Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Henry S. Thompson]

Something funny going on: the normal/bold contrast is much less than it
should be (used to be).  I'm on a new laptop with clean install of
Windows 10, Cygwin (64-bit), etc.

To see what I'm talking about, look at whatever test page you like which
includes both bold and non-bold text (I use gnus).  Compare either

  * the bold / non-bold contrast for Deja Vu Sans Mono vs. e.g. Lucida
     Sans Typewriter
  * the appearance of any text if you select DVSM Bold as the font,
     vs. the bold you get with emboldened text with DVSM Book as the font.

The setting of 'Show bold as font' has no effect.

Can anyone reproduce / suggest a fix?

ht

[The only easily-available cross-check is with an old install with
2.1.5, which shows good contrast (and slightly _different_ depending on
Show bold as font on or off).
-- 
       Henry S. Thompson, School of Informatics, University of Edinburgh
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                       URL: http://www.ltg.ed.ac.uk/~ht/
 [mail from me _always_ has a .sig like this -- mail without it is forged spam]

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Thomas Wolff]

Am 11.07.2016 um 20:52 schrieb Henry S. Thompson:
> Something funny going on: the normal/bold contrast is much less than it
> should be (used to be).  I'm on a new laptop with clean install of
> Windows 10, Cygwin (64-bit), etc.
>
> To see what I'm talking about, look at whatever test page you like which
> includes both bold and non-bold text (I use gnus).
What "gnus" test page?
>    Compare either
>
>    * the bold / non-bold contrast for Deja Vu Sans Mono vs. e.g. Lucida
>       Sans Typewriter
>    * the appearance of any text if you select DVSM Bold as the font,
>       vs. the bold you get with emboldened text with DVSM Book as the font.
>
> The setting of 'Show bold as font' has no effect.
It's using shadowed (self-drawn) bold mode, not the bold font, with 
DejaVu; with other fonts it's using the bold font.
> Can anyone reproduce / suggest a fix?
Looking for the root cause.
Thomas

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Warren Young]

On Jul 11, 2016, at 12:52 PM, Henry S. Thompson <ht@inf.ed.ac.uk> wrote:
> 
> To see what I'm talking about, look at whatever test page you like which
> includes both bold and non-bold text (I use gnus).

I used vttest:

  http://invisible-island.net/vttest/vttest.html

About halfway through menu option 2, you get this screen:

  https://imgur.com/AvZ0XwS

That’s with the latest version of Deja Vu Sans Mono on 64-bit Windows 10.  Looks fine to me.

Are you sure you have MinTTY configured to display bold with fonts, rather than colors?  Check Options > Text > Show bold as colour.  If that’s checked, it tries to fake bold by changing regular white (more like 75% gray) to bright white (100% white) and so forth.  I’m not sure what it does with displays that are already using the bright colors.



--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Warren Young]

On Jul 11, 2016, at 2:43 PM, Thomas Wolff <towo@towo.net> wrote:
> 
> Am 11.07.2016 um 20:52 schrieb Henry S. Thompson:
>> 
>> To see what I'm talking about, look at whatever test page you like which
>> includes both bold and non-bold text (I use gnus).
> What "gnus" test page?

gnus is a full-screen terminal newsreader add-on for Emacs:

  http://www.gnus.org/

Presumably it uses bold fonts.  (I wouldn’t know, I cannot Emacs.)

I think the vttest program I pointed to in my other response is more enlightening.

By the way, while going through menu item 2 of vttest to produce the screen shot I posted, I noticed that its notion of “last line” appears to be line 24.  Is it a weakness of MinTTY, vttest, or the VT-whatever protocol that it can’t tell that my MinTTY window is 60 lines high?

>> The setting of 'Show bold as font' has no effect.
> It's using shadowed (self-drawn) bold mode, not the bold font, with DejaVu; with other fonts it's using the bold font.

If you open the Fonts control panel, in the default Large Icons view, do you see all four variants of Deja Vu Sans Mono stacked together?

I ask because it is possible that the fonts weren’t constructed in a way that lets Windows understand that these four separate font files differ only in weight.  Since this font family doesn’t use the traditional Regular / Bold / Italic / Bold Italic scheme, it is possible that some software wouldn’t know how to move between weights and styles automatically.

There’s a lot of naive font handling code out in the world which only believes in those four variants.
--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Henry S. Thompson]

Thomas Wolff writes:

> What "gnus" test page?

Sorry, missed that question first time.  I first noticed the problem
when using gnus via ssh, not a very helpful test.

Much easier:

  > export PS1='\[\033[1m\]\h\[\033[0m\]<\!>:'

and your machine name should be bold.

