dev@glassfish.java.net

Re: build broken?

From: Byron Nevins <Byron.Nevins_at_Sun.COM>
Date: Tue, 31 Oct 2006 10:44:28 -0800

jws: my building dir
ee: my install dir

I use a batch file on Windows, below. I do this first:
mv jws_prev junk1
mv ee_prev junk2
mv jws jws_prev
mv ee ee_prev

and then run the script.
If anything fails -- I go back to my last build like so:

rm -rf jws ee
mv jws_prev jws
mv ee_prev ee

With this technique, I always have a working build I can go back to...
-- 
=================================================
c:
cd \
if EXIST ee goto error
if EXIST jws goto error
md ee
md jws
cd \jws
del c:\temp\build.out
md m_jws
cd m_jws
cvs co appserv appserv-ee bootstrap
cd appserv-ee
call maven checkout bootstrap-all build configure-runtime >>  
c:\temp\build.out
cd \jws\m_jws
cvs co packager installer packager-ee installer-ee
cd installer-ee
call maven  -Dnative.build=true bootstrap build create-self-extract >>  
c:\temp\build.out
cd  \jws
cvs -d %CVSAS% co glassfish-samples
cvs -d %CVSAS% co glassfish/appserv-tests
goto end
:error
echo **** wipeout ee and jws first!!!  ****
:end
Lloyd L Chambers wrote:
> How do I do a build "from scratch"?
>
> I thought 'maven bootstrap-all checkout build' was a full build from 
> scratch.
>
> Thanks,
> Lloyd
>
> On Oct 27, 2006, at 4:56 PM, Byron Nevins wrote:
>
>> If you had done a build from scratch, you would have had no errors.
>> I never do incremental builds.  I do the full build in an empty 
>> directory, and then make changes locally and just build the module or 
>> 2 that I'm working on.  After a few days or so, after I've checked-in 
>> my changes, I start the process over again with a fresh new build...
>>
>>
>> Lloyd L Chambers wrote:
>>> I don't understand what you think I should do?
>>>
>>> On Oct 27, 2006, at 2:48 PM, Byron Nevins wrote:
>>>
>>>> IMHO, this technique is too fragile and (human) time-consuming.  
>>>> What I do is spend the computer's time instead doing a full build 
>>>> from scratch (< 1 hour).  Then I do edit-maven-test a million times 
>>>> from the module I'm working on.  I think it saves time & 
>>>> frustration overall because the full-build almost always goes 
>>>> through...
>>>>
>>>> Lloyd L Chambers wrote:
>>>>> I've done a 'maven clobber bootstrap-all' followed by 'maven 
>>>>> checkout' and I still get this (Mac OS X):
>>>>>
>>>>>     [java]
>>>>>     [java] compile-server:
>>>>>     [java]      [echo] Compiling com/sun/ejb, com/sun/ejb
>>>>>     [java]     [javac] Compiling 1 source file to 
>>>>> /gf-trunk/build/glassfish/appserv-core/build
>>>>>     [java]      [echo] Compiling com/sun/web, com/sun/jndi, 
>>>>> com/sun/enterprise
>>>>>     [java]     [javac] Compiling 528 source files to 
>>>>> /gf-trunk/build/glassfish/appserv-core/build
>>>>>     [java]     [javac] 
>>>>> /gf-trunk/build/glassfish/appserv-core/src/java/com/sun/enterprise/webservice/ClientPipeCreator.java:27: 
>>>>> cannot find symbol
>>>>>     [java]     [javac] symbol  : class ClientPipelineHook
>>>>>     [java]     [javac] location: package com.sun.xml.ws.assembler
>>>>>     [java]     [javac] import 
>>>>> com.sun.xml.ws.assembler.ClientPipelineHook;
>>>>>     [java]     [javac]                                 ^
>>>>>     [java]     [javac] 
>>>>> /gf-trunk/build/glassfish/appserv-core/src/java/com/sun/enterprise/webservice/ClientPipeCreator.java:37: 
>>>>> cannot find symbol
>>>>>     [java]     [javac] symbol: class ClientPipelineHook
>>>>>     [java]     [javac] public class ClientPipeCreator extends 
>>>>> ClientPipelineHook {
>>>>>     [java]     [javac]                                        ^
>>>>>     [java]     [javac] 
>>>>> /gf-trunk/build/glassfish/appserv-core/src/java/com/sun/enterprise/tools/verifier/VerifierResultsImpl.java:64: 
>>>>> warning: non-varargs call of varargs method with inexact argument 
>>>>> type for last parameter;
>>>>>     [java]     [javac] cast to java.lang.Object for a varargs call
>>>>>     [java]     [javac] cast to java.lang.Object[] for a 
>>>>> non-varargs call and to suppress this warning
>>>>>     [java]     [javac]         Object result = 
>>>>> failedCountMethod.invoke(resultObj,null);
>>>>>     [java]     [javac]
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>>>>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>>>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
>> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: dev-help_at_glassfish.dev.java.net
>