Login | Register
My pages Projects Community openCollabNet

Discussions > dev > Re: [sourcecast-dev] Tonight's Build BROKEN

Project highlights: Architectural Overview

joist
Discussion topic

Hide all messages in topic

All messages in topic

Re: [joist-dev] Re: [sourcecast-dev] Tonight's Build BROKEN

Author dlr
Full name Daniel Rall
Date 2000-10-21 19:01:54 PDT
Message "Daniel L. Rall" wrote:
>
> Ed Korthof wrote:
> >
> > Looking at the code, I think javac is incorrect in throwing an error --
> > even more so because it will subsequently compile the file without
> > complaint.
>
> I totally agree that javac does the wrong thing. I wonder if this
> problem exists in the javac that comes with the IBM JDK?

I meant to refer to version 1.3.
--

Daniel Rall <dlr at finemaltcoding dot com>

Re: [joist-dev] Re: [sourcecast-dev] Tonight's Build BROKEN

Author dlr
Full name Daniel Rall
Date 2000-10-21 19:00:49 PDT
Message Ed Korthof wrote:
>
> Looking at the code, I think javac is incorrect in throwing an error --
> even more so because it will subsequently compile the file without
> complaint.

I totally agree that javac does the wrong thing. I wonder if this
problem exists in the javac that comes with the IBM JDK?
--

Daniel Rall <dlr at finemaltcoding dot com>

Re: [sourcecast-dev] Tonight's Build BROKEN

Author edk
Full name Ed Korthof
Date 2000-10-21 17:31:29 PDT
Message On Sat, Oct 21, 2000 at 11:26:08AM -0700, Daniel L. Rall wrote:
> We now have build under javac (there was never a problem under jikes)!
> The problem was that jikes and javac handle initialization of final
> variables differently (I switched a few static final variables to just
> static). Most of the developers use jikes to compile, so we weren't
> experiencing this problem (though both Elicia and Ed did reported a
> couple cases of it, for some reason we weren't able to reproduce it at
> the time--standardization of the lower level Makefiles would help a lot
> with this). Thanks.

I was able to reproduce the problem reliably. However -- if you compile
a second time, the error goes away. That's right -- on the first
compile, with 118 files, it fails; on the second, with just 1, it works.

Looking at the code, I think javac is incorrect in throwing an error --
even more so because it will subsequently compile the file without
complaint.

I described these issues on irc; but when I found that jikes doesn't
care, and javac only cares the first time around, I decided that I
didn't want to spend the time trying to figure out what javac's problem
(or bug) with that is, so I just started setting JIKES=yes.

It's quite ugly, but running make twice in helm will "solve" this, as
will using jikes.

cheers --

Ed
--
   +=-=+=-=+=-=+=-=+=-=​+=-=+=-=+=-=+=-=+=-=​+=-=+=-=+=-=+=-=
   | Ed Korthof | edk at collab dot net | 415-247-1690 |
   +=-=+=-=+=-=+=-=+=-=​+=-=+=-=+=-=+=-=+=-=​+=-=+=-=+=-=+=-=

Re: [sourcecast-dev] Tonight's Build BROKEN

Author dlr
Full name Daniel Rall
Date 2000-10-21 11:26:08 PDT
Message We now have build under javac (there was never a problem under jikes)!
The problem was that jikes and javac handle initialization of final
variables differently (I switched a few static final variables to just
static). Most of the developers use jikes to compile, so we weren't
experiencing this problem (though both Elicia and Ed did reported a
couple cases of it, for some reason we weren't able to reproduce it at
the time--standardization of the lower level Makefiles would help a lot
with this). Thanks.

Josh Lucas wrote:
>
> I am getting compile errors in the newhelm directory tonight. The errors
> are shown below.
>
> I realize that we are all under pressure to get things done but if we are
> going to do nightly builds then build MUST compile. If you are working on
> code close to the 9pm deadline then PLEASE make sure that your code will
> compile and not just within your own sandbox. Instead, commit your code
> and do a clean checkout and try to compile. That way you can see your
> errors and I won't.
>
> If this gets fixed over the weekend then I will do a build. If not, then
> the build won't happen until Monday.
>
> Let me know if you have any questions.
>
> josh
>
> Errors:
>
> /home/lucas/sandbox/​newhelm/dest/src/org​/tigris/helm/databas​e/ProjectManager.jav​a:225:
> Blank final variable 'HOST_PROJECT_ID' may not have been initialized. It
> must be assigned a value in an initializer, or in every constructor.
> static
> ^
> /home/lucas/sandbox/​newhelm/dest/src/org​/tigris/helm/databas​e/ProjectManager.jav​a:225:
> Blank final variable 'PROJECT_ID_BY_DOMAIN' may not have been initialized.
> It must be assigned a value in an initializer, or in every constructor.
> static
> ^
> /home/lucas/sandbox/​newhelm/dest/src/org​/tigris/helm/databas​e/ProjectManager.jav​a:225:
> Blank final variable 'PUBLIC_PROJECT_ID_BY_DOMAIN' may not have been
> initialized. It must be assigned a value in an initializer, or in every
> constructor.
> static
> ^
> 3 errors
>
> BUILD FAILED

--

Daniel Rall <dlr at finemaltcoding dot com>
Messages per page: