[file.streams] Change sub-clauses to rSec2, similar to [string.streams] #385
Conversation
|
@zygoloid, are you OK with this change? IMHO only the [c.files] relocation is non-obvious. |
|
I agree that promoting [filebuf], [ifstream], [ofstream] and [fstream] to Promoting [filebuf] to The [c.files] change needs a corresponding change to Table 121, and we'll need to check for other cross references that assume |
|
I've pushed another commit to the branch with the requested changes. I moved [fstreams]/2 before the existing paragraphs in [iostream.objects.overview](but after the synopsis) as it didn't seem to make sense elsewhere in that sub-clause. |
|
Can the commits be rebased to remove the merge? |
|
@zygoloid: Ping |
[file.streams] Change sub-clauses to rSec2, similar to [string.streams]
[string.streams] and [file.streams] follow very similar structure (an introduction followed by streambuf, input stream, output stream and bidirectional stream class templates) but are nested entirely differently. Under [fstreams] we have an entirely flat hierarchy where the definition of
basic_streambufis at the same level as its own members.This change makes [file.streams] consistent with [string.streams] by promoting [filebuf], [ifstream], [ofstream] and [fstream] to
rSec2.Also promote [c.files] to
rSec1, I don't think<cstdint>should be nested underneath the sub-clause forbasic_fstreambufandbasic_fstreametc.