"""
For example, this code:

 f'abc{expr1:spec1}{expr2!r:spec2}def{expr3:!s}ghi'

Might be be evaluated as:

 'abc' + format(expr1, spec1) + format(repr(expr2)) + 'def' + format(str(expr3)) + 'ghi'
"""

format(repr(expr2)) should be format(repr(expr2), spec2)



"""
>>> f'x={x'
  File "<stdin>", line 1
SyntaxError: missing '}' in format string expression
"""

Actually implemented exception message is:
SyntaxError: f-string: expecting '}'



"""
>>> f'x={!x}'
  File "<fstring>", line 1
    !x
    ^
SyntaxError: invalid syntax
"""

Actually implemented exception message is:
SyntaxError: f-string: invalid conversion character: expected 's', 'r', or 'a'

New changeset 831276834e4b by Eric V. Smith in branch 'default':
Partial fix for issue 25206: Fixed format() call.
https://hg.python.org/peps/rev/831276834e4b

Also in the first example, the colon (format specifier) and exclamation mark (conversion) are in the wrong order. It should either be

f"{expr3:!s}" → format(expr3, "!s")

or

f"{expr3!s:}" → format(str(expr3), "")

New changeset df19ba9217e4 by Eric V. Smith in branch 'default':
More on issue 25206: typo.
https://hg.python.org/peps/rev/df19ba9217e4

Thanks for the feedback.

On the 
f"{expr3!s}" → format(str(expr3))
issue, I'm trying to show the 1 argument version of format (which is what the code generator actually calls). Although I realize that's not a great example, since the string conversion is implicit in format(). Maybe I'll have the !r example be the 1 argument format() call.

I'll fix the exceptions, eventually. But I don't want the PEP to become a exact specification of what the exception text will actually be.

On this one:
f'x={!x}'
I'll probably change it to be an error with the empty expression, not with the invalid conversion character.

New changeset 7e32da8da58e by Eric V. Smith in branch 'default':
Issue 25206: fix f-string exceptions to match the code.
https://hg.python.org/peps/rev/7e32da8da58e