Merge pull request #935 from dbcli/RW/utf8mb4-default-charset

Default to standards-compliant utf8mb4 character set

Author: rolandwalker

CONTRIBUTING.md

@@ -80,10 +80,10 @@ $ export PYTEST_HOST=localhost
 $ export PYTEST_USER=mycli
 $ export PYTEST_PASSWORD=myclirocks
 $ export PYTEST_PORT=3306
-$ export PYTEST_CHARSET=utf8
+$ export PYTEST_CHARSET=utf8mb4
 ```
 
-The default values are `localhost`, `root`, no password, `3306`, and `utf8`.
+The default values are `localhost`, `root`, no password, `3306`, and `utf8mb4`.
 You only need to set the values that differ from the defaults.
 
 If you would like to run the tests as a user with only the necessary privileges,

README.md

@@ -142,6 +142,14 @@ Thanks to [PyMysql](https://github.com/PyMySQL/PyMySQL) for a pure python adapte
 
 Mycli is tested on macOS and Linux, and requires Python 3.10 or better.
 
+To connect to MySQL versions earlier than 5.5, you may need to set the following in `~/.myclirc`:
+
+```
+# character set for connections without --charset being set at the CLI
+default_character_set = utf8
+```
+
+or set `--charset=utf8` when invoking MyCLI.
 
 ### Configuration and Usage
 

changelog.md

@@ -4,6 +4,7 @@ TBD
 Features
 --------
 * Add `--unbuffered` mode which fetches rows as needed, to save memory.
+* Default to standards-compliant `utf8mb4` character set.
 
 
 Bug Fixes

mycli/main.py

@@ -514,7 +514,7 @@ def connect(
                 socket = socket or cnf["socket"] or cnf["default_socket"] or guess_socket_location()
 
         passwd = passwd if isinstance(passwd, str) else cnf["password"]
-        charset = charset or cnf["default-character-set"] or "utf8"
+        charset = charset or self.config["main"].get("default_character_set") or cnf["default-character-set"] or "utf8mb4"
 
         # Favor whichever local_infile option is set.
         use_local_infile = False

mycli/myclirc

@@ -131,6 +131,9 @@ enable_pager = True
 # Choose a specific pager
 pager = 'less'
 
+# character set for connections without --charset being set at the CLI
+default_character_set = utf8mb4
+
 [keys]
 # possible values: auto, fzf, reverse_isearch
 control_r = auto

test/myclirc

@@ -129,6 +129,9 @@ enable_pager = True
 # Choose a specific pager
 pager = less
 
+# character set for connections without --charset being set at the CLI
+default_character_set = utf8mb4
+
 [keys]
 # possible values: auto, fzf, reverse_isearch
 control_r = auto

test/utils.py

@@ -16,7 +16,7 @@
 USER = os.getenv("PYTEST_USER", "root")
 HOST = os.getenv("PYTEST_HOST", "localhost")
 PORT = int(os.getenv("PYTEST_PORT", "3306"))
-CHARSET = os.getenv("PYTEST_CHARSET", "utf8")
+CHARSET = os.getenv("PYTEST_CHARSET", "utf8mb4")
 SSH_USER = os.getenv("PYTEST_SSH_USER", None)
 SSH_HOST = os.getenv("PYTEST_SSH_HOST", None)
 SSH_PORT = int(os.getenv("PYTEST_SSH_PORT", "22"))