Patch by Caelyn McAulay.
Neither Caelyn nor I could find any cases in 2.7 or 3.4/5 where an
operation on a closed stream raised IOError; generally the C
implementation have a macro to check for the stream being closed, and
these macros all raised ValueError. If we find any, a new bug should
be opened.
read or readall methods no longer lose data when an underlying read
system call is interrupted within an io module object. IOError is no
longer raised due to a read system call returning EINTR from within
these methods.
This is a backport of changeset 781b95159954 from 3.2.
The earlier 2.7 changeset 67dc99a989cd already fixed this for the
builtin python 2.x file object.
file position when calling `truncate()`. It would previously change the
file position to the given argument, which goes against the tradition of
ftruncate() and other truncation APIs. Patch by Pascal Chambon.
This is for consistency with imitation file objects like StringIO and BytesIO.
This commit also adds a few tests, where they were lacking for concerned
methods.
Andrew Kuchling
Terry Jan Reedy
Gregory P. Smith
Ezio Melotti
Victor Stinner
Antoine Pitrou
Benjamin Peterson