no message

This commit is contained in:
Yann Esposito (Yogsototh) 2010-04-15 11:20:27 +02:00
parent 56bb9b5699
commit 0adfc39d8d
7 changed files with 758 additions and 0 deletions

2
.gitignore vendored
View file

@ -6,3 +6,5 @@ output/Scratch/en
output/Scratch/fr
output/Scratch/map.yrf
tmp/
recupen.pl
recupfr.pl

View file

@ -0,0 +1,118 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-02-23T10:09:52+02:00
title: When regexp is not the best solution
multiTitle:
fr: When regexp is not the best solution
en: When regexp is not the best solution
multiDescription:
fr: pas de description.
en: no description.
tags:
- programming
- regexp
- regular expression
- extension
- file
-----
Regular expression are really useful. Unfortunately, they are not always the best way of doing things.
Particularly when transformations you want to make are easy.
I wanted to know how to get file extension from filename the fastest way possible. There is 3 natural way of doing this:
<div><code class="ruby">
# regexp
str.match(/[^.]*$/);
ext=$&
# split
ext=str.split('.')[-1]
# File module
ext=File.extname(str)
</code></div>
At first sight I believed that the regexp should be faster than the split because it could be many `.` in a filename. But in reality, most of time there is only one dot and I realized the split will be faster. But not the fastest way. There is a function dedicated to this work in the `File` module.
Here is the Benchmark ruby code:
<div><code class="ruby" file="regex_benchmark_ext.rb">
#!/usr/bin/env ruby
require 'benchmark'
n=80000
tab=[ '/accounts/user.json',
'/accounts/user.xml',
'/user/titi/blog/toto.json',
'/user/titi/blog/toto.xml' ]
puts "Get extname"
Benchmark.bm do |x|
x.report("regexp:") { n.times do
str=tab[rand(4)];
str.match(/[^.]*$/);
ext=$&;
end }
x.report(" split:") { n.times do
str=tab[rand(4)];
ext=str.split('.')[-1] ;
end }
x.report(" File:") { n.times do
str=tab[rand(4)];
ext=File.extname(str);
end }
end
</code></div>
And here is the result
<pre class="twilight">
Get extname
user system total real
regexp: 2.550000 0.020000 2.570000 ( 2.693407)
split: 1.080000 0.050000 1.130000 ( 1.190408)
File: 0.640000 0.030000 0.670000 ( 0.717748)
</pre>
Conclusion of this benchmark, dedicated function are better than your way of doing stuff (most of time).
## file path without the extension.
<div><code class="ruby" file="regex_benchmark_strip.rb">
#!/usr/bin/env ruby
require 'benchmark'
n=80000
tab=[ '/accounts/user.json',
'/accounts/user.xml',
'/user/titi/blog/toto.json',
'/user/titi/blog/toto.xml' ]
puts "remove extension"
Benchmark.bm do |x|
x.report(" File:") { n.times do
str=tab[rand(4)];
path=File.expand_path(str,File.basename(str,File.extname(str)));
end }
x.report("chomp:") { n.times do
str=tab[rand(4)];
ext=File.extname(str);
path=str.chomp(ext);
end }
end
</code></div>
and here is the result:
<pre class="twilight">
remove extension
user system total real
File: 0.970000 0.060000 1.030000 ( 1.081398)
chomp: 0.820000 0.040000 0.860000 ( 0.947432)
</pre>
Conclusion of the second benchmark. One simple function is better than three dedicated functions. No surprise, but it is good to know.

View file

@ -0,0 +1,78 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-03-22T10:42:27+02:00
title: Git Tips
multiTitle:
fr: Git Tips
en: Git Tips
multiDescription:
fr: pas de description.
en: no description.
tags:
- git
- tip
-----
## clone from github behind an evil firewall
Standard:
<div><code class="zsh">
git clone git@github.com:yogsototh/project.git
</code></div>
Using HTTPS port:
<div><code class="zsh">
git clone git+ssh://git@github.com:443/yogsototh/project.git
</code></div>
## clone all branches
`git clone` can only fetch the master branch.
If you don't have much branches, you can simply use clone your project and then use the following command:
<div><code class="zsh">
git branch --track local_branch remote_branch
</code></div>
for example:
<div><code class="zsh">
$ git clone git@github:yogsototh/example.git
$ git branch
master *
$ git branch -a
master *
remotes/origin/HEAD -> origin/master
remotes/origin/experimental
$ git branch --track experimental remotes/origin/experimental
$ git branch
master *
experimental
</code></div>
If you have many branches it can be useful to use the following *script*/*long command line*.
<div><code class="zsh">
# first clone your project
$ git clone git@github.com:yogsototh/project.git
# copy all branches
$ zsh
$ cd project
$ for br in $( git br -a ); do
case $br in
remotes/*) print $br ;
case ${br:t} in
master|HEAD) continue ;;
*) git branch --track ${br:t} $br ;;
esac ;;
esac
done
</code></div>

