Don't update width of prompt area

Fixes gh-9331 (roughly)

The input prompts are now displayed through prompt_toolkit, so this code
doesn't know the width it's trying to match.

IPython/core/prompts.py

@@ -1,412 +1,412
 # -*- coding: utf-8 -*-
 """Classes for handling input/output prompts."""
 
 # Copyright (c) 2001-2007 Fernando Perez <[e-mail]>
 # Copyright (c) IPython Development Team.
 # Distributed under the terms of the Modified BSD License.
 
 import os
 import re
 import socket
 import sys
 import time
 
 from string import Formatter
 
 from traitlets.config.configurable import Configurable
 from IPython.core import release
 from IPython.utils import coloransi, py3compat
 from traitlets import Unicode, Instance, Dict, Bool, Int, observe, default
 
 from IPython.utils.PyColorize import LightBGColors, LinuxColors, NoColor
 
 #-----------------------------------------------------------------------------
 # Color schemes for prompts
 #-----------------------------------------------------------------------------
 
 InputColors = coloransi.InputTermColors  # just a shorthand
 Colors = coloransi.TermColors  # just a shorthand
 
 color_lists = dict(normal=Colors(), inp=InputColors(), nocolor=coloransi.NoColors())
 
 #-----------------------------------------------------------------------------
 # Utilities
 #-----------------------------------------------------------------------------
 
 class LazyEvaluate(object):
     """This is used for formatting strings with values that need to be updated
     at that time, such as the current time or working directory."""
     def __init__(self, func, *args, **kwargs):
         self.func = func
         self.args = args
         self.kwargs = kwargs
 
     def __call__(self, **kwargs):
         self.kwargs.update(kwargs)
         return self.func(*self.args, **self.kwargs)
 
     def __str__(self):
         return str(self())
 
     def __unicode__(self):
         return py3compat.unicode_type(self())
 
     def __format__(self, format_spec):
         return format(self(), format_spec)
 
 def multiple_replace(dict, text):
     """ Replace in 'text' all occurrences of any key in the given
     dictionary by its corresponding value.  Returns the new string."""
 
     # Function by Xavier Defrang, originally found at:
     # http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/81330
 
     # Create a regular expression  from the dictionary keys
     regex = re.compile("(%s)" % "|".join(map(re.escape, dict.keys())))
     # For each match, look-up corresponding value in dictionary
     return regex.sub(lambda mo: dict[mo.string[mo.start():mo.end()]], text)
 
 #-----------------------------------------------------------------------------
 # Special characters that can be used in prompt templates, mainly bash-like
 #-----------------------------------------------------------------------------
 
 # If $HOME isn't defined (Windows), make it an absurd string so that it can
 # never be expanded out into '~'.  Basically anything which can never be a
 # reasonable directory name will do, we just want the $HOME -> '~' operation
 # to become a no-op.  We pre-compute $HOME here so it's not done on every
 # prompt call.
 
 # FIXME:
 
 # - This should be turned into a class which does proper namespace management,
 # since the prompt specials need to be evaluated in a certain namespace.
 # Currently it's just globals, which need to be managed manually by code
 # below.
 
 # - I also need to split up the color schemes from the prompt specials
 # somehow.  I don't have a clean design for that quite yet.
 
 HOME = py3compat.str_to_unicode(os.environ.get("HOME","//////:::::ZZZZZ,,,~~~"))
 
 # This is needed on FreeBSD, and maybe other systems which symlink /home to
 # /usr/home, but retain the $HOME variable as pointing to /home
 HOME = os.path.realpath(HOME)
 
 # We precompute a few more strings here for the prompt_specials, which are
 # fixed once ipython starts.  This reduces the runtime overhead of computing
 # prompt strings.
 USER           = py3compat.str_to_unicode(os.environ.get("USER",''))
 HOSTNAME       = py3compat.str_to_unicode(socket.gethostname())
 HOSTNAME_SHORT = HOSTNAME.split(".")[0]
 
