<span style=" font-size:10pt;font-family:sans-serif">Hi,</span>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">I tried to use
the IJG's contact page, but no joy. Seems broken; there a spinning icon
when you hit "send", but nothing happens.</span>
<br>
<br><a href=http://jpegclub.org/reference/contact/><span style=" font-size:10pt;color:blue;font-family:sans-serif">http://jpegclub.org/reference/contact/</span></a>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">So I used a slightly
older mailing list on sourceforge. The request to update their code has
been sent, and I hope it will appear on the mailing list soon.</span>
<br>
<br><a href="https://sourceforge.net/p/libjpeg/mailman/libjpeg-devel-6x/"><span style=" font-size:10pt;color:blue;font-family:sans-serif">https://sourceforge.net/p/libjpeg/mailman/libjpeg-devel-6x/</span></a>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">However, that
list seems fairly idle too.</span>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">Does anyone know
of another method we can use to communicate this fix upstream?</span>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">Best Regards</span>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">Adam</span>
<br>
<br><span style=" font-size:10pt;font-family:sans-serif">> Re: [OpenJDK
2D-Dev] RFR(xxxs): 8200052: libjavajpeg: Fix compile warning in jchuff.cPhilip
Race to: Adam Farley8 17/05/2018 03:32</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Cc: 2d-dev,
Andrew Leonard, build-dev, Magnus Ihse Bursie, "Thomas Stüfe"</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> From: Philip
Race <philip.race@oracle.com></span>
<br><span style=" font-size:10pt;font-family:sans-serif">> To: Adam
Farley8 <adam.farley@uk.ibm.com></span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Cc: 2d-dev
<2d-dev@openjdk.java.net>, Andrew Leonard <andrew_m_leonard@uk.ibm.com>,
build-dev <build-dev@openjdk.java.net>, Magnus Ihse Bursie <magnus.ihse.bursie@oracle.com>,
"Thomas Stüfe" <thomas.stuefe@gmail.com></span>
<br><span style=" font-size:10pt;font-family:sans-serif">>  </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Hi,</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> OK .. if
you can convince upstream this is worth doing, then we can accept it</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> as we would
not regress when updating. As I noted previously :</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span><a href="http://mail.openjdk.java.net/pipermail/2d-dev/2018-March/009086.html"><span style=" font-size:10pt;color:blue;font-family:sans-serif">http://mail.openjdk.java.net/pipermail/2d-dev/2018-March/009086.html</span></a>
<br><span style=" font-size:10pt;font-family:sans-serif">> this is still
an issue in the currently being developed 9c train.</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> -phil.</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> On 5/14/18,
3:06 AM, Adam Farley8 wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Hi Phil,
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Would an
acceptable compromise be to deliver the source code change </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> and send
the code to the upstream community, allowing them to include </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> the fix if/when
they are able? </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> I believe
Magnus was advocating this idea as well. See below. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Best Regards
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Adam Farley
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > Same
here. I would like to have this fix in, but do not want to go </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > over
Phils head. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > I think
Phil was the main objector, maybe he could reconsider?` </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > Thanks,
Thomas </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > On Thu,
Apr 26, 2018 at 10:39 AM, Magnus Ihse Bursie </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > <magnus.ihse.bursie@oracle.com>
wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
I don't object, but it's not build code so I don't have the final say.
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
/Magnus </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
On 2018-04-25 17:43, Adam Farley8 wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
Hi All, </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
Does anyone have an objection to pushing this tiny change in? </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
It doesn't break anything, it fixes a build break on two supported </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
platforms, and it seems </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
like we never refresh the code from upstream. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
- Adam </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
I also advocate the source code fix, as Make isn't meant to use the sort
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
of logic required </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
to properly analyse the toolchain version string. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
e.g. An "eq" match on 4.8.5 doesn't protect the user who is using
4.8.6, </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
and Make doesn't </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
seem to do substring stuff unless you mess around with shells. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Plus, as people have said, it's better to solve problem x (or work around
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
a specific </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
instance of x) than to ignore the exception, even if the ignoring code
is </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
toolchain- </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
specific. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
- Adam Farley </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> On 2018-03-27 18:44, Phil Race wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
>> As I said I prefer the make file change, since this is a change
to an </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
>> upstream library. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Newtons fourth law: For every reviewer, there's an equal and opposite
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> reviewer. :) </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Here I am, advocating a source code fix. As Thomas says, the compiler
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> complaint seems reasonable, and disabling it might cause us to miss
other </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> future errors. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Why can't we push the source code fix, and also submit it upstream?