View file

@ -0,0 +1,182 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-03-23T22:37:36+02:00
title: Encapsulate git
multiTitle:
fr: Encapsuler git
en: Encapsulate git
multiDescription:
fr: pas de description.
en: no description.
tags:
- git
- protection
- branches
- diverged
-----
<div class="small">
Here is a solution to maintain divergent branches in git. Because it is easy to merge by mistake. I give a script that encapsulate git in order to forbid some merge and warn you some merge should be dangerous.
</div>
## how to protect against your own dumb
I work on a project in which some of my git branches should remain divergent. And divergences should grow.
I also use some branch to contain what is common between projects.
Say I have some branches:
master: common to all branches
dev: branch devoted to unstable development
client: branch with features for all client but not general enough for master
clientA: project adapted for client A
clientB: project adapted for client B
Here how I want to work:
<%= blogimage("Dynamic branching","dynamic_branching.png") %>
And more precisely the branch hierarchy:
<%= blogimage("Branch hierarchy","branch_hierarchy.png") %>
An arrow from A to B means, you can merge A in B. If there is no arrow from A to B that means it is *forbidden* to merge A in B. Here is the corresponding rubycode:
<div><code class="ruby">
$architecture={
:master => [ :dev, :client ],
:dev => [ :master ],
:client => [ :clientA, :clientB ] }
</code></div>
Having a `:master => [ :dev, :client ]` means you can merge `master` branch into `dev` and `client`.
If by mistake I make a `git checkout master && git merge clientA`, I made a mistake. This is why I made a script which encapsulate the git behaviour to dodge this kind of mistake.
But this script do far more than that. It also merge from top to down. The action `allmerges` will do:
<div><code class="zsh">
git co dev && git merge master
git co client && git merge master
git co clientA && git merge client
git co clientB && git merge client
</code></div>
That means, I can update all branches. The algorithm will not make loop even if there is a cycle in the branch hierarchy.
Here it is:
<div class="small"><code class="ruby" file="eng">
#!/usr/bin/env ruby
# encoding: utf-8
# architecture
#
# master <-> dev
# master -> client
# clien -> clientA | clientB
#
# merge using two of these branches should be
# restricted to these rules
# merge to one of these branch and an unknown one should
# raise a warning, and may the option to add this new branch
# to the hierarchy
$architecture={
:master => [ :dev, :client ],
:dev => [ :master ],
:client => [ :clientA, :clientB ] }
def get_current_branch()
(`git branch --no-color | awk '$1 == "*" {print $2}'`).chop.intern
end
if ARGV.length == 0
puts %{usage: $0:t [git_command or local_command]
local commands:
allmerges: merge from top to down}
exit 0
end
require 'set'
$known_branches=Set.new
$architecture.each do |k,v|
$known_branches.add(k)
v.each { |b| $known_branches.add(b) }
end
def rec_merge(branch)
if $architecture[branch].nil?
return
end
$architecture[branch].each do |b|
if $flag.has_key?(b.to_s + branch.to_s)
next
end
flagname=branch.to_s + b.to_s
if $flag.has_key?(flagname)
next
end
if system %{eng checkout #{b}}
if get_current_branch != b
puts "Can't checkout to #{b}"
exit 2
end
if system %{eng merge #{branch}}
$flag[flagname]=true
rec_merge(b)
else
exit 1
end
else
exit 1
end
end
end
def do_all_merges
puts 'Will merge from father to sons'
current_branch=get_current_branch
$flag={}
rec_merge(:master)
system %{git co #{current_branch}}
end
def do_merge
current_branch=get_current_branch
src_branch=ARGV[1].intern
puts %{do_merge: #{src_branch} => #{current_branch}}
if $known_branches.include?(current_branch)
if $known_branches.include?(src_branch)
if $architecture.has_key?(src_branch) and
$architecture[src_branch].include?(current_branch)
system %{git merge #{src_branch}}
else
puts %{Forbidden merge: #{src_branch} => #{current_branch}}
end
else
puts %{Warning! #{src_branch} not mentionned in rb configuration}
sleep 2
f system %{git merge #{src_branch}}
puts %{Warning! #{src_branch} not mentionned in rb configuration}
end
end
end
case ARGV[0]
when 'allmerges' then do_all_merges
when 'merge' then do_merge
else system %{git #{ARGV.join(' ')}}
end
</code></div>
All you need to do to make it work is simply to copy eng in a directory contained in your PATH.
Of course try to use as few as possible `cherry-pick` and `rebase`. This script was intended to work with workflow using `pull` and `merge`.

