Synchronizing two CVS repositories using Mercurial

We're in a situation where we have access to some of our partner's source code. We need to keep in sync with their repo, otherwise our copy would quickly become stale.

Currently, we can do this manually because our changes are still small. A few months from now, however, our changes will eventually drift away from their code and merging it with our tree would become difficult.

A distributed version control system is in order. Here's how I did it:

+-----------------+         +------------+                 +------------+         +----------------+
| Upstream        |         | Immutable  |                 | Working    |         | Local          |
| CVS repository  |-------->| Mercurial  |---------------->| Mercurial  |-------->| CVS repository |
+-----------------+ import  | repository | Mercurial pull  | repository | export  +----------------+
                           +------------+                 +------------+
                                                               |^
                                                               ||
                                                               ||
                                                               v|
                                                              User

In theory, exporting it back to CVS is no longer needed. In practice, however, I needed to maintain "compatibility" with my fellow developers who only use CVS. I hope we'll eventually migrate away from CVS though because merging lots of files is really time consuming. Merging/branching is CVS' main weakness.

MySQL: How to merge two databases

I'm currently managing a couple of databases in a production server. The problem: I need to know what new tables were created from Database 1 and merge them to Database 2.

Here's the solution:

0. Backup the target database.
% mysqldump -h [HOST] -P [PORT] -u [USERNAME] -p[PASSWORD] Database2 > Database2.sql

1. Extract the table names from the source and target databases.
% echo show tables | mysql -h [HOST] -P [PORT] -u [USERNAME] -p[PASSWORD] Database1 | sed '1d' > one.tables
% echo show tables | mysql -h [HOST] -P [PORT] -u [USERNAME] -p[PASSWORD] Database2 | sed '1d' > two.tables

2. Perform the merge.
% mysqldump -h [HOST] -P [PORT] -u [USERNAME] -p[PASSWORD] Database1 \
`diff --unchanged-line-format='' --new-line-format='%l ' two.tables one.tables` | \
mysql [HOST] -P [PORT] -u [USERNAME] -p[PASSWORD] Database2

Unix/Linux Tip: Syntax check your PHP files

If you wish to syntax check (lint) all the PHP files in your current directory, simply do:
    for i in `ls *.php`; do
        php -l $i
    done

Fedora/Redhat Linux: C/C++ : unfound library "mysql.h"

If you encounter this problem while compiling an application that links to MySQL, you can fix this by simply installing mysql-dev:
    # /etc/init.d/mysqld stop
    # yum install mysql-dev

FreeBSD Tip: How to delete all IPv6 address in an interface

You can delete all IPv6 addresses in an interface using the following command (vanilla Bourne shell):

while i="`ifconfig le0 | grep inet6 | grep -m 1 -v '%'`"; do
    ifconfig le0 $i delete
done

Note: replace le0 with your interface name.

Bridge problem with FreeBSD 7.0-RELEASE

Problem
The FreeBSD Handbook states that "if the bridge host needs an IP address, then the correct place to set this is on the bridge interface itself rather than one of the member interfaces."

With FreeBSD 6.x and FreeBSD 8.0-CURRENT, setting the IP address on the bridge works without any problem. With FreeBSD 7.0-RELEASE, however, this does not work. The OS doesn't even properly boot when the IP address is set via /etc/rc.conf.

Solution
This is probably a temporary solution, but try setting the IP address to one of the member interface. This is in direct opposition to the advice given in the Handbook, but it works for me.

If this is a bug (which I suspect it is), a patch probably exists somewhere. Please feel free to leave a comment.

JPEG support for ImageMagick

Problem
On Debian Linux, there seems to be no out-of-the-box support for JPEG in ImageMagick.

Solution
First, install libjpeg62-dev library. libjpeg62 doesn't work because it doesn't include the necessary development files. Next, download ImageMagick's source and compile it manually. Use ./configure --with-jpeg=yes for configuring.

How to Show All CVS Tags

Here is a useful Perl script for displaying all tags in a CVS Repository.

OpenLaszlo IDEforLaszlo Eclipse plugin

Syte GMBH has provided an update for IDEforLaszlo plugin to support OpenLaszlo 4.0.10 and OpenLaszlo 3.4. IDEforLaszlo has been stale for quite some time now, and I'm glad that someone has offered their time and effort for this project. I'll be checking it out later.

IDEforLaszlo can be found at http://www.syte.ch/en/laszlo.xml.

Technical Humor

Yeesh, I found the links for the April 1 RFCs. Here's a dose of highly technical humor:

• RFC 5241 — Naming Rights in IETF Protocols. A. Falk, S. Bradner.
• RFC 5242 — A Generalized Unified Character Code: Western European and CJK Sections. J. Klensin, H. Alvestrand.
  

WARNING: Reading the above documents may result in nosebleed.

Morph eXchange

Our friends from Morph Labs has just launched version 2.0 of their Morph eXchange website. Congratulations, butterfly!

