Rename MemoryIO to IO::Memory #3563
Conversation
|
My PR also did a |
|
|
||
| class MemoryIO < IO::Memory | ||
| def self.new(*args, **nargs) | ||
| {{ puts "Warning: MemoryIO is deprecated and will be removed after 0.20.0, use IO::Memory instead".id }} |
There was a problem hiding this comment.
Maybe a line number of the caller?
There was a problem hiding this comment.
There's no way to get that info right now. In the future we could add it, but it will always be the one of the first call site, because later method calls with the same argument types are not re-analyzed
asterite
force-pushed
the
feature/io_memory
branch
from
c880cbc to
0da8a24
Compare
|
@RX14 Done the a/an replacement, good catch :-) |
| @@ -46,7 +46,7 @@ module IO::Buffered | |||
| # We first check, after filling the buffer, if the delimiter | |||
| # is already in the buffer. In that case it's much faster to create | |||
| # a String from a slice of the buffer instead of appending to a | |||
| # raw IO. Optimizations for specific IOs are tested separately | ||
| # (for example in buffered_io_spec) | ||
| private class SimpleMemoryIO | ||
| private class SimpleIO::Memory |
| @@ -397,8 +397,8 @@ module Crystal | |||
| produce_tuple = node.names.first == "T" | |||
| when GenericInstanceType | |||
| produce_tuple = ((splat_index = path_lookup.splat_index) && | |||
| path_lookup.type_vars.keys.index(node.names.first) == splat_index) || | |||
| (path_lookup.double_variadic? && path_lookup.type_vars.first_key == node.names.first) | |||
| path_lookup.type_vars.keys.index(node.names.first) == splat_index) || | |||
|
Specs |
|
I spent quite a bit of time on the old PR, it's not as easy as you would think at first! |
|
@RX14 Do you want to rebase your PR instead? It will probably be easier :-) |
|
(well, rebase or do it from scratch, whatever is easier) |
|
Well you've done a lot with this PR too so why not go ahead with this. |
asterite
force-pushed
the
feature/io_memory
branch
3 times, most recently
from
334a4c4 to
ef0bf43
Compare
MemoryIO is still available but produces a deprecation warning at compile-time, to ease migration.
asterite
force-pushed
the
feature/io_memory
branch
from
ef0bf43 to
7c24948
Compare
| @@ -3,6 +3,7 @@ | |||
| * **(breaking change)** Removed `ifdef` from the language | |||
| * **(breaking change)** Removed `PointerIO` | |||
| * **(breaking change)** The `body` property of `HTTP::Request` is now an `IO?` (previously it was `String`). Use `request.body.try(&.gets_to_end)` if you need the entire body as a String. | |||
| * **(breaking change)** `MemoryIO` has been renamed to `IO::Memory`. The old name can still be used but will produce a compile-time warning. `MemoryIO` will be removed immediately after 0.20.0. | |||
There was a problem hiding this comment.
Do you mean removed in 0.21.0 or 0.20.1? This comment seems to imply the latter, which I don't think is what you meant.
There was a problem hiding this comment.
Yes, that's what I meant. MemoryIO is gone but it's still there to ease migration. I want to remove it as soon as possible. Or we can do it in 0.21.0. In any case we are at 0.x so semver doesn't truly applies.
There was a problem hiding this comment.
Since we added a deprecation, we can wait for 0.21 to be released :)
|
@asterite there are |
|
@Sija Thanks. It seems they got added after I merged an old PR |
Similar to #3401. MemoryIO is still available but produces a deprecation warning at compile-time, to ease migration.