HP3000-L Archives

May 1997, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Lars Appel <[log in to unmask]>
Reply To:
Lars Appel <[log in to unmask]>
Date:
Thu, 8 May 1997 12:35:45 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (42 lines)
At 09:09 08.05.1997 +1000, you wrote:
>Should Samba/iX be expected to handle MPE fixed length ASCII records?

The current version still has problems with MPE file types like FA,
which might depend on the very PC editor you use. The cause seems to
be that "reported" file size (i.e. EOF times recsize) does not match
"perceived" file size (when reading thru bytestream emulation which
removes trailing blanks from records and "fakes" LF delimiters).

Just compare EOF times recsize to what Posix wc command reports.

If your editor retrieves file attributes (Posix will use stat() for
this) and then reads N bytes it will likely choke or get garbage chars
at the end of the text. If your editor reads until end-of-file reported
then it should be fine. Unfortunately most PC editors seem to take the
first route.

I believe it is possible to tweak the sys_stat() routine in system.c
to detect a FA type file and perform a "silent" wc-like scan to report
a better file-size but that would cost some performance penality.

Another "workaround" might be switching your FA files to bytestream
format using TOBYTE.HPBIN.SYS (and FROMBYTE.HPBIN.SYS for the reverse)
but that might only be viable in a few contexts/situations.

Another issue (when being able to read FA files some day) will be
that files saved from the PC tend to become bytestream type (depends
on the very PC program behaviour re rewrite vs purge/rewrite again).

The 3000 can handle bytestream files quite nicely. PC or Unix on the
other hand have their fair share of trouble understanding concepts like
record-oriented file types (at the file-system level at least).

>Sam "hoping I don't have to read all that source code" Calvert

I hope, reading the new ReadMe file will be anough pain already...
By the way, I hope that I did not forget the FA-file-type issue in
the ReadMe file. Well, at least I was not as detailed as with this
reply here...

Sorry, Lars.

ATOM RSS1 RSS2