In MX, I'm not sure what underlying classes are being used (java.io vs. java.nio), so you'll probably have to do the same thing in MX.
<cffile> sucks. What CF needs is the ability to read/write file streams.
----- Original Message -----
From: Chris More <[EMAIL PROTECTED]>
Date: Thursday, November 13, 2003 11:12 am
Subject: CFFILE BUG CF5.0 (DEADLOCK)
> I have been experiencing a CF Bug for quite some time now and I
> wrote some code to get around the problem, but I am amazed that
> Macromedia never posted a fix other then saying to upgrade to MX.
>
> I am writing a text file from a very large query. It takes a few
> minutes to run and exports of the order of 16,000 rows of data to
> a text file. The first CFFILE is a write that only added one row
> of column header descriptions. As I loop through the query I
> append on one row of data at a time. Well I have experienced on
> large queries that CF doesn't always close the file before the
> loop continues. So the next time the CFFILE append is executed an
> error is generated because the file is currently in use. It
> happens at random times and it takes a random amount of time for
> CF to release the file and allow it to continue.
>
> I put a TRY/CATCH statement around the CFFILE append and I could
> catch the error. I then jumped into a loop that would continuously
> try to append to the file up to 500 retries. After the first
> successful retry it jumps out of the loop and continues to read
> the query and append the file. 500 retries seem to work as at 300
> tries I still had a few times that the file was still open when
> the next append occurred.
>
> Other people have complained about this problem, but Macromedia
> never did anything about it.
>
> No one else is using the file when the error occurs since the
> filename is always dynamic and unique when created.
>
> Does anyone know if someone wrote a CFX version of CFFILE that
> gets around this bug?
>
> Thanks,
> Chris More
>
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings]
