Added code to detect UTF32 files for NSString contentsOfURL:usedEncoding #928
Conversation
|
Please can you squash the history of this, so that we don't get noisy messages like "Resolved merge conflicts"? and instead of having "Merge remote-tracking branch X" messages, you can do a You can also configure the 'rebase' to be the default by doing |
|
@alblue usually any merges we accept are squash merged anyhow as a default rule so that isn't too bad. |
| @@ -462,6 +462,8 @@ | |||
| 'TestFoundation/Resources/NSStringTestData.txt', | |||
| 'TestFoundation/Resources/NSString-UTF16-BE-data.txt', | |||
| 'TestFoundation/Resources/NSString-UTF16-LE-data.txt', | |||
| 'TestFoundation/Resources/NSString-UTF32-BE-data.txt', | |||
There was a problem hiding this comment.
are these files missing from the commit? I don't see them currently on master
There was a problem hiding this comment.
They're actually there. Because these files have BOM in them, Github is detecting them to be binary files. If you go ahead and click on "Files Changed" above, you would see them listed as binary files..
| XCTFail("Unable to init NSString from contentOf:usedEncoding:") | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
Do these pass on macOS using the system Foundation?
There was a problem hiding this comment.
@mohitathwani can you double check on this part?
Foundation/NSString.swift
Outdated
| @@ -1301,12 +1301,18 @@ extension NSString { | |||
| let readResult = try NSData(contentsOf: url, options:[]) | |||
|
|
|||
| let bytePtr = readResult.bytes.bindMemory(to: UInt8.self, capacity:readResult.length) | |||
| if readResult.length >= 2 && bytePtr[0] == 254 && bytePtr[1] == 255 { | |||
| if readResult.length >= 4 && bytePtr[0] == 255 && bytePtr[1] == 254 && bytePtr[2] == 0 && bytePtr[3] == 0 { | |||
There was a problem hiding this comment.
might be good to use the hex values for these since most sources of the unicode standards refer to 0xFF and 0xFE etc. (this is just a style nit not a functionality commentary)
There was a problem hiding this comment.
Updated my code with hex values
Foundation/NSString.swift
Outdated
| enc?.pointee = String.Encoding.utf16BigEndian.rawValue | ||
| } | ||
| else if readResult.length >= 2 && bytePtr[0] == 255 && bytePtr[1] == 254 { | ||
| enc?.pointee = String.Encoding.utf16LittleEndian.rawValue | ||
| } | ||
| else if readResult.length >= 4 && bytePtr[0] == 0 && bytePtr[1] == 0 && bytePtr[2] == 254 && bytePtr[3] == 255 { | ||
| enc?.pointee = String.Encoding.utf32BigEndian.rawValue | ||
| } |
There was a problem hiding this comment.
nit: We seem to be using some Int literals. Will it make sense to replace them with constants so that their contextual meaning is clear?
There was a problem hiding this comment.
What kind of naming conventions do you suggest for these literals?
mohitathwani
force-pushed
the
String_used_encoding
branch
from
f59f577
to
fcd4693
Compare
…L:usedEncoding: This pull request is in continuation to PR swiftlang#893
mohitathwani
force-pushed
the
String_used_encoding
branch
from
fcd4693
to
18cafb1
Compare
|
@swift-ci please test |
|
This looks good to me - @phausler WDYT? |
|
looks good to me @swift-ci please test and merge |
This pull request is in continuation to PR #893