize-options saved).

> 3. I evaluated (set-background-color "white").  This caused the
> background to turn gray.  The mode line and menu bar didn't visibly
> change.

That's the main problem and I can't figure out why it associates white with
anything that doesn't have all the bits set in the RGB colorspec.

> 4. I evaluated (set-background-color "#F8F8F8").  The background
> stayed gray.

That's most likely a result of your terminal not being recognized as
being able to use 256 colors.

> I need to investigate 3 and 4, but first I'd like to understand why
> you and I are seeing different things.  One difference is the value of
> TERM. You say "mintty reports itself as xterm-256color".  Do you mean
> it does this by default, or do you have something in your startup
> files that causes this?  When I start mintty, I get the following:

That's an option you can set in mintty or just change the TERM variable
before you start emacs.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf rackAttack:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/15/2017 2:43 PM, Achim Gratz wrote:
> Ken Brown writes:
>> 3. I evaluated (set-background-color "white").  This caused the
>> background to turn gray.  The mode line and menu bar didn't visibly
>> change.
> 
> That's the main problem and I can't figure out why it associates white with
> anything that doesn't have all the bits set in the RGB colorspec.

OK, let's focus on this problem.

Your original report, written on February 22, said that the problem 
started after the latest Cygwin update.  That means you would have 
installed emacs-25.1, if Emacs was updated at all.  But I've just 
checked that the problem already exists in emacs-24.5.  So maybe the 
problem was triggered by an update of some package other than Emacs, 
Mintty being the most obvious guess.  Can you check your 
/var/log/setup.log to see what was updated around February 22?

Ken


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Ken Brown writes:
> Your original report, written on February 22, said that the problem
> started after the latest Cygwin update.  That means you would have
> installed emacs-25.1, if Emacs was updated at all.

No, it really was a Cygwin update (to 2.7.0-1).  The Emacs update was
before that and it had worked right up to that update.

> But I've just checked that the problem already exists in emacs-24.5.
> So maybe the problem was triggered by an update of some package other
> than Emacs, Mintty being the most obvious guess.  Can you check your
> /var/log/setup.log to see what was updated around February 22?

The only other two things that were updated that day (along with some
perl distributions that can't possibly be related, if only because I've
locally updated them before I got them back from the offical Cygwin
packages) was coreutils and vim.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/16/2017 2:27 PM, Achim Gratz wrote:
> Ken Brown writes:
>> Your original report, written on February 22, said that the problem
>> started after the latest Cygwin update.  That means you would have
>> installed emacs-25.1, if Emacs was updated at all.
> 
> No, it really was a Cygwin update (to 2.7.0-1).  The Emacs update was
> before that and it had worked right up to that update.

Sorry, I misunderstood what you meant by "Cygwin update".  I understand 
now.  I'll try to do a bisection of the Cygwin sources.  It may be a few 
days before I get to this.

Ken


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Ken Brown writes:
>> No, it really was a Cygwin update (to 2.7.0-1).  The Emacs update was
>> before that and it had worked right up to that update.
>
> Sorry, I misunderstood what you meant by "Cygwin update".  I
> understand now.  I'll try to do a bisection of the Cygwin sources.  It
> may be a few days before I get to this.

It may be possible that the coreutils update messed up things in a way
that I don't understand, but that's unlikely.  Or vim might have
delivered a file that conflicts with something from Emacs, but Emacs has
since been updated again and so that conflict would have flipped in
favor of Emacs at those times, which I didn't observe (and no such
conflicts seem to exist in the package file listings at the current
moment).


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Terratec KOMPLEXER:
http://Synth.Stromeko.net/Downloads.html#KomplexerWaves

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/17/2017 4:20 AM, Achim Gratz wrote:
> Ken Brown writes:
>>> No, it really was a Cygwin update (to 2.7.0-1).  The Emacs update was
>>> before that and it had worked right up to that update.
>>
>> Sorry, I misunderstood what you meant by "Cygwin update".  I
>> understand now.  I'll try to do a bisection of the Cygwin sources.  It
>> may be a few days before I get to this.
> 
> It may be possible that the coreutils update messed up things in a way
> that I don't understand, but that's unlikely.  Or vim might have
> delivered a file that conflicts with something from Emacs, but Emacs has
> since been updated again and so that conflict would have flipped in
> favor of Emacs at those times, which I didn't observe (and no such
> conflicts seem to exist in the package file listings at the current
> moment).

I just tried downgrading cygwin1.dll to the 2016-12-14 snapshot, emacs 
to 25.1-1 (released 2016-09-17), and coreutils to 8.26-1 (released 
2016-12-16).  The problem is still there with that configuration.  And 
starting with an off-white background is not necessary.  Even with no 
.minttrc, 'emacs-nox -Q' followed by 'M-x set-background-color RET white 
RET' turns the background gray.  I can also reproduce the problem in 
xterm, so I don't think Mintty is the culprit.

What about ncurses?  That was updated on 2017-01-27?  I haven't tried 
downgrading it yet, but I can probably try it later today.

Ken

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Ken Brown writes:
> Even with no .minttrc, 'emacs-nox -Q' followed by 'M-x
> set-background-color RET white RET' turns the background gray.

I can't rule out that there was a change to the configuration of Emacs
on that day that effectively added the set-background operation on that
day.

> I can also reproduce the
> problem in xterm, so I don't think Mintty is the culprit.

I already said in my original report that I can reproduce this from my
Linux box via konsole.

