On Thu, Nov 4, 2010 at 12:42 PM, Rick Hillegas <[email protected]> wrote: > Myrna van Lunteren wrote: >> >> On Thu, Nov 4, 2010 at 11:12 AM, Rick Hillegas <[email protected]> >> wrote: >> >>> >>> Myrna van Lunteren wrote: >>> >>>> >>>> Hi, >>>> >>>> My main build on windows is broken after the latest updates. On my >>>> Linux machine the build is fine, but I don't have the j2me stuff >>>> there, probably some other stuff is missing. Perhaps I should've done >>>> clobber before svn update again, but I tried to clean everything up >>>> and start over... >>>> >>>> I'm still trying to see what's happening/happened/fixing it up... >>>> >>>> e.g.: >>>> compile_util: >>>> [javac] Compiling 8 source files to C:\derbyt\svn\trunk\classes >>>> [javac] >>>> >>>> C:\derbyt\svn\trunk\java\engine\org\apache\derby\iapi\types\SqlXmlUtil.java:40: >>>> package org.w3c.dom does not exist >>>> [javac] import org.w3c.dom.Attr; >>>> [javac] ^ >>>> >>>> and, more seriously: >>>> [javac] >>>> >>>> C:\derbyt\svn\trunk\java\engine\org\apache\derby\impl\jdbc\LOBStoredProcedure.java:27: >>>> cannot find symbol >>>> [javac] symbol : class DriverManager >>>> [javac] location: package java.sql >>>> [javac] import java.sql.DriverManager; >>>> [javac] ^ >>>> >>>> >>>> >>>> Myrna >>>> >>>> >>>> >>> >>> I am seeing pages of these build errors as well, when I do "ant clobber" >>> followed by "ant build". I tried backing out my recent change to the >>> build >>> (1031069, see DERBY-4845), but that did not fix the problem. I think that >>> someone else's changes have caused this. >>> >>> Regards, >>> -Rick >>> >>> >> >> I was doing "ant all". I noticed (will do some more checking) that the >> classes it's complaining about seem to get created at some point >> anyway, and if I run ant all a second time, there are no errors... >> >> Myrna >> >> > > I have checked in a patch which fixes the build for me at least. The commit > stamp is 1031174 and the explanation of the fix can be found on DERBY-4845. > I believe that some recent submission introduced a dependency of the > ...iapi.util package on ...iapi.types. Probably that change was not tested > by a full "ant clobber" followed by "ant all". > > Hope this helps and doesn't break someone else's build in another way... > > Regards, > -Rick >
Last 2 pieces of info: - the build failed when I had the property jsr169compile.classpath set; it passed if it was not set (silly not to try this first, as it was the one thing I knew was different between my linux and windows build). - after those latest changes there are no further failures with or without the jsr169compile.classpath property set. Thanks Rick!
