status check on PEP 517
Nick: Do you have any update on when the PEP will be completed? Regards, xoviat
On 15 August 2017 at 15:11, 12345 67890 <xoviat@gmail.com> wrote:
Do you have any update on when the PEP will be completed?
From my perspective, the main item from the last round of discussions that has yet to be explicitly added to the PEP is the ability for build_sdist and build_wheel to raise NotImplementedError to indicate that an operation has failed for aniticipated reasons and to provide an error message explaining the problem.
There were a couple of other items where the PEP authors weren't necessarily sure that what was in the PEP was exactly what they wanted to propose implementing, but my own view on those aspects is that what's currently in the PEP is fine, as the last round of discussions cleaned up some key problems where the way the explicitly out-of-tree build support was designed in earlier versions of the API didn't align with the way most build systems actually work (i.e. they expect to be given the target directory as a build setting, allowing the build system itself to decide exactly how to handle that). Cheers, Nick. -- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia
I'm also willing to bet that Thomas is busy prepping for JupyterCon. On Tue, 15 Aug 2017 at 03:14 Nick Coghlan <ncoghlan@gmail.com> wrote:
On 15 August 2017 at 15:11, 12345 67890 <xoviat@gmail.com> wrote:
Do you have any update on when the PEP will be completed?
From my perspective, the main item from the last round of discussions that has yet to be explicitly added to the PEP is the ability for build_sdist and build_wheel to raise NotImplementedError to indicate that an operation has failed for aniticipated reasons and to provide an error message explaining the problem.
There were a couple of other items where the PEP authors weren't necessarily sure that what was in the PEP was exactly what they wanted to propose implementing, but my own view on those aspects is that what's currently in the PEP is fine, as the last round of discussions cleaned up some key problems where the way the explicitly out-of-tree build support was designed in earlier versions of the API didn't align with the way most build systems actually work (i.e. they expect to be given the target directory as a build setting, allowing the build system itself to decide exactly how to handle that).
Cheers, Nick.
-- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia _______________________________________________ Distutils-SIG maillist - Distutils-SIG@python.org https://mail.python.org/mailman/listinfo/distutils-sig
participants (4)
-
12345 67890 -
Brett Cannon -
Nick Coghlan -
Thomas Kluyver