mgedmin

A list of interesting GH repos:

• https://github.com/vim-scripts/rest.vim (rest.vim from 2004, "derived somewhat from Nikolai Weibull's" syntax file)
• https://github.com/mitsuhiko/vim-rst (syntax, based on Nikolai Weibull's rst syntax, no documentation)
• https://github.com/mrsipan/vim-rst (mappings, folding, syntax based on Nikolai Weibull's rst syntax, no documentation)
• https://github.com/khorser/vim-rst-ftplugin (no syntax, ftplugin for folding, mappings for navigation, modification)
• https://github.com/veselosky/vim-rst (a ftplugin that sets nolist and a snippets file with one snippet)
• https://github.com/edthedev/rst-syntax.vim (sets filetype=rst for *.txt)
• https://github.com/yang-ling/rst/blob/master/ftplugin/rst.vim (a ftplugin that sets indent options)

mgedmin

#!/usr/bin/python
"""
Parse email messages with Python tracebacks, add email headers to group them.

Usage with procmail:

    # Traceback emails!
    :0 fw
    * ^From: [email protected]
    * ^TO_.*[email protected]

mgedmin

style "menu_item"
{
	xthickness	= 2
	ythickness	= 4

        # HACK: Gtk doesn't actually read this value
        # while rendering the menu items, but Libreoffice
        # does; setting this value equal to the one in
        # fg[PRELIGHT] ensures a code path in the LO theming code
        # that falls back to a dark text color for menu item text

mgedmin

<mgedmin> I'm seeing very weird segfaults in glib
 in g_closure_invoke(), to be precise
 this is on ubuntu gnome 14.10 with their gnome3-staging ppa enabled, so who knows what got broken where
 but basically all the apps are segfaulting left and right when widgets lose focus
 the stack trace is this: https://gist.github.com/0c80f6e81d9dfc4a4fd7
 I have a core dump and I can poke around it with gdb
 and gdb tells me closure->marshal is 0x7fd4ba283040 <g_cclosure_marshal_VOID__BOOLEAN>
 and real_closure->meta_marshal is 0x7fd4ba27fb10 <g_type_class_meta_marshal>
 the segfault is here: https://github.com/GNOME/glib/blob/glib-2-42/gobject/gclosure.c#L768
 it's invoking a local variable 'marshal' which is "optimized out" in the stack trace's locals

mgedmin

mg@platonas: ~/src/apt-sources/glib2.0-2.42.0 $ valgrind gnome-calculator 
==9270== Memcheck, a memory error detector
==9270== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==9270== Using Valgrind-3.10.0 and LibVEX; rerun with -h for copyright info
==9270== Command: gnome-calculator
==9270== 
==9270== Conditional jump or move depends on uninitialised value(s)
==9270==    at 0x7871324: composite_traps (cairo-xlib-render-compositor.c:1882)
==9270==    by 0x78591AE: composite_traps (cairo-traps-compositor.c:1098)
==9270==    by 0x78591AE: composite_mask_clip (cairo-traps-compositor.c:2029)

mgedmin

# This is what I need to do if I want to import a .zexp from Zope 2.12 because there's apparently some object in the ZODB that got removed in 2.11 days
    import mock
    sys.modules['Products.PluginIndexes.TextIndex'] = mock.Mock()
    sys.modules['Products.PluginIndexes.TextIndex.Splitter'] = mock.Mock()
    sys.modules['Products.PluginIndexes.TextIndex.Splitter.ZopeSplitter'] = mock.Mock()
    from persistent import Persistent
    sys.modules['Products.PluginIndexes.TextIndex.Splitter.ZopeSplitter'].ZopeSplitter = type('ZopeSplitter', (Persistent,), {'__module__': 'Products.PluginIndexes.TextIndex.Splitter.ZopeSplitter'})

mgedmin

from gi.repository import GLib, Gtk

def make_option(long_name, short_name=None, flags=0, arg=GLib.OptionArg.NONE,
                arg_data=None, description=None, arg_description=None):
    # surely something like this should exist inside PyGObject itself?!
    option = GLib.OptionEntry()
    option.long_name = long_name.lstrip('-')
    option.short_name = 0 if not short_name else short_name.lstrip('-')
    option.flags = flags
    option.arg = arg

mgedmin

#!/bin/sh
filename=${1:-bootstrap.py}
test -f "$filename" || {
    echo "there is no $filename here"
    exit 1
}
checksum=$(sha1sum "$filename")
case "$checksum" in
    0b63906b1c5d2eccbc9f727b0c1374c7c29f06a2*)
        echo "you've got the obsolete v1.x bootstrap.py"

mgedmin

#!/bin/bash
echo "Using ~/bin/transset instead of the broken /usr/bin/transset"
echo "See https://bugs.freedesktop.org/show_bug.cgi?id=76958"
opacity=${1:-0.75}
scaled=$(echo "$opacity*4294967295"|bc)
hex=$(printf "0x%x" ${scaled%.*})
echo "Setting opacity to $opacity"
xprop -f _NET_WM_WINDOW_OPACITY 32c -set _NET_WM_WINDOW_OPACITY $hex

mgedmin

#!/bin/bash
# Grep for a string in all the Python packages in the buildout path
# Assumes . is the directory with a buildout.cfg
# Note: there is no one true buildout path -- every part has its own.  So
# we look for a python interpreter in bin/python (or bin/py) and use its path.
# If that fails, we try to find a bin/test script and scrape the path from it.
# Should also work with virtualenv (just run it somewhere with a bin/python)
# Based on buildout-tags, which lives at https://gist.github.com/mgedmin/5152189

progname=${0##*/}