References: <56C41C5B.9000809@lucee.org> <56C4D516.9010403@hogranch.com>
 <56C4D8C4.4010809@anastigmatix.net> <56C50E87.4070008@lucee.org>
 <56C51150.1070708@anastigmatix.net> <56C54EE0.2080905@lucee.org>
 <56C553C3.40408@lucee.org> <56C556BD.1010907@anastigmatix.net>
 <56C55824.1060502@lucee.org> <56C564F4.1050909@anastigmatix.net>
 <56C5E1B3.8090204@lucee.org> <56C5E795.4080405@anastigmatix.net>
 <56C5F11F.9050709@lucee.org> <56C5F47C.2070906@anastigmatix.net>
 <56C5FCEB.4000803@lucee.org>
From: Chapman Flack <chap@anastigmatix.net>
Message-ID: <56C61189.4040807@anastigmatix.net>
Date: Thu, 18 Feb 2016 13:46:33 -0500
MIME-Version: 1.0
In-Reply-To: <56C5FCEB.4000803@lucee.org>
Subject: Re: [Pljava-dev] Build Fails on Windows 7
Content-Type: text/plain; charset="us-ascii"

On 02/18/2016 12:18 PM, Igal @ Lucee.org wrote:

>> I guess the printf format warnings are about this:
>> https://sourceforge.net/p/mingw-w64/mailman/message/24633511/
>>
>> ugh...
> Any thoughts what the implications of that would be?
> Arithmetic overflows?  Unexpected results?

Possibly bogus values appearing in messages printed from those
format strings (which I think only appear at debug levels anyway).

The most common solution I've seen other places is to define
macros for the correct printf formats that go with each type,
and ANSI-cpp-concat them into the format strings. Which is
clever but not very compatible with message localization.
So far there hasn't ever been any localization work on PL/Java,
but that doesn't mean I'm eager to make it harder....

Maybe it's better to just always use the printf format for the
widest known type, and widen-cast every corresponding argument.
That would be annoying but not get in the way of localization.
Of course that leads to the question, will every platform that
might want PL/Java have a printf that likes %lld?

The other warnings I saw were innocuous. Apparently PL/Java
just doesn't use the statementcancel, die, or quickdie handlers
when built on Windows. There is probably some sensible Windows
reason for that and I don't know what it is, but that #ifdef was
added in 2005 so if your build from today doesn't work, I don't
think that will be the reason. ;)

It's looking like you ought to start this thing up and take it
for a spin. If I were you I would initdb a new play database
for it before trying it out in, you know, payroll, but you
can't put off trying it forever. :)

-Chap
_______________________________________________
Pljava-dev mailing list
Pljava-dev@lists.pgfoundry.org
http://lists.pgfoundry.org/mailman/listinfo/pljava-dev