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> /Magnus </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> I've looked at jpeg-9c and it still looks identical to what we have
in </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> 6b, so this code </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> seems to have stood the test of time. I'm also unclear why the compiler
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> would </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> complain about that and not the one a few lines later </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
>  819   while (bits[i] == 0)        
 /* find largest codelength still in </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> use */ </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
>  820     i--; </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> A push to jchuff.c will get blown away if/when we upgrade. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> A tool-chain specific fix in the makefile with an appropriate comment
is </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> more targeted. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Phil, </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Returning to this. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
While I understand your reluctance to patch upstream code, let me point
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
out that we have not updated libjpeg a single time since the JDK was open
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
sourced. We're using 6b from 27-Mar-1998. I had a look at the Wikipedia
page </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
on libjpeg, and this is the latest "uncontroversial" version
of the source. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Versions 7 and up have proprietary extensions, which in turn has resulted
in </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
multiple forks of libjpeg. As it stands, it seems unlikely that we will
ever </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
replace libjpeg 6b with a simple upgrade from upstream. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
I therefore maintain my position that a source code fix would be the best
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
way forward here. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
/Magnus </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> -phil. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> On 03/27/2018 05:44 AM, Thomas Stüfe wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Hi all, </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> just a friendly reminder. I would like to push this tiny fix because
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> tripping over this on our linux s390x builds is annoying (yes, we
can </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> maintain patch queues, but this is a constant error source). </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> I will wait for 24 more hours until a reaction. If no serious objections
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> are forcoming, I want to push it (tier1 tests ran thru, and me and
Christoph </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> langer are both Reviewers). </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Thanks! Thomas </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> On Wed, Mar 21, 2018 at 6:20 PM, Thomas Stüfe <thomas.stuefe@gmail.com>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> wrote: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Hi all, </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> may I please have another review for this really trivial change. It
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> fixes a gcc warning on s390 and ppc. Also, it is probably a good idea
to fix </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> this. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> bug: </span><a href="https://bugs.openjdk.java.net/browse/JDK-8200052"><span style=" font-size:10pt;color:blue;font-family:sans-serif">https://bugs.openjdk.java.net/browse/JDK-8200052</span></a><span style=" font-size:10pt;font-family:sans-serif">
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> webrev: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span><a href="http://cr.openjdk.java.net/~stuefe/webrevs/8200052-fix-warning-in-jchuff.c/webrev.00/webrev/"><span style=" font-size:10pt;color:blue;font-family:sans-serif">http://cr.openjdk.java.net/~stuefe/webrevs/8200052-fix-warning-in-jchuff.c/webrev.00/webrev/</span></a><span style=" font-size:10pt;font-family:sans-serif">
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> This was contributed by Adam Farley at IBM. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> I already reviewed this. I also test-built on zlinux and it works.
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> Thanks, Thomas </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Unless stated otherwise above: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
IBM United Kingdom Limited - Registered in England and Wales with number
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
741598. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >>
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
Unless stated otherwise above: </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
IBM United Kingdom Limited - Registered in England and Wales with number
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
741598. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> > >
</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Unless stated
otherwise above:</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> IBM United
Kingdom Limited - Registered in England and Wales with number 741598. </span>
<br><span style=" font-size:10pt;font-family:sans-serif">> Registered
office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU</span>
<br><span style=" font-size:10pt;font-family:sans-serif">> </span>
<br><span style=" font-size:10pt;font-family:sans-serif"><br>
Unless stated otherwise above:<br>
IBM United Kingdom Limited - Registered in England and Wales with number
741598. <br>
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU<br>
</span>