support malformed char literals, the second #44989
Conversation
simeonschaub
added
parser
simeonschaub
changed the title
| @test_parseerror "'\\xff\\xff\\xff\\xff'" "character literal contains multiple characters" # == reinterpret(Char, 0xffffffff) | ||
| @test '\uffff' == Char(0xffff) | ||
| @test '\U00002014' == Char(0x2014) | ||
| @test '\100' == reinterpret(Char, UInt32(0o100) << 24) |
There was a problem hiding this comment.
Yeah, perhaps this makes it more obvious:
| @test '\100' == reinterpret(Char, UInt32(0o100) << 24) | |
| @test '\100' == reinterpret(Char, UInt32(0o100) << 24) == '@' |
| @test_parseerror "'\\U00002014a'" "character literal contains multiple characters" | ||
| @test_parseerror "'\\1000'" "character literal contains multiple characters" | ||
| @test Meta.isexpr(Meta.parse("'a"), :incomplete) | ||
| @test ''' == "'"[1] |
There was a problem hiding this comment.
Huh, I never realized that would work, but it already does so carry on...
|
I think this implements exactly what I had in mind. My only quibble is with the commit message, which says:
That isn't strictly true since it allows invalid UTF-8, so I would describe this perhaps this way:
|
This is pulled out from #44989. Making `test_parseerror` a macro leads to much more helpful stacktraces if something fails.
JeffBezanson
force-pushed
the
sds/char_literals2
branch
from
71683df to
16e37c2
Compare
|
Tidied up and added news, should be ready to merge. |
JeffBezanson
force-pushed
the
sds/char_literals2
branch
from
16e37c2 to
a1ce793
Compare
|
I see this changed the flisp parser, so I guess JuliaSyntax.jl needs to be updated too, I assume forgotten. I'm also thinking, invalid UTF-8 means overlong, and it can be overlong to arbitrary degree (in Strings), while Char is fixed, was that an( unsolvable?) problem thought of? |
Alternative to #44765. This disallows character literals that can not be
created from iterating a UTF-8 string.
fixes #25072