[issue3560] redundant "base" field in memoryview objects

Antoine Pitrou <[EMAIL PROTECTED]> added the comment: Well, the PEP needs some love generally speaking; many things are un- or under-specified (and I don't understand better than anyone else). As for docs, there aren't any ("grep -ri memoryview Doc/" yields only the two lines referencing the "mem

[issue3560] redundant "base" field in memoryview objects

Guido van Rossum <[EMAIL PROTECTED]> added the comment: On 2008/08/19 17:18:45, GvR wrote: > PS. The PEP probably needs an update. And the docs. And Misc/NEWS http://codereview.appspot.com/3003 ___ Python tracker <[EMAIL PROTECTED]>

[issue3560] redundant "base" field in memoryview objects

Guido van Rossum <[EMAIL PROTECTED]> added the comment: PS. The PEP probably needs an update. And the docs. http://codereview.appspot.com/3003 ___ Python tracker <[EMAIL PROTECTED]> ___ __

[issue3560] redundant "base" field in memoryview objects

Guido van Rossum <[EMAIL PROTECTED]> added the comment: Reviewers: GvR, Message: Looks good. Go ahead and submit, assuming you've run full tests. Description: http://bugs.python.org/issue3560 Please review this at http://codereview.appspot.com/3003 Affected files: Include/memoryobject.h

[issue3560] redundant "base" field in memoryview objects

Barry A. Warsaw <[EMAIL PROTECTED]> added the comment: I don't have time to review the patch, but based on the description, I'm +1 on landing this before beta 3. I'd like to get the API right for 3.0 and it will be too late to land it after the release candidates start. -- nosy: +barry

[issue3560] redundant "base" field in memoryview objects

Antoine Pitrou <[EMAIL PROTECTED]> added the comment: Ok, here is a simple patch. It: - moves the struct definition at the end of memoryobject.h with a comment that the definition should not be considered public - adds two macros for accessing the underlying Py_buffer* and PyObject*, respectively

[issue3560] redundant "base" field in memoryview objects

Changes by Antoine Pitrou <[EMAIL PROTECTED]>: -- assignee: -> pitrou ___ Python tracker <[EMAIL PROTECTED]> ___ ___ Python-bugs-list m

[issue3560] redundant "base" field in memoryview objects

Martin v. Löwis <[EMAIL PROTECTED]> added the comment: > I've filed this as critical because it is a new API and, if we change > it, we'd better change it before 3.0 is released. I don't think it is API. The structure may be defined in a public header, but it is not intended to be used directly.

[issue3560] redundant "base" field in memoryview objects

Antoine Pitrou <[EMAIL PROTECTED]> added the comment: Le vendredi 15 août 2008 à 19:10 +, Martin v. Löwis a écrit : > Martin v. Löwis <[EMAIL PROTECTED]> added the comment: > > > Because it should be fixed before 3.0 final? > > And why should that be done? IMO, this can still > be fixed in

[issue3560] redundant "base" field in memoryview objects

Martin v. Löwis <[EMAIL PROTECTED]> added the comment: > Because it should be fixed before 3.0 final? And why should that be done? IMO, this can still be fixed in 3.1, or not a fixed at all - I fail to see the true bug (apart from the minor redundancy). ___ P

[issue3560] redundant "base" field in memoryview objects

Georg Brandl <[EMAIL PROTECTED]> added the comment: Because it should be fixed before 3.0 final? -- nosy: +georg.brandl ___ Python tracker <[EMAIL PROTECTED]> ___ _

[issue3560] redundant "base" field in memoryview objects

Martin v. Löwis <[EMAIL PROTECTED]> added the comment: Why is this a "critical" bug? -- nosy: +loewis ___ Python tracker <[EMAIL PROTECTED]> ___ ___

[issue3560] redundant "base" field in memoryview objects

New submission from Antoine Pitrou <[EMAIL PROTECTED]>: PyMemoryViewObject has a "base" field which points to the originator of the buffer. However, this field has become redundant now that the Py_buffer struct has received an "obj" field which also points to the originator of the buffer (this ha