Re: RFC: Proposal to halt automatic builds of Windows BaseTools executables

["Tim Lewis" <tim.lewis@insyde.com>]

Laszlo, Erik --

I understand this dislike from some open source developers. I respect that and am glad that EDK2 provides a way to accommodate this preference. But "most" is a strong term. I would venture to say that a good number (and probably the majority) of the people using EDK2-derived code and tools are fine with the current situation.

The only reason I would have Python on most of my company's dev systems would be for EDK2. Since (a) the current system is working and (b) since the possibility for rebuild is available for those who want it, it doesn't weigh heavily enough IMO to change the current situation. 

Regards,

Tim

-----Original Message-----
From: Laszlo Ersek <lersek@redhat.com> 
Sent: Thursday, March 8, 2018 1:19 PM
To: Tim Lewis <tim.lewis@insyde.com>; 'Bjorge, Erik C' <erik.c.bjorge@intel.com>; edk2-devel@lists.01.org
Subject: Re: [edk2] RFC: Proposal to halt automatic builds of Windows BaseTools executables

On 03/08/18 19:05, Tim Lewis wrote:
> Erik --
> 
> What is the justification? Moving from more immediately usable to less 
> immediately usable doesn't seem, on the surface, to be  a good direction.
> Why not go the other direction and pre-build the binaries for the 
> other environments?

I'd just like to offer one data point for the last question: most open source developers *really* dislike running any native binaries that were built by neither (a) themselves nor (b) the provider of their OS distribution.

To give you an example for (b), Fedora provides the "edk2-tools" package (built from the "edk2" source package), and "edk2-tools" definitely installs native binaries:

https://koji.fedoraproject.org/koji/rpminfo?rpmID=13354362

The difference is that these binaries were built in a build environment that matches the rest of Fedora [*] and is generally trusted by Fedora users.

[*] For example, binaries could be instrumented for security purposes system-wide; some buffer overflows in a native (C) application could be caught automatically as a result.

Thanks,
Laszlo