#5184: createDirectory does not retry on EINTR (interrupted)
------------------------------------------+---------------------------------
  Reporter:  CoreyOConnor                 |          Owner:  simonmar
      Type:  bug                          |         Status:  closed  
  Priority:  high                         |      Milestone:  7.2.1   
 Component:  libraries/unix               |        Version:  7.0.3   
Resolution:  fixed                        |       Keywords:          
  Testcase:                               |      Blockedby:          
Difficulty:                               |             Os:  MacOS X 
  Blocking:                               |   Architecture:  x86     
   Failure:  Incorrect result at runtime  |  
------------------------------------------+---------------------------------

Comment(by CoreyOConnor):

 I think there is a deeper issue here and the EINTR error might just be a
 red herring.

 I still cannot create directories with "createDirectoryIfMissing True" as
 expected. I'm going to leave my comments attached to this bug for now.

 In my application I call:
 {{{
 createDirectoryIfMissing True "/home/coconnor/Development/dev-
 system/dist/build"
 }}}

 on a system where "/home/coconnor/Development/dev-system/dist" already
 exists. Instead of creating the "build' directory under "dist" as expected
 the application fails with:
 {{{
 dev: /home/coconnor: createDirectory: failed (Operation not supported)
 }}}

 It does not make sense that "createDirectory" would ever be called with an
 existing directory by "createDirectoryIfMissing".

 I'm not sure how this can occur. If there was a bug in getFileStatus or
 isDirectory then this could occur.

 The only thing I can think of has a weak argument.
 The stat structure on OS X comes in two varieties. The define
 _DARWIN_FEATURE_64_BIT_INODE is used to select between the two varieties.
 If the fstat call used at runtime expects a different stat structure than
 System/Posix/Files.hsc is compiled against then getFileStatus can succeed
 but the status would be junk.

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5184#comment:4>
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

Reply via email to