ant-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Duffey, Kevin" <>
Subject RE: Trying to get rid of makefiles
Date Thu, 24 Jul 2003 00:10:39 GMT
First, find another job! My God that is ridiculous. I can't stand the resistance to change.
I mean, I am all for watching a product gain a little maturity, but gees, Ant is the defacto
standard. Who the hell heard of building java code with makefiles anyway? Do you guys do any
packaging into jar files, ear files, war files, zip files? First, I strongly recommend getting
the book Java Development with Ant. It has proved very helpful in coming up with a very powerful
template script to do just about anything. I am not even 1/2 way through the book and already
I have learned a lot. My template script does various tasks, including clean builds, development
builds, runs unit tests and generates test reports, packages into jar and zips, ear/wars,
pulls "shared" files from various dirs into the zip/jar files in specific locations I want
them in, filters various text files for specific inclusion/exclusion as I need it, executes
my application during development so I don't need a separate batch file, does QA/nightly builds,
integration testing and more and more. And the script is like 300 lines long and I re-use
the script almost as is for every project.

Anyway, I feel really bad that you are in such an environment. It is sad to see such stunted
creativity and such closed mindedness. It is often a fear of loss of control that stimulates
those unwilling to change to do as they do. Either that, or lack of understanding and thus
unwilling to give in to change.

What you talk about as your last issue, can you provide ANY info as to why they care that
deleting .class files is a problem? One of my build scripts deletes over 2000 files from various
directories, runs on a very slow machine and it takes less than 2 minutes for a full clean
nightly build. During development, yeah, it would be a nightmare waiting that long. But for
a clean build, that is what is involved. When you set it up for a nightly build or when a
developer is done for the day and wants to do a clean build to make sure they haven't broken
anything, what is the big deal? Is your build taking hours on clean?

Are you guys running junit and jfcunit tests? If not, show them that one! Tell them they need
to update their development habits... on second thought, better not, they may fire you!

-----Original Message-----
From: Leonardo Abreu de Barros [] 
Sent: Wednesday, July 23, 2003 4:02 PM
Subject: Trying to get rid of makefiles

Hi all,

I work in a company that has a long history using makefiles. Since I 
joined it, I proposed to migrate to Ant. Some people resist to 
changes, some people agree, as long as all benefits provided by the 
makefiles are covered by Ant.

I've already reproduced all behaviors but one: currently, all java 
classes names are declared explicitly in the makefile, and sent to 
the java compiler. It rebuilds all the class files, at each build. 

They want this behavior in order to detect broken code, due to 
changes on a base class, for example.

The problem is that Ant, specifically the <javac> task, only 
recompiles ".class" if the timestamp differs from the related ".java" 
file. I've read on documentation that to detect this kind of broken 
dependencies, you should perform "clean builds" from time to time.

I proposed this solution, but it wasn't accepted. They think it's a 
loss of time (even if it's ridiculous) to delete all ".class" files 
before calling the compiler. If I'm not able to reproduce this 
behavior with Ant, they prefer to keep the makefiles.

Does anyone have any idea of how I could reproduce this behavior 
using Ant? Any other possible solution?

Thanks in advance,
Leonardo Barros

To unsubscribe, e-mail:
For additional commands, e-mail:

Incoming mail is certified Virus Free.
Checked by AVG anti-virus system (
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003

Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (
Version: 6.0.501 / Virus Database: 299 - Release Date: 7/14/2003

"The information contained in this e-mail message  may be confidential and 
protected from disclosure.  If you are not the intended recipient, any 
dissemination, distribution or copying is strictly prohibited.  If you think 
that you have received this e-mail message in error, please e-mail the 
sender at"

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message