 # IronPython doesn't currently have os.getuid() even if
 # os.name == 'posix'; 2/8/2014
 ROOT_SYMBOL    = "#" if (os.name=='nt' or sys.platform=='cli' or os.getuid()==0) else "$"
 
 prompt_abbreviations = {
     # Prompt/history count
     '%n' : '{color.number}' '{count}' '{color.prompt}',
     r'\#': '{color.number}' '{count}' '{color.prompt}',
     # Just the prompt counter number, WITHOUT any coloring wrappers, so users
     # can get numbers displayed in whatever color they want.
     r'\N': '{count}',
 
     # Prompt/history count, with the actual digits replaced by dots or
     # spaces.  Used mainly in continuation prompts (prompt_in2).
     r'\D': '{dots}',
     r'\S': '{spaces}',
 
     # Current time
     r'\T' : '{time}',
     # Current working directory
     r'\w': '{cwd}',
     # Basename of current working directory.
     # (use os.sep to make this portable across OSes)
     r'\W' : '{cwd_last}',
     # These X<N> are an extension to the normal bash prompts.  They return
     # N terms of the path, after replacing $HOME with '~'
     r'\X0': '{cwd_x[0]}',
     r'\X1': '{cwd_x[1]}',
     r'\X2': '{cwd_x[2]}',
     r'\X3': '{cwd_x[3]}',
     r'\X4': '{cwd_x[4]}',
     r'\X5': '{cwd_x[5]}',
     # Y<N> are similar to X<N>, but they show '~' if it's the directory
     # N+1 in the list.  Somewhat like %cN in tcsh.
     r'\Y0': '{cwd_y[0]}',
     r'\Y1': '{cwd_y[1]}',
     r'\Y2': '{cwd_y[2]}',
     r'\Y3': '{cwd_y[3]}',
     r'\Y4': '{cwd_y[4]}',
     r'\Y5': '{cwd_y[5]}',
     # Hostname up to first .
     r'\h': HOSTNAME_SHORT,
     # Full hostname
     r'\H': HOSTNAME,
     # Username of current user
     r'\u': USER,
     # Escaped '\'
     '\\\\': '\\',
     # Newline
     r'\n': '\n',
     # Carriage return
     r'\r': '\r',
     # Release version
     r'\v': release.version,
     # Root symbol ($ or #)
     r'\$': ROOT_SYMBOL,
     }
 
 #-----------------------------------------------------------------------------
 # More utilities
 #-----------------------------------------------------------------------------
 
 def cwd_filt(depth):
     """Return the last depth elements of the current working directory.
 
     $HOME is always replaced with '~'.
     If depth==0, the full path is returned."""
 
     cwd = py3compat.getcwd().replace(HOME,"~")
     out = os.sep.join(cwd.split(os.sep)[-depth:])
     return out or os.sep
 
 def cwd_filt2(depth):
     """Return the last depth elements of the current working directory.
 
     $HOME is always replaced with '~'.
     If depth==0, the full path is returned."""
 
     full_cwd = py3compat.getcwd()
     cwd = full_cwd.replace(HOME,"~").split(os.sep)
     if '~' in cwd and len(cwd) == depth+1:
         depth += 1
     drivepart = ''
     if sys.platform == 'win32' and len(cwd) > depth:
         drivepart = os.path.splitdrive(full_cwd)[0]
     out = drivepart + '/'.join(cwd[-depth:])
 
     return out or os.sep
 
 #-----------------------------------------------------------------------------
 # Prompt classes
 #-----------------------------------------------------------------------------
 
 lazily_evaluate = {'time': LazyEvaluate(time.strftime, "%H:%M:%S"),
                    'cwd': LazyEvaluate(py3compat.getcwd),
                    'cwd_last': LazyEvaluate(lambda: py3compat.getcwd().split(os.sep)[-1]),
                    'cwd_x': [LazyEvaluate(lambda: py3compat.getcwd().replace(HOME,"~"))] +\
                             [LazyEvaluate(cwd_filt, x) for x in range(1,6)],
                    'cwd_y': [LazyEvaluate(cwd_filt2, x) for x in range(6)]
                    }
 