But I have a further (bizarre) observation: This problem is
size and/or resolution-related:

       Show bold as font
       On            Off
Size

 8pt   Works         Doesn't work

12pt   Doesn't work  Doesn't work

The above is on a large screen (20" diagonal, resolution 1680 x 1050).

On a smaller screen (13" diagonal, resolution 1920 x 1080) I have to go
up to 16pt before bold and regular are indistinguishable...

ht
-- 
       Henry S. Thompson, School of Informatics, University of Edinburgh
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                       URL: http://www.ltg.ed.ac.uk/~ht/
 [mail from me _always_ has a .sig like this -- mail without it is forged spam]

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Henry S. Thompson]

Warren Young writes:

> If you open the Fonts control panel, in the default Large Icons view,
> do you see all four variants of Deja Vu Sans Mono stacked together?

Yes.

ht
-- 
       Henry S. Thompson, School of Informatics, University of Edinburgh
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                       URL: http://www.ltg.ed.ac.uk/~ht/
 [mail from me _always_ has a .sig like this -- mail without it is forged spam]

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Warren Young]

On Jul 12, 2016, at 9:09 AM, Henry S. Thompson <ht@inf.ed.ac.uk> wrote:
> 
> Much easier:
> 
>> export PS1='\[\033[1m\]\h\[\033[0m\]<\!>:'
> 
> and your machine name should be bold.

That is indeed what I see.

> But I have a further (bizarre) observation: This problem is
> size and/or resolution-related:
> 
>       Show bold as font
>       On            Off
> Size
> 
> 8pt   Works         Doesn't work
> 
> 12pt   Doesn't work  Doesn't work

The screenshot I posted was 12pt.  Bold and Book weights are clearly distinguished.

I will concede that there does seem to be some kind of size dependence, as the difference is less clear at 72pt, though I could not say that the difference “vanishes.”

It isn’t specific to Deja Vu, either, as the same thing happens with Lucida Console and Consolas.

That said, if you create a Wordpad document with the same faces, sizes, and weights, bold is more clearly distinguished at high point sizes.

So, perhaps MinTTY is indeed using some kind of faux bold here.
--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Brian Inglis]

Warren Young <wyml <at> etr-usa.com> writes:
> On Jul 11, 2016, at 2:43 PM, Thomas Wolff <towo <at> towo.net> wrote:
>> Am 11.07.2016 um 20:52 schrieb Henry S. Thompson:
>>> To see what I'm talking about, look at whatever test page you like which
>>> includes both bold and non-bold text (I use gnus).
> I think the vttest program I pointed to in my other response is more
enlightening.
> By the way, while going through menu item 2 of vttest to produce the
screen shot I posted, I noticed that its notion of “last line” appears to be
line 24. Is it a weakness of MinTTY, vttest, or the VT-whatever protocol
that it can’t tell that my MinTTY window is 60 lines high?

IIRC vttest is designed to test text emulations of real DEC/Cpq/HP VT###
terminals, which mainly offered only 24/25 lines but did support a
"condensed" font 132 column mode as well as 80 columns, and some quirky
features. 

man mintty may be a more canonical test of bold font handling; on my setup
(Win 7 x64, latest Cygwin curr), DejaVu Sans Mono/Book/9, with Show bold as
font selected, shows BOLD; with Show bold as font deselected, and Show bold
as colour selected, bold shows as grey; with Show bold as colour deselected,
bold shows only somewhat darker. 

>>> The setting of 'Show bold as font' has no effect.
>> It's using shadowed (self-drawn) bold mode, not the bold font, with
DejaVu; with other fonts it's using the bold font.
> If you open the Fonts control panel, in the default Large Icons view, do
you see all four variants of Deja Vu Sans Mono stacked together?

If you select mintty/Options/Text/Font/Select/Font/DejaVu Sans Mono shows
Font Styles Book, Oblique, Bold, Bold Oblique, so there should be no font or
handling problems with mintty. 

Font aliasing selection within mintty, Windows ClearType settings, and their
interactions, may cause more (suppression of?) variations in font display,
and be worth checking. 

Re: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Thomas Wolff]

Am 12.07.2016 um 17:27 schrieb Warren Young:
> ...
Windows reports the font metrics of the bold font sometimes slightly 
differently than the normal font.
Due to rounding, it depends on the font size, e.g. with DejaVu it 
happens with 12pt, 14pt, and larger, but not with 13pt, 11pt, and 
smaller. I do not know why that happens.
If mintty detects the bold fontsize to be different from the normal 
fontsize, it does not use the bold font. I could tweak that to allow 
some small deviation, perhaps.
> It isn’t specific to Deja Vu, either, as the same thing happens with Lucida Console and Consolas.
For Lucida Console, it's clear as there is no bold weight. Mintty could 
use the Windows-generated fake bold in that case but it rather generates 
its own.
> That said, if you create a Wordpad document with the same faces, sizes, and weights, bold is more clearly distinguished at high point sizes.
Mintty generates "faux-bold" by overstriking with a pixel offset of 1. 
Maybe it should scale the thickness with the font size (like it does now 
for manual underline and VT100 line drawing graphics), at increased risk 
of clipping, however.

