mismatch between pyglet/font/win32.py and pyglet/font/win32query.py - examples/html_label.py tracebacks

[GoogleCodeExporter]

Seen on last pyglet from repo, rev 2014 01 03 2866:c43bc57c954f, but is present 
from before, see below.

What steps will reproduce the problem?
1. run examples/html_label.py
2. it tracebaks

Paste in the traceback or error message:
D:\hg_externals\pyglet_dev\examples>py -2.6 html_label.py
Traceback (most recent call last):
  File "html_label.py", line 62, in <module>
    multiline=True, anchor_y='center')
  File "D:\hg_externals\pyglet_dev\pyglet\text\__init__.py", line 492, in __init
__
    multiline, dpi, batch, group)
  File "D:\hg_externals\pyglet_dev\pyglet\text\__init__.py", line 257, in __init
__
    dpi=dpi, batch=batch, group=group)
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 791, in __init__

    self.document = document
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 874, in _set_doc
ument
    self._init_document()
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 977, in _init_do
cument
    self._update()
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 911, in _update
    lines = self._get_lines()
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 887, in _get_lin
es
    glyphs = self._get_glyphs()
  File "D:\hg_externals\pyglet_dev\pyglet\text\layout.py", line 1012, in _get_gl
yphs
    for start, end, (font, element) in runs.ranges(0, len(text)):
  File "D:\hg_externals\pyglet_dev\pyglet\text\runlist.py", line 404, in ranges
    starts[i], ends[i], values[i] = iterator.next()
  File "D:\hg_externals\pyglet_dev\pyglet\text\document.py", line 716, in ranges

    dpi=self.dpi)
  File "D:\hg_externals\pyglet_dev\pyglet\font\__init__.py", line 627, in load
    if _font_class.have_font(n):
  File "D:\hg_externals\pyglet_dev\pyglet\font\win32.py", line 267, in have_font

    return win32query.have_font(name, vector_only=True)
TypeError: have_font() got an unexpected keyword argument 'vector_only'

Tested in py2.6 and py3.3, win xp 32bits

The problem was introduced at r2817:eea2bb3a18b0 , the previous revision 
(2816:7c011471fd2b) works ok in py 2.6

Original issue reported on code.google.com by ccanepacc@gmail.com on 18 Feb 2014 at 8:16

[GoogleCodeExporter]

Original comment by useboxnet on 19 Feb 2014 at 9:06

[GoogleCodeExporter]

I am not sure that reea2bb3a18b0 is the source of problem. I see that offending 
line is introduced in rc698662310e6. Juan, where did you get it?

Original comment by techtonik@gmail.com on 20 Feb 2014 at 7:37

[GoogleCodeExporter]

techtonik: oh, that merge had some issues. After 
8d8d4656883005b1c60b6e9fc5729c07b4d756e3 it should be fine though.

Can you confirm that the issue can be reproduced in tip? ccanepacc report seems 
using that version.

Original comment by useboxnet on 20 Feb 2014 at 6:03

[GoogleCodeExporter]

Yes, look at: 
https://code.google.com/p/pyglet/source/detail?r=8d8d4656883005b1c60b6e9fc5729c0
7b4d756e3

The /pyglet/font/win32.py diff shows that vector_only=True and hav_font in 
/pyglet/font/win32query.py doesn't have that parameter; I'm sorry I can't help 
further on this.

Original comment by useboxnet on 20 Feb 2014 at 6:08

[GoogleCodeExporter]

If I understand correctly, the changesets related to this issue have the 
purpose of integrate https://bitbucket.org/techtonik/fontquery

Looking at that source, the extra kw-param vector_only=True in the call to 
have_font is spurious (the code for have_font both in pyglet and fontquery 
don't uses that param)

The other function provided by fontquery, font_query, does in fact use that 
kw-param.
Searching the pyglet repo tip for 'font_query', the only hit outside 
win32query.py is in a commented out line in examples/font_comparison.py :

#FONTS = list(set(pyglet.font.win32query.font_list(vector_only=True, 
monospace_only=True)))

May I suggest:
    - delete vector_only=True in the call to have_font (fixes this issue)
    - uncomment the call to font_list in examples, so that it gets exercised (it runs in py26 win32 xp)
?

Original comment by ccanepacc@gmail.com on 21 Feb 2014 at 4:59

[GoogleCodeExporter]

Also, maybe rename
font_comparison.py -> font_comparison_windows.py

Original comment by ccanepacc@gmail.com on 21 Feb 2014 at 5:02

[GoogleCodeExporter]

I am not sure where r8d8d46568830 merge comes from.

UPD: https://code.google.com/r/techtonik-fontquery/source/list

I confirm that it is my fault. There is a bug with loading monospace fonts with 
pyglet on Windows, so I to have to add the parameter, but failed to push into 
fontquery. To get back the previous behavior just remove vector_only=True

And if somebody can test the bug with monospace font and list it here - this 
would be awesome too.

Original comment by techtonik@gmail.com on 23 Feb 2014 at 6:56

[GoogleCodeExporter]

Original comment by useboxnet on 6 Mar 2014 at 9:44

• Changed state: Accepted

[GoogleCodeExporter]

Juan, do you think you can
    delete vector_only=True in the call to have_font (fixes the traceback)
    apply the patch in #704 ?

I think they are not controversial, and I plan to release cocos 0.6.0 in a few 
days; it would be good to suggest a 'known good' pyglet revision to use with 
cocos.

Original comment by ccanepacc@gmail.com on 19 Mar 2014 at 6:57

[GoogleCodeExporter]

ccanepacc: sure! This part was contributed by techtonik, that's why I tried to 
involve him; but meanwhile I can commit the workaround.

Original comment by useboxnet on 19 Mar 2014 at 8:17

[GoogleCodeExporter]

Done in https://code.google.com/p/pyglet/source/detail?r=82693a20ece0

Original comment by useboxnet on 19 Mar 2014 at 8:22

[GoogleCodeExporter]

thanks !

Original comment by ccanepacc@gmail.com on 20 Mar 2014 at 1:48

[GoogleCodeExporter]

A bit of investigation as asked by Anatoly in comment 7.

I see three different issues here

tabs looking bad (issue #727)
-----------------------------

This is an 'expectations not matching the actual pyglet spec' issue 

In windows, some fonts expected to be monospace render as proportional under 
pyglet
--------------------------------------------------------------------------------
---

This is a fontquery bug. The autoritative info to decide monospace / not 
monospace is FIXED_PITCH, not FF_MODERN
I would had filled an issue in fontquery with details and code but his issue 
tracker is not public.
Should I post details as a new issue in pyglet tracker ?

Expectations about supported font formats #731
----------------------------------------------

related to bad rendering of some fonts in examples/font_comparison.py

Original comment by ccanepacc@gmail.com on 7 Apr 2014 at 2:25

[GoogleCodeExporter]

opened issue #744 for the last bit remaining of this issue.
I think this issue can be closed now.

Original comment by ccanepacc@gmail.com on 21 May 2014 at 3:04

[GoogleCodeExporter]

Thanks Claudio!

Original comment by useboxnet on 21 May 2014 at 3:08

• Changed state: Fixed