 def _lenlastline(s):
     """Get the length of the last line. More intelligent than
     len(s.splitlines()[-1]).
     """
     if not s or s.endswith(('\n', '\r')):
         return 0
     return len(s.splitlines()[-1])
 
 
 invisible_chars_re = re.compile('\001[^\001\002]*\002')
 def _invisible_characters(s):
     """
     Get the number of invisible ANSI characters in s. Invisible characters
     must be delimited by \001 and \002.
     """
     return _lenlastline(s) - _lenlastline(invisible_chars_re.sub('', s))
 
 class UserNSFormatter(Formatter):
     """A Formatter that falls back on a shell's user_ns and __builtins__ for name resolution"""
     def __init__(self, shell):
         self.shell = shell
 
     def get_value(self, key, args, kwargs):
         # try regular formatting first:
         try:
             return Formatter.get_value(self, key, args, kwargs)
         except Exception:
             pass
         # next, look in user_ns and builtins:
         for container in (self.shell.user_ns, __builtins__):
             if key in container:
                 return container[key]
         # nothing found, put error message in its place
         return "<ERROR: '%s' not found>" % key
 
 
 class PromptManager(Configurable):
     """This is the primary interface for producing IPython's prompts."""
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC', allow_none=True)
 
     color_scheme_table = Instance(coloransi.ColorSchemeTable, allow_none=True)
     color_scheme = Unicode('Linux').tag(config=True)
 
     @observe('color_scheme')
     def _color_scheme_changed(self, change):
         self.color_scheme_table.set_active_scheme(change['new'])
         for pname in ['in', 'in2', 'out', 'rewrite']:
             # We need to recalculate the number of invisible characters
             self.update_prompt(pname)
 
     lazy_evaluate_fields = Dict(help="""
         This maps field names used in the prompt templates to functions which
         will be called when the prompt is rendered. This allows us to include
         things like the current time in the prompts. Functions are only called
         if they are used in the prompt.
         """)
 
     in_template = Unicode('In [\\#]: ',
         help="Input prompt.  '\\#' will be transformed to the prompt number"
         ).tag(config=True)
     in2_template = Unicode('   .\\D.: ',
         help="Continuation prompt.").tag(config=True)
     out_template = Unicode('Out[\\#]: ',
         help="Output prompt. '\\#' will be transformed to the prompt number"
         ).tag(config=True)
 
     @default('lazy_evaluate_fields')
     def _lazy_evaluate_fields_default(self):
         return lazily_evaluate.copy()
 
     justify = Bool(True, help="""
         If True (default), each prompt will be right-aligned with the
         preceding one.
         """).tag(config=True)
 
     # We actually store the expanded templates here:
     templates = Dict()
 
     # The number of characters in the last prompt rendered, not including
     # colour characters.
     width = Int()
     txtwidth = Int()   # Not including right-justification
 
     # The number of characters in each prompt which don't contribute to width
     invisible_chars = Dict()
 
     @default('invisible_chars')
     def _invisible_chars_default(self):
         return {'in': 0, 'in2': 0, 'out': 0, 'rewrite':0}
 
     def __init__(self, shell, **kwargs):
         super(PromptManager, self).__init__(shell=shell, **kwargs)
 
         # Prepare colour scheme table
         self.color_scheme_table = coloransi.ColorSchemeTable([NoColor,
                                     LinuxColors, LightBGColors], self.color_scheme)
 
         self._formatter = UserNSFormatter(shell)
         # Prepare templates & numbers of invisible characters
         self.update_prompt('in', self.in_template)
         self.update_prompt('in2', self.in2_template)
         self.update_prompt('out', self.out_template)
         self.update_prompt('rewrite')
         self.observe(self._update_prompt_trait,
             names=['in_template', 'in2_template', 'out_template'])
 
     def update_prompt(self, name, new_template=None):
         """This is called when a prompt template is updated. It processes
         abbreviations used in the prompt template (like \#) and calculates how
         many invisible characters (ANSI colour escapes) the resulting prompt
         contains.
 