View file

@ -0,0 +1,118 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-02-23T10:09:52+02:00
title: When regexp is not the best solution
multiTitle:
fr: When regexp is not the best solution
en: When regexp is not the best solution
multiDescription:
fr: pas de description.
en: no description.
tags:
- programming
- regexp
- regular expression
- extension
- file
-----
Regular expression are really useful. Unfortunately, they are not always the best way of doing things.
Particularly when transformations you want to make are easy.
I wanted to know how to get file extension from filename the fastest way possible. There is 3 natural way of doing this:
<div><code class="ruby">
# regexp
str.match(/[^.]*$/);
ext=$&
# split
ext=str.split('.')[-1]
# File module
ext=File.extname(str)
</code></div>
At first sight I believed that the regexp should be faster than the split because it could be many `.` in a filename. But in reality, most of time there is only one dot and I realized the split will be faster. But not the fastest way. There is a function dedicated to this work in the `File` module.
Here is the Benchmark ruby code:
<div><code class="ruby" file="regex_benchmark_ext.rb">
#!/usr/bin/env ruby
require 'benchmark'
n=80000
tab=[ '/accounts/user.json',
'/accounts/user.xml',
'/user/titi/blog/toto.json',
'/user/titi/blog/toto.xml' ]
puts "Get extname"
Benchmark.bm do |x|
x.report("regexp:") { n.times do
str=tab[rand(4)];
str.match(/[^.]*$/);
ext=$&;
end }
x.report(" split:") { n.times do
str=tab[rand(4)];
ext=str.split('.')[-1] ;
end }
x.report(" File:") { n.times do
str=tab[rand(4)];
ext=File.extname(str);
end }
end
</code></div>
And here is the result
<pre class="twilight">
Get extname
user system total real
regexp: 2.550000 0.020000 2.570000 ( 2.693407)
split: 1.080000 0.050000 1.130000 ( 1.190408)
File: 0.640000 0.030000 0.670000 ( 0.717748)
</pre>
Conclusion of this benchmark, dedicated function are better than your way of doing stuff (most of time).
## file path without the extension.
<div><code class="ruby" file="regex_benchmark_strip.rb">
#!/usr/bin/env ruby
require 'benchmark'
n=80000
tab=[ '/accounts/user.json',
'/accounts/user.xml',
'/user/titi/blog/toto.json',
'/user/titi/blog/toto.xml' ]
puts "remove extension"
Benchmark.bm do |x|
x.report(" File:") { n.times do
str=tab[rand(4)];
path=File.expand_path(str,File.basename(str,File.extname(str)));
end }
x.report("chomp:") { n.times do
str=tab[rand(4)];
ext=File.extname(str);
path=str.chomp(ext);
end }
end
</code></div>
and here is the result:
<pre class="twilight">
remove extension
user system total real
File: 0.970000 0.060000 1.030000 ( 1.081398)
chomp: 0.820000 0.040000 0.860000 ( 0.947432)
</pre>
Conclusion of the second benchmark. One simple function is better than three dedicated functions. No surprise, but it is good to know.

View file

@ -0,0 +1,78 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-03-22T10:42:27+02:00
title: Git Tips
multiTitle:
fr: Git Tips
en: Git Tips
multiDescription:
fr: pas de description.
en: no description.
tags:
- git
- tip
-----
## clone from github behind an evil firewall
Standard:
<div><code class="zsh">
git clone git@github.com:yogsototh/project.git
</code></div>
Using HTTPS port:
<div><code class="zsh">
git clone git+ssh://git@github.com:443/yogsototh/project.git
</code></div>
## clone all branches
`git clone` can only fetch the master branch.
If you don't have much branches, you can simply use clone your project and then use the following command:
<div><code class="zsh">
git branch --track local_branch remote_branch
</code></div>
for example:
<div><code class="zsh">
$ git clone git@github:yogsototh/example.git
$ git branch
master *
$ git branch -a
master *
remotes/origin/HEAD -> origin/master
remotes/origin/experimental
$ git branch --track experimental remotes/origin/experimental
$ git branch
master *
experimental
</code></div>
If you have many branches it can be useful to use the following *script*/*long command line*.
<div><code class="zsh">
# first clone your project
$ git clone git@github.com:yogsototh/project.git
# copy all branches
$ zsh
$ cd project
$ for br in $( git br -a ); do
case $br in
remotes/*) print $br ;
case ${br:t} in
master|HEAD) continue ;;
*) git branch --track ${br:t} $br ;;
esac ;;
esac
done
</code></div>

