Rake project structure,? good?


#1

So more Raking - wanting to divide up projects in subdirs into
sandbox directories but also have parent rakefile actually include their
“children” into the dependency graph to be more efficient and accurate.
and yet allow each “sandbox” rakefile.

I have come up with the below and want to know if this is a ‘best’ way
or if people see it as silly and there is a more “normal” method
(I am a recent Ruby/Rake newbee)

pardon tabs expand big here

a ‘parent’ rakefile

task :default => [:test]

load all rakefiles for sub projects, create tasks

and dependecy graph

rakefiles = FileList.new(File.expand_path(’**/rakefile.rb’))
myDir = pwd # save here
rakefiles.each do |rakefile|
cd File.dirname(rakefile)
load ‘rakefile.rb’
end
cd myDir # restore here

:subtest is in a subdirectory module

task :test => [:subtest] do |t|
p “root test”
end

rakefile in a subdirectory

subdir/rakefile

while “file” defacto anonymous module

Module.new do
public # defacto private nut public for internal blocks

variables are “instance” variables

@me = object_id # unique for this Module and small
def doIt
puts “doing it in #{@me}”
end
module_function :doIt # is there an easier way to do this?

namespace @me do # unique (private) namespace for this module
# private task definitions for this module
task :subtest do |t|
doIt
puts “submake subtest in #{@me}”
end
end

all file tasks used to create fules and tasks

must be full paths so they are all unique

and no file task will be messed up ???

files = FileList.new(File.expand_path(’*’));

public tasks (targets) seen by parent namespace

task :subtest => [ “#{@me}:subtest”, files ].flatten do |t|
puts “submake test”
end

end # end private module