Read The Text

[readthetxt]

Pure AWK Base64 decoder for LDIF

What if your LDIF content contains BASE64 encoded values and you want it more human readable?

It can be done using pure AWK. Feel free to use the following AWK script.

/^\w+:.+/ || /^$/ { if(b64) { obc=0 printf attr" " for(i=1;i<=length(b64);i++) { c=index( \ "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/", \ substr(b64,i,1) \ ) if(c--) { for(b=0;b<6;b++) { o=o*2+int(c/32) c=(c*2)%64 if(++obc==8) { if (o>31 || o==9) { printf "%c",o } else { printf "." } obc=0 o=0 } } } } b64="" print b46 } } /^\w+:: .+/ { attr=$1; b64=$NF } /^ .+/ { if (b64){ b64=b64$NF } } { if(!b64) { print } }

I have shell alias:

alias un64='awk "/^\w+:.+/ || /^$/ { if(b64) { obc=0; printf attr\" \"; for(i=1;i31 || o==9 {printf \"%c\",o} else {printf \".\"} obc=0; o=0; }} }}; b64=\"\"; \\ print b46} } /^\w+:: .+/ {attr=\$1; b64=\$NF} /^ .+/ {if (b64){b64=b64\$NF} } \\ { if(!b64) {print} }"'

and LDIF with encoded attributes:

# extended LDIF # # LDAPv3 # base with scope subtree # filter: (...) # requesting: displayName # # postgres, Services, UNIX, test dn: CN=postgres,OU=Services,OU=UNIX,DC=test displayName:: 0J/RgNC+0LLQtdGA0LrQsA== # search result search: 2 result: 0 Success # numResponses: 2 # numEntries: 1

Now you can decode on the fly:

$ ldapsearch ... displayName | un64 # extended LDIF # # LDAPv3 # base with scope subtree # filter: (...) # requesting: displayName # # postgres, Services, UNIX, test dn: CN=postgres,OU=Services,OU=UNIX,DC=test displayName:: Проверка # search result search: 2 result: 0 Success # numResponses: 2 # numEntries: 1

[readthetxt]

Simple ruby native configurator

If you need a simple ruby native configurator for your ruby application, use the following construction

module SomeConfig extend self attr_accessor :variable def define yield self end end SomeConfig.define do |config| config.variable = "value" end puts SomeConfig.variable

[readthetxt]

Quick maintenance page with nginx

There are often situations when the web-service should display maintenance page, something like "Maintenance, come back later!". It is not always an application that implements this functionality. But if we have nginx "before" the our application then you can do something like this:

location @ban try_files /stub.html =404; } error_page 418 =200 @ban; location / { if ( -f /etc/stub ) { return 418; } proxy_pass http://backends:8080; }

Now you can run touch /etc/stub and your nginx will response with stub.html fro root dir. Turn on proxying again by /etc/stub removing. And you don't need restart your nginx.

[readthetxt]

Unix: recursive grep in source code

Big repository, a lot of files. How can I do recursive grep more clean? I don't want match files ander .svn or other "waste" directories. I want search only in my code! Let's do it in shell! Put this code into your .bashrc or .zshrc:

g() { exclude_regex=".*\.(svn|git|hg)/.*" grep_dir=${2:-./} case $(hostname|tr '[:upper:]' '[:lower:]') in linux) find -L $grep_dir -regextype posix-extended -type f ! -regex $exclude_regex -exec grep $1 {} + ;; freebsd) find -EL $grep_dir -type f ! -regex "$exclude_regex" -exec grep $1 {} + ;; *) #POSIX.1 compat. for others find -L $grep_dir -type f | egrep -v $exclude_regex | xargs grep $1 ;; esac }

And now, for grep from current directory: bash$ g someMethod or for grep from ./mycode directory: bash$ g "class MyClass" ./mycode You can also make grep output more comfortable for your eyes. Just put this line in .bashrc/.zshrc:

[ -z "`grep --help 2>&1|grep -- '--color'`" ] || export GREP_OPTIONS='--color=auto'