RE: "Official" binary release for Windows

[John D]

-----Original Message-----
From: John W. Eaton [mailto:address@hidden 
Sent: Tuesday, March 18, 2014 2:26 PM
To: John Donoghue
Cc: address@hidden
Subject: Re: "Official" binary release for Windows

On 03/07/2014 07:10 PM, John Donoghue wrote:
> On 03/07/2014 06:08 PM, address@hidden wrote:
>> Message: 3
>> Date: Fri, 07 Mar 2014 17:21:01 -0500
>> From: "John W. Eaton"<address@hidden>
>> To: octave maintainers mailing list<address@hidden>
>> Subject: "Official" binary release for Windows 
>> Message-ID:<address@hidden>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>
>> I'd like to start making the Windows installer available from 
>> ftp.gnu.org.  Is it ready for that now?  Or are there some 
>> significant problems that need to be fixed before we should make
"official" releases?
>>
>> jwe
>
> I havent had any issues recently with the build except with gcc.
>
> I had added a readme html file for the install that may want to be 
> looked at / modified prior to distribution, depending on what is 
> distributed or not.
>
> Would it also make sense to list in it all the packages that are known 
> to work and compile in windows ?

Sure.

Is there any good reason that openblas should not be used by default?

Are there any significant issues remaining?  Is it time to make the
3.8.1 binary available?

jwe

---

For openblas: No reason that I know of, I had just set the default blas to
the reference in the NSIS script since I knew that the refwerence would
always be available, but it could be changed to default to openblas if
available on creating the installer.

For significant issues: Theres a few minor ones hanging around, but I don't
know of any major issues preventing a binary. 

Building of native gcc for some reason doesn't work without patching on some
linux systems, but if it works for your binary build, that shouldn't prevent
a distribution going ahead.

JD