Suggestions welcome.
Thomas

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Michael Enright]

On Tue, Jul 12, 2016 at 12:26 PM, Thomas Wolff wrote:
>
> Mintty generates "faux-bold" by overstriking with a pixel offset of 1. Maybe
> it should scale the thickness with the font size (like it does now for
> manual underline and VT100 line drawing graphics), at increased risk of
> clipping, however.
>
> Suggestions welcome.
> Thomas
>

I assume you have already tried quite a few suggestions. Did you try
digitally inflating the bitmap image of the normal weight characters?
Did you try using the metrics from the normal weight while rendering
with the bold weight?

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Thomas Wolff]

Am 12.07.2016 um 22:35 schrieb Michael Enright:
> On Tue, Jul 12, 2016 at 12:26 PM, Thomas Wolff wrote:
>> Mintty generates "faux-bold" by overstriking with a pixel offset of 1. Maybe
>> it should scale the thickness with the font size (like it does now for
>> manual underline and VT100 line drawing graphics), at increased risk of
>> clipping, however.
>>
>> Suggestions welcome.
>> Thomas
> I assume you have already tried quite a few suggestions. Did you try
> digitally inflating the bitmap image of the normal weight characters?
No; apart from how to find a suitable algorithm, this would bypass the
system font rendering, including smoothing/Cleartype, so likely not have
good results.

> Did you try using the metrics from the normal weight while rendering
> with the bold weight?
The bold font is requested to be created with the same metrics but turns
out to be created with different metrics.
There is no adjustment option during text output (which uses the Windows
function ExtTextOut), and "rendering" is done by Windows.

Thomas

---
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus


--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Thomas Wolff]

Am 13.07.2016 um 07:56 schrieb Thomas Wolff:
> Am 12.07.2016 um 22:35 schrieb Michael Enright:
>> On Tue, Jul 12, 2016 at 12:26 PM, Thomas Wolff wrote:
>>> Mintty generates "faux-bold" by overstriking with a pixel offset of 
>>> 1. Maybe
>>> it should scale the thickness with the font size (like it does now for
>>> manual underline and VT100 line drawing graphics), at increased risk of
>>> clipping, however.
>>>
>>> Suggestions welcome.
>>> Thomas
>> I assume you have already tried quite a few suggestions. Did you try
>> digitally inflating the bitmap image of the normal weight characters?
> No; apart from how to find a suitable algorithm, this would bypass the
> system font rendering, including smoothing/Cleartype, so likely not have
> good results.
>
>> Did you try using the metrics from the normal weight while rendering
>> with the bold weight?
> The bold font is requested to be created with the same metrics ...
Actually, it isn't. Found the root cause.
Thomas

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Thomas Wolff]

Am 13.07.2016 um 21:30 schrieb Thomas Wolff:
> Am 13.07.2016 um 07:56 schrieb Thomas Wolff:
>> Am 12.07.2016 um 22:35 schrieb Michael Enright:
>>> On Tue, Jul 12, 2016 at 12:26 PM, Thomas Wolff wrote:
>>>> Mintty generates "faux-bold" by overstriking with a pixel offset of 
>>>> 1. Maybe
>>>> it should scale the thickness with the font size (like it does now for
>>>> manual underline and VT100 line drawing graphics), at increased 
>>>> risk of
>>>> clipping, however.
>>>>
>>>> Suggestions welcome.
>>>> Thomas
>>> I assume you have already tried quite a few suggestions. Did you try
>>> digitally inflating the bitmap image of the normal weight characters?
>> No; apart from how to find a suitable algorithm, this would bypass the
>> system font rendering, including smoothing/Cleartype, so likely not have
>> good results.
>>
>>> Did you try using the metrics from the normal weight while rendering
>>> with the bold weight?
>> The bold font is requested to be created with the same metrics ...
> Actually, it isn't. Found the root cause.
Fixed in https://github.com/mintty/mintty/archive/master.zip
> Thomas

--
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: Mintty 2.4.0 and Deja Vu Sans Mono 2.35: issue with bold

[Henry S. Thompson]

Thomas Wolff writes:

> Fixed in https://github.com/mintty/mintty/archive/master.zip

Confirmed.

Thanks!

ht
-- 
       Henry S. Thompson, School of Informatics, University of Edinburgh
      10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
                Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
                       URL: http://www.ltg.ed.ac.uk/~ht/
 [mail from me _always_ has a .sig like this -- mail without it is forged spam]

--
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