[bug] character width calculate error with unicode

[gsw945]

Description

Chinese characters lead to line breaks or print repeatedly, because of the characters' width calculated error.

Code

1. use FormatCustomText:

   import progressbar
   import time
   
   
   total = 12
   current = 0
   format_custom_text = progressbar.FormatCustomText(
       '%(progress)6s: ',
       dict(
           progress='进度'
       ),
   )
   bar = progressbar.ProgressBar(
       widgets=[
           format_custom_text,
           progressbar.Bar(),
           ' ',
           progressbar.Counter(),
           ' / {0}'.format(total)
       ],
       max_value=total,
       redirect_stdout=True
   )
   bar.start()
   bar.update(current)
   while current < total:
       current += 1
       bar.update(current)
       time.sleep(0.5)
   bar.finish()

   output:
   

2. use raw text:

   import progressbar
   import time
   
   
   total = 12
   current = 0
   tip = '进度: '
   bar = progressbar.ProgressBar(
       widgets=[
           tip,
           progressbar.Bar(),
           ' ',
           progressbar.Counter(),
           ' / {0}'.format(total)
       ],
       max_value=total,
       redirect_stdout=True
   )
   bar.start()
   bar.update(current)
   while current < total:
       current += 1
       bar.update(current)
       time.sleep(0.5)
   bar.finish()

   output:
   

3. fill length of str with 6(a chinese character's width is 2) by using \0

   import progressbar
   import time
   
   
   total = 12
   current = 0
   tip = '进\0度\0: '
   bar = progressbar.ProgressBar(
       widgets=[
           tip,
           progressbar.Bar(),
           ' ',
           progressbar.Counter(),
           ' / {0}'.format(total)
       ],
       max_value=total,
       redirect_stdout=True
   )
   bar.start()
   bar.update(current)
   while current < total:
       current += 1
       bar.update(current)
       time.sleep(0.5)
   bar.finish()

   output:
   

ps: 3 is the effect i expected.

Versions

• Python version: 3.5.2
• Python distribution/environment: CPython
• Operating System: Ubuntu 16.04 LTS
• Package version: 3.33.2

maybe helpful

wcwidth Python library that measures the width of unicode strings rendered to a terminal

[wolph]

Thanks for the very detailed bug report, it's greatly appreciated :)

I'll add support for a pluggable len() function so libraries such as wcwidth can be used. Since the width greatly depends on the font as well I imagine that even libraries such as wcwidth don't work correctly in all cases.

[wolph]

I've added support for a custom_len function to the progressbar. This makes it possible to specify your own custom length function, it seems the wcwidth library doesn't support these characters either so I've added an example to the readme to show at least one workaround.

# vim: fileencoding=utf-8
import time
import progressbar


def custom_len(value):
    # These characters take up more space
    characters = {
        '进': 2,
        '度': 2,
    }

    total = 0
    for c in value:
        total += characters.get(c, 1)

    return total


bar = progressbar.ProgressBar(
    widgets=[
        '进度: ',
        progressbar.Bar(),
        ' ',
        progressbar.Counter(format='%(value)02d/%(max_value)d'),
    ],
    len_func=custom_len,
)
for i in bar(range(10)):
    time.sleep(0.1)