Re: Finalization of I/O Channels

From: John Skaller (skaller@maxtal.com.au)
Date: Fri Jul 30 1999 - 17:22:13 MET DST


Message-Id: <3.0.6.32.19990731012213.0098dcf0@mail.triode.net.au>
Date: Sat, 31 Jul 1999 01:22:13 +1000
To: "David McClain" <dmcclain@azstarnet.com>, <caml-list@inria.fr>
From: John Skaller <skaller@maxtal.com.au>
Subject: Re: Finalization of I/O Channels
In-Reply-To: <000101bed925$88815ff0$210148bf@dylan>

At 11:17 28/07/99 -0700, David McClain wrote:
>In attempting to provide finalizable I/O channels to my users, I discovered
>that OCAML 2.02 is already finalizing them... But in so doing you do not
>flush or close the file descriptors. Is this an oversight, or is there some
>good reason for not doing so?

        Perhaps worse, I found that explicitly closing
a file does not always work. I do not understand why,
but this is surely a bug. This occured in the bytecode
interpreter, and went away in with the native code
compiler. [A program that wrote a file could not read it
after closing the output file and reopening it input]

-------------------------------------------------------
John Skaller email: skaller@maxtal.com.au
                http://www.maxtal.com.au/~skaller
                phone: 61-2-96600850
                snail: 10/1 Toxteth Rd, Glebe NSW 2037, Australia



This archive was generated by hypermail 2b29 : Sun Jan 02 2000 - 11:58:24 MET