##// END OF EJS Templates
revset: special case commonancestors(none()) to be empty set...
revset: special case commonancestors(none()) to be empty set This matches the behavior of ancestor(none()). From an implementation perspective, ancestor() and commonancestors() are intersection, and ancestors() is union, so it would make some sense that commonancestors(none()) returned all revisions. However, ancestor(none()) isn't implemented as such, which breaks ancestor(x) == max(commonancestors(x)). From a user perspective, ancestors of nothing is nothing whichever type of operation the ancestor predicate does.

File last commit:

r38359:9abe91a5 default
r38727:e4b270a3 default
Show More
beautifygraph.py
93 lines | 2.9 KiB | text/x-python | PythonLexer
# -*- coding: UTF-8 -*-
# beautifygraph.py - improve graph output by using Unicode characters
#
# Copyright 2018 John Stiles <johnstiles@gmail.com>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
'''beautify log -G output by using Unicode characters (EXPERIMENTAL)
A terminal with UTF-8 support and monospace narrow text are required.
'''
from __future__ import absolute_import
from mercurial.i18n import _
from mercurial import (
encoding,
extensions,
graphmod,
templatekw,
)
# Note for extension authors: ONLY specify testedwith = 'ships-with-hg-core' for
# extensions which SHIP WITH MERCURIAL. Non-mainline extensions should
# be specifying the version(s) of Mercurial they are tested with, or
# leave the attribute unspecified.
testedwith = 'ships-with-hg-core'
def prettyedge(before, edge, after):
if edge == '~':
return '\xE2\x95\xA7' # U+2567 ╧
if edge == 'X':
return '\xE2\x95\xB3' # U+2573 ╳
if edge == '/':
return '\xE2\x95\xB1' # U+2571 ╱
if edge == '-':
return '\xE2\x94\x80' # U+2500 ─
if edge == '|':
return '\xE2\x94\x82' # U+2502 │
if edge == ':':
return '\xE2\x94\x86' # U+2506 ┆
if edge == '\\':
return '\xE2\x95\xB2' # U+2572 ╲
if edge == '+':
if before == ' ' and not after == ' ':
return '\xE2\x94\x9C' # U+251C ├
if after == ' ' and not before == ' ':
return '\xE2\x94\xA4' # U+2524 ┤
return '\xE2\x94\xBC' # U+253C ┼
return edge
def convertedges(line):
line = ' %s ' % line
pretty = []
for idx in xrange(len(line) - 2):
pretty.append(prettyedge(line[idx], line[idx + 1], line[idx + 2]))
return ''.join(pretty)
def getprettygraphnode(orig, *args, **kwargs):
node = orig(*args, **kwargs)
if node == 'o':
return '\xE2\x97\x8B' # U+25CB ○
if node == '@':
return '\xE2\x97\x8D' # U+25CD ◍
if node == '*':
return '\xE2\x88\x97' # U+2217 ∗
if node == 'x':
return '\xE2\x97\x8C' # U+25CC ◌
if node == '_':
return '\xE2\x95\xA4' # U+2564 ╤
return node
def outputprettygraph(orig, ui, graph, *args, **kwargs):
(edges, text) = zip(*graph)
graph = zip([convertedges(e) for e in edges], text)
return orig(ui, graph, *args, **kwargs)
def extsetup(ui):
if encoding.encoding != 'UTF-8':
ui.warn(_('beautifygraph: unsupported encoding, UTF-8 required\n'))
return
if 'A' in encoding._wide:
ui.warn(_('beautifygraph: unsupported terminal settings, '
'monospace narrow text required\n'))
return
if ui.plain('graph'):
return
extensions.wrapfunction(graphmod, 'outputgraph', outputprettygraph)
extensions.wrapfunction(templatekw, 'getgraphnode', getprettygraphnode)