Problem about pattern matching.

newer
Cabal package with BSD3 library...

Magicloud Magiclouds

28 Apr 2010 28 Apr '10

8:49 a.m.

Hi, I have code as below. How come "case" version works wrong and gives me "overlap" compiling warning? Thanks. if dayOfMonth == firstDayOfMonth then v day (x, y) else if dayOfMonth == lastDayOfMonth then not_ $ v day (x, y) else Mider day (x, y) case dayOfMonth of firstDayOfMonth -> v day (x, y) lastDay -> not_ $ v day (x, y) _ -> Mider day (x, y) -- 竹密岂妨流水过山高哪阻野云飞

Show replies by date

minh thu

28 Apr 28 Apr

9:02 a.m.

2010/4/28 Magicloud Magiclouds :

...

Hi, I have code as below. How come "case" version works wrong and gives me "overlap" compiling warning? Thanks. if dayOfMonth == firstDayOfMonth then v day (x, y) else if dayOfMonth == lastDayOfMonth then not_ $ v day (x, y) else Mider day (x, y)

case dayOfMonth of firstDayOfMonth -> v day (x, y) lastDay -> not_ $ v day (x, y) _ -> Mider day (x, y)

Hi, firstDayIfMonth and lastDay are new variables, not previously bound variable. This means that trying to pattern match (the value of) dayOfMonth will always succeed with the first alternative. Think of parsing command line argument, using the last alternative as a fall-through : args <- getArgs case args of ["--help"] -> ... ["--run-tests"] -> ... x -> putStrLn $ "unknow args " ++ concat x What would happen if the last alternative (with the x) was put in first position ? Cheers, Thu

Henning Thielemann

5:51 p.m.

minh thu schrieb:

...

2010/4/28 Magicloud Magiclouds :

...
Hi, I have code as below. How come "case" version works wrong and gives me "overlap" compiling warning? Thanks. if dayOfMonth == firstDayOfMonth then v day (x, y) else if dayOfMonth == lastDayOfMonth then not_ $ v day (x, y) else Mider day (x, y)

case dayOfMonth of firstDayOfMonth -> v day (x, y) lastDay -> not_ $ v day (x, y) _ -> Mider day (x, y)

Hi,

firstDayIfMonth and lastDay are new variables, not previously bound variable. This means that trying to pattern match (the value of) dayOfMonth will always succeed with the first alternative.

You can however simulate a 'case' with predefined variables: http://haskell.org/haskellwiki/Case

Magicloud Magiclouds

29 Apr 29 Apr

8:48 a.m.

Sorry folks, I did not make it clear. The firstDayOfMonth and lastDayOfMonth was defined earlier in let ... in structure. The problem here is, ghc warned me that "_" and "lastDayOfMonth" was overlapped. And the results showed that, everything that should be worked with "Mider" was done by "not_ ...". On Thu, Apr 29, 2010 at 1:51 AM, Henning Thielemann wrote:

...

minh thu schrieb:

...
2010/4/28 Magicloud Magiclouds :

...
Hi, I have code as below. How come "case" version works wrong and gives me "overlap" compiling warning? Thanks. if dayOfMonth == firstDayOfMonth then v day (x, y)S else if dayOfMonth == lastDayOfMonth then not_ $ v day (x, y) else Mider day (x, y)

case dayOfMonth of firstDayOfMonth -> v day (x, y) lastDay -> not_ $ v day (x, y) _ -> Mider day (x, y)

Hi,

firstDayIfMonth and lastDay are new variables, not previously bound variable. This means that trying to pattern match (the value of) dayOfMonth will always succeed with the first alternative.

You can however simulate a 'case' with predefined variables: http://haskell.org/haskellwiki/Case

-- 竹密岂妨流水过山高哪阻野云飞

Brandon S. Allbery KF8NH

9 a.m.

On Apr 29, 2010, at 04:48 , Magicloud Magiclouds wrote:

...

Sorry folks, I did not make it clear. The firstDayOfMonth and lastDayOfMonth was defined earlier in let ... in structure.

Doesn't matter. Pattern matches *create* local bindings; they don't pay any attention to existing ones. So you have a match-anything-and- bind-to-"firstDayOfMonth" pattern and a match-anything-and-discard pattern; the original firstDayOfMonth is hidden by the new one created by the pattern match, and since both patterns match the same thing (that is, anything) they overlap. -- brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery@kf8nh.com system administrator [openafs,heimdal,too many hats] allbery@ece.cmu.edu electrical and computer engineering, carnegie mellon university KF8NH

minh thu

9:01 a.m.

Hi, Precisely, the fact that you already defined them is not important: variables that appear in the left-hand-side of a production are new, local variables, just as it would be the case with nested lets: let x = 1 in let x = 2 in x -- what's the value of x ? let x = 1 in case 2 of x -> x -- what's the value of x ? now, in the expression case 1 of x -> ... y -> ... _ -> ... every left-hand-side can pattern-match succesfully against the value 1, but since only the first match is used, y and _ are useless. Since they match identical values (in this case, they match anything), they are overlapped. Make sure you can tell what the value of the x above are and explain why. Cheers, Thu 2010/4/29 Magicloud Magiclouds :

...

Sorry folks, I did not make it clear. The firstDayOfMonth and lastDayOfMonth was defined earlier in let ... in structure. The problem here is, ghc warned me that "_" and "lastDayOfMonth" was overlapped. And the results showed that, everything that should be worked with "Mider" was done by "not_ ...".

On Thu, Apr 29, 2010 at 1:51 AM, Henning Thielemann wrote:

...
minh thu schrieb:

...
2010/4/28 Magicloud Magiclouds :

...
Hi, I have code as below. How come "case" version works wrong and gives me "overlap" compiling warning? Thanks. if dayOfMonth == firstDayOfMonth then v day (x, y)S else if dayOfMonth == lastDayOfMonth then not_ $ v day (x, y) else Mider day (x, y)

case dayOfMonth of firstDayOfMonth -> v day (x, y) lastDay -> not_ $ v day (x, y) _ -> Mider day (x, y)

Hi,

firstDayIfMonth and lastDay are new variables, not previously bound variable. This means that trying to pattern match (the value of) dayOfMonth will always succeed with the first alternative.

You can however simulate a 'case' with predefined variables: http://haskell.org/haskellwiki/Case

-- 竹密岂妨流水过山高哪阻野云飞

5548

Age (days ago)

5549

Last active (days ago)

List overview

Download

5 comments

4 participants

participants (4)

Brandon S. Allbery KF8NH
Henning Thielemann
Magicloud Magiclouds
minh thu