In article <[EMAIL PROTECTED]>, Paul Sander wrote:
>>--- Forwarded mail from [EMAIL PROTECTED]
>
>>>>>>> "Paul" == Paul Sander <[EMAIL PROTECTED]> writes:
>>Paul>
>>Paul> Unfortunately, if this is what your build procedure consists of,
>
>>Don't be silly. We have our own make tool (written in Java in fact) that
>>enforces various packaging layers during designer compilation and
>>loadbuild.
>
>>Paul> then you lose traceability between your sources and shippables, and
>>Paul> you can't assess the impact of any change you make to your source
>>Paul> code. That makes it really really hard to accomplish the common task
>>Paul> of shipping minimal patches when bugs are found in the product.
>
>>I believe that you would be insane to handle Java source in the fashion
>>described in my previous posting. My point was, however, that the language
>>does not *force* you to keep the source in some sort of coherent order so
>>that it's incorrect for people (including me) to claim that it does.
>
>I'm glad that we're in agreement. I do know some Java programmers who
>literally do use "javac `find . -name '*.java' -print`" as their build
>procedures, so I assume the worst when somebody mentions the practice.
That's terrible! What if the resulting command line violates the
systems's environment passing limit? Of course, you want:
find . -name '*.java' -print | xargs javac
hopefully, none of the names contain spaces and newlines, but I wouldn't
put any such stupidity past Java programmers, so better use GNU tools:
find . -name '*.java' -print0 | xargs -0 javac
--
Meta-CVS: version control with directory structure versioning over top of CVS.
http://users.footprints.net/~kaz/mcvs.html
_______________________________________________
Info-cvs mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/info-cvs
