From d58d44f30b8a39a40716903df645767b12445b09 Mon Sep 17 00:00:00 2001
From: Iustin Pop <iustin@google.com>
Date: Thu, 28 Feb 2013 11:05:45 +0100
Subject: [PATCH] Switch Attoparsec parser from double to rational
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
According to the documentation, βThis function is almost ten times
faster than rational, but is slightly less accurate. For 94.2% of
numbers, this function and rational give identical results, but for
the remaining 5.8%, this function loses precision around the 15th
decimal place. For 0.001% of numbers, this function will lose
precision at the 13th or 14th decimal place.β. What happens is that
for our tests, it can happen that βAttoparsec.double (show a_double)β
is quite different from βa_doubleβ, such that we have much more than
1e-12 absolute difference.
Since our xm lists should not be too big, I think switching to
rational is better. Next patch also changes the way we compare
doubles, so maybe this patch is not really neededβ¦
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>
---
src/Ganeti/Hypervisor/Xen/XmParser.hs | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/Ganeti/Hypervisor/Xen/XmParser.hs b/src/Ganeti/Hypervisor/Xen/XmParser.hs
index a5459f29d..c24d11147 100644
--- a/src/Ganeti/Hypervisor/Xen/XmParser.hs
+++ b/src/Ganeti/Hypervisor/Xen/XmParser.hs
@@ -59,8 +59,8 @@ lispConfigParser =
where listConfigP = LCList <$> (A.char '(' *> liftA2 (++)
(many middleP)
(((:[]) <$> finalP) <|> (rparen *> pure [])))
- doubleP = LCDouble <$> A.double <* A.skipSpace <* A.endOfInput
- innerDoubleP = LCDouble <$> A.double
+ doubleP = LCDouble <$> A.rational <* A.skipSpace <* A.endOfInput
+ innerDoubleP = LCDouble <$> A.rational
stringP = LCString . unpack <$> A.takeWhile1 (not . (\c -> isSpace c
|| c `elem` "()"))
wspace = AC.many1 A.space
--
GitLab