#5375: Regression in newName
---------------------------------+------------------------------------------
Reporter: reinerp | Owner:
Type: bug | Status: new
Priority: highest | Milestone: 7.2.1
Component: Template Haskell | Version: 7.3
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: GHC rejects valid program
---------------------------------+------------------------------------------
Comment(by reinerp):
I should point out another (perhaps more common) need for the "totally
fresh" semantics. When I write a quasiquoter which antiquotes to Haskell,
for instance for interpolated strings (see
[http://hackage.haskell.org/package/interpolatedstring-perl6-0.8.1]),
consider the following:
{{{
f x = [qc|The value of x is {x}|]
}}}
(here antiquotation is delimited by {}s.) In this case, I want the
antiquoted {{{x}}} to refer to {{{f}}}'s parameter. My quasiquotation will
evaluate to a syntax tree as follows:
{{{
f x = $( ... (VarE (mkName "x")) ... )
}}}
What happens if I want to bind a variable somewhere in the syntax tree? I
might for instance want to produce the following:
{{{
f x = let theString = "The value of x is "
in theString ++ show x
}}}
In this case the name {{{theString}}} should be totally fresh: it's an
implementation detail of the quasiquoter, and we don't want its binding to
capture the variable {{{x}}}. So in this case we want:
{{{
f x =
$(do
theString <- newName "theString" -- old semantics are needed here
return $ LetE [ValD (VarP theString) (NormalB (LitE "The value of x
is")) []]
(InfixE (VarE theString) '(++) (VarE (mkName "x")))
)
}}}
Cheers,
Reiner
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5375#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