         It is also called for each prompt on changing the colour scheme. In both
         cases, traitlets should take care of calling this automatically.
         """
         if new_template is not None:
             self.templates[name] = multiple_replace(prompt_abbreviations, new_template)
         # We count invisible characters (colour escapes) on the last line of the
         # prompt, to calculate the width for lining up subsequent prompts.
         invis_chars = _invisible_characters(self._render(name, color=True))
         self.invisible_chars[name] = invis_chars
 
     def _update_prompt_trait(self, changes):
         traitname = changes['name']
         new_template = changes['new']
         name = traitname[:-9]   # Cut off '_template'
         self.update_prompt(name, new_template)
 
     def _render(self, name, color=True, **kwargs):
         """Render but don't justify, or update the width or txtwidth attributes.
         """
         if name == 'rewrite':
             return self._render_rewrite(color=color)
 
         if color:
             scheme = self.color_scheme_table.active_colors
             if name=='out':
                 colors = color_lists['normal']
                 colors.number, colors.prompt, colors.normal = \
                         scheme.out_number, scheme.out_prompt, scheme.normal
             else:
                 colors = color_lists['inp']
                 colors.number, colors.prompt, colors.normal = \
                         scheme.in_number, scheme.in_prompt, scheme.in_normal
                 if name=='in2':
                     colors.prompt = scheme.in_prompt2
         else:
             # No color
             colors = color_lists['nocolor']
             colors.number, colors.prompt, colors.normal = '', '', ''
 
         count = self.shell.execution_count    # Shorthand
         # Build the dictionary to be passed to string formatting
         fmtargs = dict(color=colors, count=count,
                        dots="."*len(str(count)), spaces=" "*len(str(count)),
                        width=self.width, txtwidth=self.txtwidth)
         fmtargs.update(self.lazy_evaluate_fields)
         fmtargs.update(kwargs)
 
         # Prepare the prompt
         prompt = colors.prompt + self.templates[name] + colors.normal
 
         # Fill in required fields
         return self._formatter.format(prompt, **fmtargs)
 
     def _render_rewrite(self, color=True):
         """Render the ---> rewrite prompt."""
         if color:
             scheme = self.color_scheme_table.active_colors
             # We need a non-input version of these escapes
             color_prompt = scheme.in_prompt.replace("\001","").replace("\002","")
             color_normal = scheme.normal
         else:
             color_prompt, color_normal = '', ''
 
         return color_prompt + "-> ".rjust(self.txtwidth, "-") + color_normal
 
     def render(self, name, color=True, just=None, **kwargs):
         """
         Render the selected prompt.
 
         Parameters
         ----------
         name : str
           Which prompt to render. One of 'in', 'in2', 'out', 'rewrite'
         color : bool
           If True (default), include ANSI escape sequences for a coloured prompt.
         just : bool
           If True, justify the prompt to the width of the last prompt. The
           default is stored in self.justify.
         **kwargs :
           Additional arguments will be passed to the string formatting operation,
           so they can override the values that would otherwise fill in the
           template.
 
         Returns
         -------
         A string containing the rendered prompt.
         """
         res = self._render(name, color=color, **kwargs)
 
         # Handle justification of prompt
         invis_chars = self.invisible_chars[name] if color else 0
-        self.txtwidth = _lenlastline(res) - invis_chars
+        # self.txtwidth = _lenlastline(res) - invis_chars
         just = self.justify if (just is None) else just
         # If the prompt spans more than one line, don't try to justify it:
         if just and name != 'in' and ('\n' not in res) and ('\r' not in res):
             res = res.rjust(self.width + invis_chars)
-        self.width = _lenlastline(res) - invis_chars
+        # self.width = _lenlastline(res) - invis_chars
         return res