use python 2.6's streaming zipfile support instead of our own zipstream module
|Reported by:||kmike||Owned by:||lvh|
|Cc:||kmike, Thijs Triemstra||Branch:||
branch-diff, diff-cov, branch-cov, buildbot
As I can see the goal of zipstream is to provide support for an incremental approach to unzipping files.
But standard zipfile has this support since python 2.6. ZipFile?.open() method returns file-like object which supports 'read' method with buffer size in parameters.
Implementation in zipfile is better than in zipstream. At least it works a lot faster because it don't have 'self.buffer = self.buffer[n:]' code which kills performance on big (even tens of Kb's) buffer sizes like zipstream version does.
I think is better to deprecate zipstream and copy implementation from new zipfile. Or make zipstream a backported zipfile wrapper.
Change History (17)
comment:2 Changed 7 years ago by
|Owner:||changed from Glyph to kmike|
|Summary:||deprecate or improve zipstream → use python 2.6's streaming zipfile support instead of our own zipstream module|
comment:5 Changed 6 years ago by
|Owner:||Andy Ruse deleted|
|Status:||assigned → new|
comment:7 Changed 6 years ago by
|Cc:||Thijs Triemstra added|
|Owner:||set to Andy Ruse|