April 1, 2008 RFC

It's April 1 already, but I can't seem to find prank RFCs for this year...

IPv4/v6 Regular Expressions

I found these useful regexes in my notes. I probably found this somewhere on the Internet, but I forgot where.

IPv4 address

^(?:(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.){3}(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)$

IPv6 address

^[0-9a-fA-F]{1,4}(\:[0-9a-fA-F]{1,4}){7}$

Note that this does not work with shortened form such as dead::beef. dead:0000:0000:0000:0000:0000:0000:beef must be used instead.

dnsmasq: unknown interface xennet0

As of version 2.41, dnsmasq doesn't support Xen virtual network interfaces in NetBSD. This means that running dnsmasq inside a Xen domU results in the following error:

dnsmasq: unknown interface xennet0

Uberlord made a patch just a few moments ago:
- http://roy.marples.name/~roy/dnsmasq-netbsd.patch
- http://roy.marples.name/~roy/dnsmasq-bridge.patch

The beauty of Open Source... =)

Permission denied to call method XMLHttpRequest.open

Problem
When deploying DHTML OpenLaszlo applications in SOLO mode, attempts to fetch datasets results in the following error:

ERROR: uncaught exception: Permission denied to call method XMLHttpRequest.openLFCdhtml-debug.js (line 1421)
uncaught exception: Permission denied to call method XMLHttpRequest.open

Solution
As a security measure, most browsers restrict XMLHttpRequest such that they don't accept XML data if the HTTP response headers are not properly set. To fix this, simply change Content-Type to:
Content-Type: application/xml; charset=UTF-8

xen Error: Device 2050 (vbd) could not be connected. Backend device not found.

Problem
NetBSD dom0 is unable to launch more domUs. Attempting to launch more domUs results in the following error:

xen Error: Device 2050 (vbd) could not be connected. Backend device not found.

Solution
The system ran out of loop devices. Create more vnode disks using the following command:

/dev/MAKEDEV vnd1
/dev/MAKEDEV vnd2
...and so on

__LZgetNodes: p is null in Datapath

I'm creating an IPv4/IPv6 OpenLaszlo widget that can bind to a datapath. I spent hours trying to find out why this.datapath.getNodeText() doesn't work inside <handler name="oninit">, although it works fine inside <method>. I kept getting an error of "p is null in Datapath."

For developers experiencing a similar problem, you might want to check out Laszlonia's entry on his blog. It turns out that I simply had to use <handler name="ondata">.

map methods execute*() methods in Agavi

Agavi is architecturally beautiful. I can really appreciate the way they designed this thing. One fundamental problem with this framework, however, is that it lacks good documentation. You need to read the code, or ask others who already did, for you to answer some very simple questions.

In my case, I was looking for a way to change the HTTP verb mappings with AgaviWebRequest methods.

Agavi maps executeCreate() with HTTP PUT and executeWrite() with HTTP POST. If I use these functions in the context of REST architecture, it makes more sense if they are mapped the other way around. Although I understand that there is no single way to implement REST, there seems to be a general agreement that update operations should be mapped to HTTP PUT while create operations should be mapped to HTTP POST.

Fortunately, Agavi provides a simple (undocumented) way to change the mappings. In config/factories.xml, simply add:

              <request class="AgaviWebRequest">
                  <parameter name="method_names">
                      <parameter name="POST">create<parameter>
                      <parameter name="GET">read<parameter>
                      <parameter name="PUT">write<parameter>
                      <parameter name="DELETE">delete<parameter>
                  <parameter>
              <request>

Thanks to the crazy guy, MikeSeth. =)

confused on implementing REST using HTTP PUT

I'm implementing a REST web service using PHP5 right now. I'm a bit confused on how to implement a write operation using HTTP PUT verb.

In PHP, data that are sent via HTTP POST are in the following format:

variable1=data&variable2=data

The data is immediately available via $_POST. Data that is sent via HTTP PUT is uploaded as a file to the server and must be parsed before reading. There's nothing wrong with it per se, except that writing data is inconsistent. There's no problem requiring my REST clients to upload a file using HTTP PUT, except that it's inconsistent.

So far, I used curl for testing my web app. I haven't tried sending data via <form> using a web browser, and I don't know whether <form method="put"> works. If it does, then the browser would probably send the data similar to POST's format, with only a different header (UPDATE - 2008/03/11: yepp, it indeed works that way). Blah, possibilities, possibilities.

I'm using Agavi framework, by the way. It's a kick-ass lightweight framework. Agavi is beautiful, but the documentation is sparse. Documentation effort is underway, and I hope to see more of it in the coming weeks. Without it, Agavi is considered "fringe" at the moment (I like fringe software; I even use fringe operating systems such as *BSD hehe).

cvsmode in csup

Nice, there are now patches for CVSMode support in csup: http://people.freebsd.org/~lulf/patches/csup/cvsmode/. This means that csup can now fetch complete CVS repositories. Thank you Ulf, whoever you are!