A merge commit has two parent sha1s, so the old branch would be accessible by its sha1 as one of the parents in the merge commit and thus wouldn't be garbage-collected by Git.
# to read it again every time we're testing the extraction of their | |
# voting history (which is still buggy) | |
if File.exists?("legislators.yml") | |
legislators = File.open("legislators.yml") { |f| YAML::load(f) } | |
else | |
page = agent.get('http://www2.camara.gov.br/deputados') | |
form = page.form('form1') | |
select = form.fields[5] |
# Grep a line in a project's entire network | |
# (or all the forks you included as remote anyway) | |
irb(main):001:0> pattern = "task :clear" | |
irb(main):002:0> `find .git/refs/remotes/ -type f`.split("\n").map{|e| e.gsub(/\.git\/refs\/remotes\//, '')}.each {|e| puts `git grep #{pattern} #{e}`} |
class AddStatsToDelayedJobs < ActiveRecord::Migration | |
def self.up | |
add_column :delayed_jobs, :first_started_at, :datetime | |
add_column :delayed_jobs, :last_started_at, :datetime | |
add_column :delayed_jobs, :finished_at, :datetime | |
end | |
def self.down | |
remove_column :delayed_jobs, :finished_at | |
remove_column :delayed_jobs, :last_started_at |
#!/usr/bin/env ruby | |
require 'rubygems' | |
require 'json' | |
require 'open-uri' | |
def google_fight(queries) | |
queries.max { |a,b| JSON.load(open("http://ajax.googleapis.com/ajax/services/search/web?v=1.0&q=#{a}"))["responseData"]["cursor"]["estimatedResultCount"].to_i <=> JSON.load(open("http://ajax.googleapis.com/ajax/services/search/web?v=1.0&q=#{b}"))["responseData"]["cursor"]["estimatedResultCount"].to_i } | |
end | |
if __FILE__ == $0 |
#!/usr/bin/env ruby | |
require 'rubygems' | |
require 'mechanize' | |
require 'ostruct' | |
require 'open-uri' | |
require 'hpricot' | |
require 'yaml' | |
# fix for screwed up net/http | |
# found at: http://pw.tech-arts.co.jp/technical/cat57/ |
class Hash | |
def assign_with_debugger(key, value) | |
debugger if key == "action_dispatch.request.path_parameters" | |
self.assign_without_debugger(key, value) | |
end | |
alias :assign_without_debugger :[]= | |
alias :[]= :assign_with_debugger | |
end |
# -*- coding: utf-8 -*- | |
class String | |
def downcase_with_accents | |
return nil if s.nil? | |
norm = self.downcase | |
norm.tr!('ÁÉÍÓÚÇ', 'aeiouç') | |
norm.tr!('ÀÈÌÒÙ', 'aeiou') | |
norm.tr!('ÄËÏÖÜ', 'aeiou') | |
norm.tr!('ÂÊÎÔÛ', 'aeiou') | |
norm.tr!('áéíóú', 'aeiou') |
Deleting branches after merging them into other branches can help unclutter your list of branches. If you use a lot of topic branches for features and don't delete them, pretty soon you'll have a very long list of old, unmaintained branches, and that's going to make it harder for you to find the ones you're currently working on.
You can always get back to old branches anyway. When you merge, a merge commit is created on the branch you merged into. Merge commits have two parents, which are referenced by their SHA1s. So if you need to see the old branch, just check out the SHA1 of the corresponding parent on the merge commit. Since it's referenced in a living branch, it's not gonna get garbage-collected by Git.
The only case where the original SHA1 wouldn't be kept around is if it was a fast-forward merge. To prevent that from happening, you can use the option --no-ff
on the git merge
command, which will create a merge commit even when a fast forward would be possible.
Git stash saves your uncommited work on a temporary stack and removes it from the current workspace, allowing you to switch branches without seeing complaints about uncommited changes.