Use Encoding::BINARY for JOHAB #472
Conversation
* A dummy encoding does not understand JOHAB anyway. * Creating an extra encoding leads to significant complexity and slowdowns in Ruby implementations: https://bugs.ruby-lang.org/issues/18949
| @conn.set_client_encoding "JOHAB" | ||
| val = @conn.exec("SELECT chr(x'3391'::int)").values[0][0] | ||
| expect( val.encoding.name ).to eq( "JOHAB" ) | ||
| end |
There was a problem hiding this comment.
Note: this test no longer makes sense/can no longer test anything useful since pg does not create encodings dynamically anymore with this PR.
There was a problem hiding this comment.
The test is actually to ensure
Line 27 in 2821bf7
|
@larskanis Could you review this? |
|
IMHO the code for JOHAB is exactly how ruby encodings are intended to be used: It tries to find a corresponding encoding in the running ruby VM, just in case a future ruby version has support for it. And if the search fails, then a placeholder encoding is created, so that the encoding information isn't lost. That's all OK, but the time of new character encodings is long over. No one invents character encodings these days. If someone would care about JOHAB, he would have implemented support for it in ruby core. So this is dead code. We already had some discussion about the uselessness of JOHAB encoding in the past #28 (comment) where we already removed the creation of aliases to the dummy encoding here. And it was my intention to remove the code entirely in the near future. |
|
Great, thank you for merging. |
in Ruby implementations: https://bugs.ruby-lang.org/issues/18949