DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11429>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11429 lcp.bat produces error: "The input line is too long." when directory structure is deep. ------- Additional Comments From [EMAIL PROTECTED] 2002-08-14 14:40 ------- I don't have a lot of time to spend on this topic, so I would appreciate it if you would please carefully re-read my prior comments. I think you'll find the answers to your questions there. >I don't see how to replicate common sense about OS limitation that we don't >know will help in any way. I assume this refers to my request for documentation, at a minimum. >> By publishing the implicit limitations of a product, you help >> encourage "design >> by contract" -- don't let your application give the impression it can do >> something it can't. What you call common sense, in this case, I call an implicit limitation. It is not common sense for me, because if I use invocation method #2, the application works fine. The problem is not with the OS, it is with ant.bat. >>... please note that there are two different invocation >>methods for ant on w2k -- 1) using ant.bat, or 2) using java invocation (with >>the proper classpath, of course) >>It is possible to run ant from a "deep" directory using method #2 -- I do it >>every day by necessity. > Command line length is different even among some > unix. The classpath does not need to be specified in the commandline or in a system variable. It can be specified in a file, too. Please read the documentation for java.exe. >> There are many ways to build up a java classpath. lcp.bat is not the only >> one. Perhaps by substituting lcp.bat with another mechanism, you will >> resolve this problem. > For me it is the same as requesting a warning in the docs about not being > able to store a 128bit integer into a 32bit one...kind of weird. You are entitled to your opinion. This is in conflict with my understanding of the Java Platform on W2k. The programs I write do not use lcp.bat. They do not have this defect. I see nothing that relates the function of lcp.bat to programming language fundamentals like memory allocation. > Seriously, assuming I can fix this to build classpath in relative, you will > come back to me because the input line is too long when you have 50000 jars > in > your classpath... It is possible to have a classpath of any length. Please read the documentation for java.exe, and see my comments above. > you will > come back to me because the input line is too long when you have 50000 jars > in your classpath... I will not do this. I promise. I am a real programmer trying to do real work. really. > Do you think it is worth wasting time to replicate the > documentation of your operating system to the java manual because you want to > encourage 'design by contract' so you don't want to read the contract of the > os you're running so you want another full contract ? Nothing is worth wasting time. Life is too short. It is not wasting time to provide a useful piece of information to the documentation. In the time we have spent writing these comments, this could have been done several times. IMO a user should *never* have to read the source code to understand why a given program errs. We are not talking about a segmentation fault or a memory leak here. This is a problem with a known and finite set of causes. I'm not trying to give anyone a hard time here. ANT is very likely the most useful, versatile, and creative program I have seen in quite a while. This is Open Source - fix the defect / don't fix it / ignore me / do what you'd like. I am honestly just glad that anyone is actually addressing this issue in any fashion. Thanks again. -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