View file

@ -0,0 +1,182 @@
-----
# Custom
isHidden: false
menupriority: 1
kind: article
created_at: 2010-03-23T22:37:36+02:00
title: Encapsulate git
multiTitle:
fr: Encapsuler git
en: Encapsulate git
multiDescription:
fr: pas de description.
en: no description.
tags:
- git
- protection
- branches
- diverged
-----
<div class="small">
Here is a solution to maintain divergent branches in git. Because it is easy to merge by mistake. I give a script that encapsulate git in order to forbid some merge and warn you some merge should be dangerous.
</div>
## how to protect against your own dumb
I work on a project in which some of my git branches should remain divergent. And divergences should grow.
I also use some branch to contain what is common between projects.
Say I have some branches:
master: common to all branches
dev: branch devoted to unstable development
client: branch with features for all client but not general enough for master
clientA: project adapted for client A
clientB: project adapted for client B
Here how I want to work:
<%= blogimage("Dynamic branching","dynamic_branching.png") %>
And more precisely the branch hierarchy:
<%= blogimage("Branch hierarchy","branch_hierarchy.png") %>
An arrow from A to B means, you can merge A in B. If there is no arrow from A to B that means it is *forbidden* to merge A in B. Here is the corresponding rubycode:
<div><code class="ruby">
$architecture={
:master => [ :dev, :client ],
:dev => [ :master ],
:client => [ :clientA, :clientB ] }
</code></div>
Having a `:master => [ :dev, :client ]` means you can merge `master` branch into `dev` and `client`.
If by mistake I make a `git checkout master && git merge clientA`, I made a mistake. This is why I made a script which encapsulate the git behaviour to dodge this kind of mistake.
But this script do far more than that. It also merge from top to down. The action `allmerges` will do:
<div><code class="zsh">
git co dev && git merge master
git co client && git merge master
git co clientA && git merge client
git co clientB && git merge client
</code></div>
That means, I can update all branches. The algorithm will not make loop even if there is a cycle in the branch hierarchy.
Here it is:
<div class="small"><code class="ruby" file="eng">
#!/usr/bin/env ruby
# encoding: utf-8
# architecture
#
# master <-> dev
# master -> client
# clien -> clientA | clientB
#
# merge using two of these branches should be
# restricted to these rules
# merge to one of these branch and an unknown one should
# raise a warning, and may the option to add this new branch
# to the hierarchy
$architecture={
:master => [ :dev, :client ],
:dev => [ :master ],
:client => [ :clientA, :clientB ] }
def get_current_branch()
(`git branch --no-color | awk '$1 == "*" {print $2}'`).chop.intern
end
if ARGV.length == 0
puts %{usage: $0:t [git_command or local_command]
local commands:
allmerges: merge from top to down}
exit 0
end
require 'set'
$known_branches=Set.new
$architecture.each do |k,v|
$known_branches.add(k)
v.each { |b| $known_branches.add(b) }
end
def rec_merge(branch)
if $architecture[branch].nil?
return
end
$architecture[branch].each do |b|
if $flag.has_key?(b.to_s + branch.to_s)
next
end
flagname=branch.to_s + b.to_s
if $flag.has_key?(flagname)
next
end
if system %{eng checkout #{b}}
if get_current_branch != b
puts "Can't checkout to #{b}"
exit 2
end
if system %{eng merge #{branch}}
$flag[flagname]=true
rec_merge(b)
else
exit 1
end
else
exit 1
end
end
end
def do_all_merges
puts 'Will merge from father to sons'
current_branch=get_current_branch
$flag={}
rec_merge(:master)
system %{git co #{current_branch}}
end
def do_merge
current_branch=get_current_branch
src_branch=ARGV[1].intern
puts %{do_merge: #{src_branch} => #{current_branch}}
if $known_branches.include?(current_branch)
if $known_branches.include?(src_branch)
if $architecture.has_key?(src_branch) and
$architecture[src_branch].include?(current_branch)
system %{git merge #{src_branch}}
else
puts %{Forbidden merge: #{src_branch} => #{current_branch}}
end
else
puts %{Warning! #{src_branch} not mentionned in rb configuration}
sleep 2
f system %{git merge #{src_branch}}
puts %{Warning! #{src_branch} not mentionned in rb configuration}
end
end
end
case ARGV[0]
when 'allmerges' then do_all_merges
when 'merge' then do_merge
else system %{git #{ARGV.join(' ')}}
end
</code></div>
All you need to do to make it work is simply to copy eng in a directory contained in your PATH.
Of course try to use as few as possible `cherry-pick` and `rebase`. This script was intended to work with workflow using `pull` and `merge`.