I don't know why I didn't think of that earlier, but I just fired up
emacs-nox from my Linux box and did 'M-x set-background-color RET white
RET' and guess what, the background turns a relatively dark shade of
grey.  I've tried to dig through the call chain that would lead to that
result and it seems that the result depends on whether brightwhite/15 or
white/7 gets picked from tty-color-alist or a value gets generated from
tty-color-standard-values (which uses the X definitions for the named
colors).


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Factory and User Sound Singles for Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Ken Brown]

On 6/17/2017 10:38 AM, Achim Gratz wrote:
> I don't know why I didn't think of that earlier, but I just fired up
> emacs-nox from my Linux box and did 'M-x set-background-color RET white
> RET' and guess what, the background turns a relatively dark shade of
> grey.  I've tried to dig through the call chain that would lead to that
> result and it seems that the result depends on whether brightwhite/15 or
> white/7 gets picked from tty-color-alist or a value gets generated from
> tty-color-standard-values (which uses the X definitions for the named
> colors).

OK, so it's not a Cygwin issue.  And I just learned about M-x 
list-colors-display.  If you try this in emacs-nox in Mintty with 
TERM=xterm-256color, you'll see that white is #e5e5e5 and brightwhite is 
#ffffff.  So I guess there's no bug here, just a surprising definition 
of "white".

Ken


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Ken Brown writes:
> OK, so it's not a Cygwin issue.  And I just learned about M-x 
> list-colors-display.  If you try this in emacs-nox in Mintty with
> TERM=xterm-256color, you'll see that white is #e5e5e5 and brightwhite
> is #ffffff.  So I guess there's no bug here, just a surprising
> definition of "white".

Yes.  I'll have to figure out just why it tries to set the background to
the named color "white" on startup, then.  That must have got saved as
an option somewhere, maybe I can still find an old enough backup to
compare the startup files to.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Terratec KOMPLEXER:
http://Synth.Stromeko.net/Downloads.html#KomplexerWaves

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Achim Gratz]

Achim Gratz writes:
> Yes.  I'll have to figure out just why it tries to set the background to
> the named color "white" on startup, then.  That must have got saved as
> an option somewhere, maybe I can still find an old enough backup to
> compare the startup files to.

Indeed on the same day that the update was done, a setting has changed
to explicitly requesting "white" as a backgroundcolor instead of leaving
the default.  I've changed that manually to "f8f8f8" in the settings
file and all is well now.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptations for KORG EX-800 and Poly-800MkII V0.9:
http://Synth.Stromeko.net/Downloads.html#KorgSDada

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

Re: Latest Cygwin update and Emacs in Mintty

[Brian Inglis]

On 2017-06-15 07:35, Ken Brown wrote:
> On 6/15/2017 6:47 AM, Ken Brown wrote:
>> On 6/14/2017 3:00 PM, Achim Gratz wrote:
>>> Achim Gratz writes:
>>>> After the latest Cygwin update I'm hitting an interesting problem with
>>>> emacs-nox running in a mintty: when Emacs starts, it decides that the
>>>> background color is gray instead of the usual white (for all but the
>>>> rightmost character in the status line, interestingly enough).  I have
>>>> the normal mintty background set slightly off-white (to #F8F8F8) and
>>>> mintty reports itself as xterm-256color (the same happens in when TERM
>>>> is set to screen-256color).  Emacs starts up with the correct
>>>> background
>>>> color, draws the status bar and the menu bar and only then switches the
>>>> background to gray.  The gray it choses is very slightly lighter than
>>>> the status bar (so lightly in fact that I can make out the
>>>> difference on
>>>> only one of my three monitors).  Emacs really thinks it is using a
>>>> white
>>>> background, as evidenced by the fact that (set-background-color
>>>> "white")
>>>> will produce exactly the same result.  I get my usual background back
>>>> with (set-background-color "#F8F8F8"), but I have no idea where the
>>>> wrong setting for the named color "white" comes from.
>>>>
>>>> I'm just trying this at home remotely logging in from a konsole
>>>> terminal
>>>> on my Linux box: Here the background chosen is slightly darker than the
>>>> status bar, the status bar seems to be #B8B8B8 and the background for
>>>> "white seems to be #B4B4B4.  Curious and curiouser...
>>>
>>> Ken, did you have a chance to look into this?
>>
>> Sorry, I must have missed this when you first sent it.  I'll take a look.
> 
> I also see some strange behavior, but I can't reproduce what you're
> seeing.  Here's what I tried:
> 
> 1. I edited .minttyrc to set BackgroundColour=#F8F8F8. (ForegroundColour
> is 0,0,0, which I guess is the default.)
> 
> 2. I ran 'emacs-nox -Q'.  The background color didn't visibly change;
> it's still off-white.  The mode line and menu line appear to be reversed
> (black background, off-white foreground, though I can't really be sure
> that the foreground is off-white rather than white).
> 
> 3. I evaluated (set-background-color "white").  This caused the
> background to turn gray.  The mode line and menu bar didn't visibly change.
> 
> 4. I evaluated (set-background-color "#F8F8F8").  The background stayed
> gray.
> 
> I need to investigate 3 and 4, but first I'd like to understand why you
> and I are seeing different things.  One difference is the value of TERM.
>  You say "mintty reports itself as xterm-256color".  Do you mean it does
> this by default, or do you have something in your startup files that
> causes this?  When I start mintty, I get the following:
> 
> $ echo $TERM
> xterm

Visible and selectable in mintty, Options, Terminal, Type or
{/etc/,~/.}minttyrc Term=
mintty has rgb.txt compiled in as rgb.t

emacs package installs /usr/share/emacs/25.2/etc/rgb.txt - and that has
the expected contents - is it found and loaded from there?

IANAEmacs guy

-- 
Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple