| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 517.1 | $ set prot=(w:r) *.* | KIRK::LONG | | Thu May 28 1987 17:02 | 1 |
| Can you set the set the protection so we can get to it ;^}
|
| 517.2 | OOPS! | SQM::JMSYNGE | James M Synge, VMS Performance Anal. | Thu May 28 1987 21:53 | 3 |
|
Boy! Thought I had the protection set up automatically :-)
Its set now.
|
| 517.3 | No it isn't | MVSUPP::RYMER | Stealthy is Healthy | Fri May 29 1987 08:43 | 12 |
|
Er .... I don't think it is set.
By the way ....
> comp.sys.amiga
What is it (Sorry to sound green, but I am). Is it a public domain
database? .... How to I get in?
Andy
|
| 517.4 | | KIRK::LONG | | Fri May 29 1987 13:11 | 17 |
| Agree with .3, not fixed yet.
RE .3:
> What is it (Sorry to sound green, but I am). Is it a public domain
> database? .... How to I get in?
Do an "add entry ROLL::USENET", the first few notes tell you how to
subscribe to a daily mailing of any number of "not-necessarily DEC"
newsgroups. The Amiga newsgroups give you access to engineers at CBM and many
others developers ( instead of the limited 100000+ people on our internal
network :^) There is also good PD software available in the postings that will
keep your plastic from going up in flames at your local software shop.
Warning: the postings are sometimes long and require a lot of spare time to
do justice to.
Dick
|
| 517.5 | | SQM::JMSYNGE | James M Synge, VMS Performance Anal. | Fri May 29 1987 15:34 | 10 |
|
I really need another account so I can log in and check to see
that the files I think are accessible really are.
I set the protection on the files to W:R, but forgot to set the
directories to W:RE.
re .-1: It takes a LOT of time to keep up with just comp.sys.amiga. An
average day has over 30 new messages, with a very high noise to
signal ratio.
|
| 517.6 | top dir too | HAZEL::MELLITZ | | Fri May 29 1987 19:59 | 2 |
| Don't for get the [000000] directory.
... Rich
|
| 517.7 | accessibility check | SAUTER::SAUTER | John Sauter | Fri May 29 1987 21:11 | 4 |
| A simple way to check for accessibility is to go through some
far-away node. E.g., COPY SAUTER::<remote file spec> *.
It's OK to put multiple node names in a file spec.
John Sauter
|
| 517.8 | | SQM::JMSYNGE | James M Synge, VMS Performance Anal. | Fri May 29 1987 21:51 | 9 |
|
Re .-1:
Good suggestion. Of course I have proxies set up from all sorts
of nodes, including my own, so I had to login into a node for which
I hadn't setup a proxy, then did a TYPE of the posting, and had
no problem. So I think everything is ok now. Sorry for tantalizing
you with the library, then not having the access setup correctly.
James
|
| 517.9 | | COOKIE::WITHERS | Le plus ca change... | Fri May 29 1987 22:53 | 6 |
| COPY 0::disk:[directory-string]files.extentions.gen []
will also check for file accessibility and all you need to do is
go as far as your local router...
BobW
|
| 517.10 | | SQM::JMSYNGE | James M Synge, VMS Performance Anal. | Sat May 30 1987 00:55 | 15 |
|
Bug report: I've already found a bug in the code, so I'll make the
change in the shar file, and I'll also place the problem file,
Protect.i, in the directory specified in .0 Don't try to run the
current version.
Re .-1
That almost works, but I have a proxy set up so that I can perform
network operations in the cluster from any node in the cluster.
COPY 0""::disk:[directory-string]file.extention;ver []
will do what you want. The 0"" specifies this node, using the
default DECnet account (Usually user DECNET).
|
| 517.11 | Removing libraries? | XSNAKE::WILSONTL | Lead Trumpet (Read that...LEED!) | Sun Dec 29 1991 16:57 | 8 |
| I checked the notes for libraries and checked in Aquarium for this
inforation and couldn't find it. Is there a utility out there that
will remove unused libraries from memory? After running several PD
things this morning, I discovered (using ARTM) that memory was very
cluttered with libraries that were no longer in use. I used ARTM to
remove the libraries one at a time.
Tony
|
| 517.12 | | STAR::GUINEAU | it's nothing personal | Sun Dec 29 1991 21:42 | 11 |
| >
> inforation and couldn't find it. Is there a utility out there that
> will remove unused libraries from memory? After running several PD
If you're running WB 2.0 you can type AVAIL FLUSH in a shell window.
With WB 1.3 (and before), you can load the workbench with the DEBUG
option (in startup-sequence, "loadwb -debug" I believe) then a menu
item will show up (DEBUG) which has a FLUSHLIBS option.
john
|
| 517.13 | Thank you, sir... | HSSTPT::WILSONTL | Lead Trumpet (Read that...LEED!) | Mon Dec 30 1991 11:22 | 0 |
| 517.14 | | TENAYA::MWM | | Tue Dec 31 1991 17:30 | 7 |
| Just curious - why do you care about the memory clutter? The first time an
allocation fails, all those libraries should be expunged and the allocation
should be retried. Of course, some of them may be for tied down by things
that don't properly close them, but that could make for a dangerous
situation.
<mike
|
| 517.15 | Mildy curious myself... | XSNAKE::WILSONTL | Lead Trumpet (Read that...LEED!) | Thu Jan 02 1992 16:34 | 10 |
| RE: .14 - No particular reason. I only have three megs right now and
was just wondering. I discovered that my memory kept shrinking and
assumed that all these libraries were using it up. I tested the theory
and found it to be the most probable culprit.
What I'm understanding from your reply is that when the system can't
allocate memory, it blows away unused libraries and tries again. 'Zat
true? I haven't delved very deeply into the internals yet.
Tony
|
| 517.16 | | TENAYA::MWM | | Thu Jan 02 1992 18:16 | 14 |
| RE .15
You have it exactly right. This is the case for libraries, devices, and fonts.
I believe it is the case for all shared resources: the system leaves them
in memory even when unused, so that you don't have to reload them from
wherever they came from if they are opened again. If a request for memory
fails, then the system asks these things to remove themselves and then
retries the allocation.
I believe that "avail flush" does the same thing as the system does after
an allocation failure. It may actually find out how much memory is available,
then ask for more than that to force an allocation failure.
<mike
|