14:47 -!- multi_io [n=olaf@port-87-193-172-154.static.qsc.de] has joined ##java

15:56 < multi_io> the ant people rightly saw that Unix make has portability issues, then wrongly decided that XML was 
                  the best way to solve them
15:56 -!- hrehf [n=fsljif@dslb-092-073-082-069.pools.arcor-ip.net] has quit [Read error: 110 (Connection timed out)]
15:56 -!- hrehf_ is now known as hrehf
15:57 < multi_io> essentially, they wrongly identified make's scriptability as being the root of the portability issues
15:57 < multi_io> which is of course wrong. The underlying shell's portability issues are the root of the problem
15:58 < waz> nothing in makefiles format make them non crossplatform
15:58 < multi_io> so, the right way would have been to standardize the scripting layer, maybe by using something 
                  Java-based for it (BeanShell or jpython or something similar comes to mind)
15:58  * tazle doesn't understand "make's scriptability"
15:59 < tazle> it just runs shell commands, how is that "scriptable"?
15:59 < multi_io> Ant is essentially a wrong solution for a correctly identified problem :P
15:59 -!- robyonrails [n=robyonra@host51-121-dynamic.51-82-r.retail.telecomitalia.it] has quit ["Leaving"]
15:59 < bearded_oneder> multi_io ->  as well as the limitations of some unnamed platforms' shell environment (cough, 
                        cough)
15:59 < multi_io> yeah, and the shell commands aren't too portable
15:59 < multi_io> as well as the shell itself
15:59 < basicer> If Ant is the solution to your problem, and you like it: use it.  If you dont like it, or it is not 
                 the solution to your problem: use or create something else you like.
15:59 < multi_io> "it's easier to port a shell than a shell script" :-P
16:00 < waz> maven does the job for us
16:01 < RenatoSilva> bearded_oneder: isn't Java human readable?
16:01 < multi_io> tazle: "make's scriptability" = make's ability to easily write (shell) code in it (the code that 
                  executes the build tasks)
16:02 < multi_io> maven isn't scriptable enough imho
16:02 < multi_io> it's too declarative, as far as I can see
16:03 < waz> Java is a LOT more readable than xml
16:03 < multi_io> so.. maven is another wrong solution to a correctly identified problem :P
16:03 < bearded_oneder> RenatoSilva ->  still think it is ill advised to mingle the build executor (whether that be an 
                        executable, script, or java class) with the build descriptoe.
16:03 < waz> multi_io: I can build a massive code base across 28 project by typing mvn install
16:04 < waz> pontificate all you like, it works
16:04 < bearded_oneder> descriptor*
16:04 < waz> every time
16:04 < basicer> multi_io -> Id like to see an example build file (in ant) that has problems because its not scritable 
                 enough, and the proposed "java" syntax to compile that same ant script.
16:04 < defendguin> basicer: how can i easily tell then
16:06 -!- Absolute0 [n=ramin@ool-182e4b01.dyn.optonline.net] has joined ##java
16:06 < basicer> System.err.prntln(System.getProperty("java.runtime.version")); at the beggining of main should tell 
                 you
16:06 < multi_io> the right way would have been to standardize the scripting layer as said before, and then implement 
                  an equivalent to maven's automatic dependency resolver as a portable library on top of that 
                  scripting layer
16:06 < bearded_oneder> RenatoSilva ->  waz had a valid point that the problem with ant was choosing XML as the 
                        descriptor format, rather than the problem being  the separation of the descriptor from the 
                        executor.
16:09 < multi_io> basicer: it's just not hackable enough. As soon as you want to do something trivial that's not 
                  directly supported, like automatically generating some text file that'll be needed at runtime, you 
                  have to switch to Java and write a specific "Ant task" for that
16:10 < multi_io> that's the opposite of hackability