I don't see the emails from the list.

[Brian K. White]

----- Original Message ----- 
From: "Mike Fedko" <filepro at adelphia.net>
To: "Filepro List" <filepro-list at lists.celestial.com>
Sent: Tuesday, September 25, 2007 9:02 PM
Subject: I don't see the emails from the list.

> I have subscribed to the filepro list. I don't get any emails from the 
> list.
> The only response I received from the list is one from Mike Schwartz
> who also sent it direct to my email also. Can someone help with this.

Your first message came through at 5pm eastern. A lot of people were 
probably simply done looking at email for the day by then. Or at least done 
working. A lot of the rest, like me, never see this kind of error as we use 
Unix and linux and almost never use the Windows version of fp. Of the 
regular Windows fp users, many probably only see this kind of thing so 
rarely that there is not enough data to suggest a cause or a fix and the 
only advice is general advice that applies to any program. Which is roughly: 
"Make sure the servers software, the servers hardware, the network, the 
clients hardware, and the clients software, are all in good shape."

However, since it happened right after a power loss that makes a pretty big 
likelyhood that the problem is bad data.
You most likely just have to locate one or more half-written or half-updated 
or otherwise corrupt records and either manually fix them or delete them, 
and then rebuild all indexes in the relevant files.

You may be better able to say how best to find bad records than we, since 
you know your app.

Short of writing custom report processes that scan through files and check 
each record for sanity according to whatever your app happens to need in 
order for a record to be sane, some ideas are:

Look through the involved files by record number (not by any index since 
even if no index file was scrambled, if a data record was scrambled, then 
the indexes no longer match the data).
Either step through the file record by record using some screen that shows a 
lot of fields.
Or enter browse mode, change the browse format to include some fields from 
both the beginning and end of each record (low & high field numbers) and 
page down through and hopefully you can spot the bad record(s) that way 
while seeing 20 at a time. Maybe use extended selection and define a temp 
selection set based on @cd or @ud to zero in on records created or updated 
today.

Once you think you have all good records (whether you find any bad ones or 
not), then rebuild all indexes in those files.

Then try the process again.

Try modifying the process such that it doesn't select any records, does it 
still crash or simply not do anything gracefully?

If you're still crashing after that then we'll have to get more systematic 
about eliminating possibilities.

Brian K. White    brian at aljex.com    http://www.myspace.com/KEYofR
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro  BBx    Linux  SCO  FreeBSD    #callahans  Satriani  Filk!