From: simonpj <unknown>
Date: Wed, 15 Sep 1999 08:00:11 +0000 (+0000)
Subject: [project @ 1999-09-15 08:00:11 by simonpj]
X-Git-Tag: Approximately_9120_patches~5813
X-Git-Url: http://git.megacz.com/?a=commitdiff_plain;h=18e952cfcc7aab9d38d66e781183d5aadae38ff0;p=ghc-hetmet.git

[project @ 1999-09-15 08:00:11 by simonpj]
Extra words about unsafePerformIO
---

diff --git a/ghc/docs/libraries/IOExts.sgml b/ghc/docs/libraries/IOExts.sgml
index 2af0789..6038706 100644
--- a/ghc/docs/libraries/IOExts.sgml
+++ b/ghc/docs/libraries/IOExts.sgml
@@ -2,11 +2,34 @@
 <label id="sec:IOExts">
 <p>
 
+(Documetation for this library is sadly inadequate.)
+
 This library provides the following extensions to the IO monad:
 <itemize>
 <item>
 The operations <tt/fixIO/, <tt/unsafePerformIO/ and <tt/unsafeInterleaveIO/
-described in <cite id="ImperativeFP">
+described in <cite id="ImperativeFP">.
+<p>
+Everyone knows that if the I/O computation wrapped in <tt/unsafePerformIO/ 
+performs side effects, then the relative order in which those side effects
+take place (relative to the main I/O trunk, or other calls to <tt/unsafePerformIO/)
+is indeterminate.  It is less well known that <tt/unsafePerformIO/ is not type safe.
+For example:
+<tscreen><verb>
+  test :: IORef [a]
+  test = unsafePerformIO $ newIORef []
+
+  main = do
+          writeIORef test [42]
+          bang <- readIORef test
+          print (bang :: [Char])
+</verb></tscreen>
+This program will core dump.  This problem with polymorphic references is
+well known in the ML community, and does not arise with normal monadic use
+of references.  There is no easy way to make it impossible once you use
+<tt/unsafePerformIO/.  Indeed, it is possible to write
+<tt>coerce :: a -> b</tt> with the help of <tt/unsafePerformIO/.
+So be careful!
 
 <item>
 References (aka mutable variables) and mutable arrays (but no form of