Comments on: Use a classpath resource or kill your application’s portability http://blog.markturansky.com/archives/141 software architecture & engineering, code hints, sometimes philosophy, photography, life, etc. Sun, 20 Jun 2010 07:51:38 -0400 http://wordpress.org/?v=2.8.6 hourly 1 By: skelter http://blog.markturansky.com/archives/141/comment-page-1#comment-1149 skelter Fri, 17 Jul 2009 05:02:42 +0000 http://blog.markturansky.com/archives/141#comment-1149 I recommend pulling from a config path set by a system property with a reasonable default. We have configuration set in resource bundles and it has been painful. Worse, somewhere along the way, multiple 'systems' were added to pull from one resource bundle or another based on machine name or user name. As we struggle to figure out which configuration bundle is the one in effect, changing it is slowed by having to go through the compile phase first. Configuration should be external. It looks like the solution we are settling on is a relative path in a reasonable default root that can be adjusted via a system property. This makes it easy to change a CI build which may be running multiple branches (and configurations) on the same machine on the same user account. So..have you encountered collisions in the java.io.temp dir yet? -s I recommend pulling from a config path set by a system property with a reasonable default.

We have configuration set in resource bundles and it has been painful. Worse, somewhere along the way, multiple ’systems’ were added to pull from one resource bundle or another based on machine name or user name. As we struggle to figure out which configuration bundle is the one in effect, changing it is slowed by having to go through the compile phase first.

Configuration should be external.

It looks like the solution we are settling on is a relative path in a reasonable default root that can be adjusted via a system property. This makes it easy to change a CI build which may be running multiple branches (and configurations) on the same machine on the same user account.

So..have you encountered collisions in the java.io.temp dir yet?

-s

]]> By: Mark Turansky http://blog.markturansky.com/archives/141/comment-page-1#comment-1144 Mark Turansky Fri, 10 Jul 2009 21:58:57 +0000 http://blog.markturansky.com/archives/141#comment-1144 That's true, Matt. I also like the forward slash trick because it works really well on Windows. I used it when I wrote an example about <a href="http://blog.markturansky.com/archives/21" rel="nofollow">Java bootstrapping</a> where the example looks for "/usr/local/src." This defaults to "C:/usr/local/src" on Windows. That’s true, Matt. I also like the forward slash trick because it works really well on Windows. I used it when I wrote an example about Java bootstrapping where the example looks for “/usr/local/src.” This defaults to “C:/usr/local/src” on Windows.

]]> By: Matt Passell http://blog.markturansky.com/archives/141/comment-page-1#comment-1143 Matt Passell Fri, 10 Jul 2009 19:48:00 +0000 http://blog.markturansky.com/archives/141#comment-1143 Even though they're not as good as your solution, relative paths are still more portable than absolute ones. From a platform-portability standpoint (which I know wasn't what you were covering), I've tried to encourage people to use forward-slashes everywhere, since Java is fine with forward-slash paths on Windows but not with backward-slash paths on Unix/Linux. You probably knew that already, but perhaps someone else reading this didn't. --Matt Even though they’re not as good as your solution, relative paths are still more portable than absolute ones. From a platform-portability standpoint (which I know wasn’t what you were covering), I’ve tried to encourage people to use forward-slashes everywhere, since Java is fine with forward-slash paths on Windows but not with backward-slash paths on Unix/Linux. You probably knew that already, but perhaps someone else reading this didn’t.

–Matt

]]>