July 2009 Archives

I recently had a need for a YUI (Yahoo! User Interface) TreeView with Drag & Drop nodes.  Here's a little background on the project.  My web application uses YUI Menu for navigation and is driven by a single database table.  Initially I threw together a quick and dirty CGI CRUD interface to manage the menu.  But, the more I had to modify the menu with this interface the more I needed something easier to use.  So I decided to scrap the CGI interface and start over with a nice 100% Ajax interface with a YUI TreeView to represent the YUI Menu.  Adding, changing, and deleting nodes is pretty easy, but to easily change the order of the Menu items, I added YUI Drag & Drop to the TreeView.  

1. Begin by following the steps for merging the branch here (assumes you are using svnmerge.py):
   http://www.orcaware.com/svn/wiki/Svnmerge.py#Merging_development_branches_back_to_trunk
   
   
2. Handle conflicts. If the branch is up-to-date, simply copy the latest revision file to the conflicted file. 
   
   
   Example: Let's say the latest revision is 100 and after merging a branch to the trunk, conflicts are present. The following shell one-liner will automatically resolve the conflicts, but only if the branch version is authoritative. If the trunk version has features not yet added to the branch version, you must manually resolve the conflict. 
   
   
   prompt:~# for a in *.r100; do b=`perl -e 'print split /\.merge-right.r100/, shift;' $a`; echo "copying $a $b"; cp $a $b; svn resolved $b; echo; done
   
   If this makes you nervous you can always consult the manual: http://svnbook.red-bean.com/en/1.5/svn.tour.cycle.html#svn.tour.cycle.resolve
   
   
3. Commit the merge: svn ci -F svnmerge-commit-message.txt

Your branch should now be successfully merged into the trunk!

I like to use join and map in order to cut down on too much static code and to simplify building SQL statements to pass to DBI.  Ideally, some sort of database abstraction should be used, but it is not always an option.

Below is an example block of code that handles a CGI form submission with an action of "Add" or "Update".  The SQL is built according to the action and passed to a single prepare and execute.

my $sql;
my @fields = qw(field1 field2 field3 field4 field5);
my @bind_values;
if ($input->{action} eq "Add") {
  $sql =
  "INSERT INTO datatable (" .
    join(", ", @fields) .
  ") VALUES (" .
    join(", ", map {"?"} @fields) .
  ")";
  map {push @bind_values, $input->{$_} } @fields;
}
elsif ($input->{action} eq "Update") {
  $sql =
    "UPDATE datatable SET " .
      join(", ", map {"$_ = ?"} @fields) .
    " WHERE primarykey = ?";
  map { push @bind_values, $input->{$_} } @fields;
  push @bind_values, $input->{primarykey};
}

if (length $sql) {
  my $sth = $dbh->prepare($sql);
  $sth->execute(@bind_values);
}

This is a "one-liner" that I use sometimes to find files that contain carriage returns.  Some files containing carriage returns were committed to a Subversion repository that I manage. These files caused the diff to be corrupt when diffing against a file without the \r.  Not a huge issue, but annoying.

This command basically uses a for loop with find to recursively list all files starting from the current directory.  In the loop, each file is dumped into a small Perl one-liner that checks each line for a \r and increments a counter.  The grep at the end filters out all files with a zero count.  You could change \r to anything you want to find.


This produces a list of files with a count of characters found.

Welcome to my new blog coderfoo.com.  My intention is to publish tips and code that I've found to make my life easier as a LAMP Developer, Linux/UNIX SA, and MySQL DBA.  Hopefully these posts will help you too.
743vidxfw6