Conversion of float 'inf' to string fails after 1.3.14

[Zac Bentley]

Read this first

Describe the bug

If I run cursor.execute("SELECT '%s' FROM dual", (float('inf'),), on 1.3.13 and earlier, I get (('inf',),) back, as expected. On later versions, I get a ProgrammingError raised by this converter indicating that infinity is not allowed.

That change was introduced with good intentions, but I do not think it makes sense. It's reasonable to want to stringify inf/nan and store that representation in a string column. Using a custom converter should not be necessary.

To Reproduce

Code

con = MySQLdb.connect(...)
cur = con.cursor()
cur.execute("SELECT '%s' FROM dual", (float('inf'),)

Environment

MySQL Server

• Server OS: MacOS 12, Ubuntu 20
• Server Version: MySQL 5.7

MySQL Client

• OS MacOS 12
• Python Pyenv 3.7.13
• Connector/C Homebrew MySQL @5.7, linked.

[Inada Naoki]

str(float('inf')) is enough.