Repository navigation
FIX: base always refers to the original subclass - #2746
Conversation
Closes numpy#470 The use case is that if b is a memmap, 'np.asarray(b).base' should be a memmap, and thus we put is to b in such a situation.
|
That branch seems to fix the originally reported issue. Thanks! |
There was a problem hiding this comment.
This comment has a grammar issue. Can you fix / rephrase?
There was a problem hiding this comment.
Maybe just inserting "when the" before "new base".
There was a problem hiding this comment.
Maybe just inserting "when the" before "new base".
Done, thanks.
There was a problem hiding this comment.
Your whitespace has gone wonky here (some tab problem perhaps)
There was a problem hiding this comment.
Your whitespace has gone wonky here (some tab problem perhaps)
Oops, sorry, messed up with my editor. Fixed
|
I haven't thought about this change carefully, but since we're talking about band-aid workarounds and the people who need the band-aid like it, I figure it's fine... and it still does collapse simple base chains, so the original bug should remain fixed. |
|
I don't think the travis failure is related to this pull request: Maybe an heisentest? |
|
On Fri, Nov 16, 2012 at 01:39:25AM -0800, Olivier Grisel wrote:
I think so too, as I didn't change anything relevent compared to the |
|
The failure is fixed by #2745. Gael, would you mind reviewing that PR for me please? As to this PR, it looks ok to me. @teoliphant, if you have a minute, would you mind checking it too, since you wrote the initial PR for this? |
|
Anybody else to review this fix before merging it to master? |
|
Works for me--this fixed a bug that was occurring in astropy/astropy#495. |
|
Since this fixes it, I am merging it. @GaelVaroquaux I think this should be backported to 1.7.x, correct? |
FIX: base always refers to the original subclass
Yes, I believe so. Thanks! |
Closes #470
The use case is that if b is a memmap, 'np.asarray(b).base' should be a
memmap, and thus we put is to b in such a situation.