Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ptp-user] re: Changing connection name for a remote project

> This still seems a little complicated and a subtle aspect of moving to a
> different host. Is there any way to make this easier to change in RDT? Is
> the hostname outside RDT's scope, where RDT only understands connection
> names?

The core of RDT is connection agnostic. All it knows is whether a given service provider is configured or not. It's up to the service provider to determine what it means to be configured. In the case of the default service provider implementations we've supplied, it generally means you've picked a connection.

Mucking with the setup of a given connection is generally under the domain of whatever tooling is providing the connection (in this case, RSE), because we're trying to be as generic as possible, and not assume anything about how a given provider works.

> Is indexing information for a project cached
> and updated only when files are created, deleted or changed?

Correct. If you do switch machines, you'll have to request a reindex, because as far as Eclipse is concerned, your resources haven't changed, so nothing should need to be reindexed.

===========================
Chris Recoskie
Team Lead, IBM CDT and RDT
IBM Toronto
Inactive hide details for Dave Wootton <dwootton@xxxxxxxxxx>Dave Wootton <dwootton@xxxxxxxxxx>


          Dave Wootton <dwootton@xxxxxxxxxx>
          Sent by: ptp-user-bounces@xxxxxxxxxxx

          01/23/2009 10:47 AM

          Please respond to
          PTP User list <ptp-user@xxxxxxxxxxx>

To

PTP User list <ptp-user@xxxxxxxxxxx>

cc

ptp-user@xxxxxxxxxxx, ptp-user-bounces@xxxxxxxxxxx

Subject

Re: [ptp-user] re: Changing connection name for a remote project

Chris
If I change the hostname in the connection properties in the remote system
view, then I can open the project on the second system. This seems to work
if I leave the project open while I disconnect, change the hostname in the
connection, then connect to the second system, but I didn't test this
extensively.

I didn't realize that I could modify the connection's hostname until you
mentioned it. I had made the association between connection name and
hostname, thinking they were fixed, possibly since when I first create a
connection and fill in the hostname, the connection name automatically
fillls in with hostname.

This still seems  a little complicated and a subtle aspect of moving to a
different host. Is there any way to make this easier to change in RDT? Is
the hostname outside RDT's scope, where RDT only understands connection
names?

I think there is a problem if a header file exists in different locations
on the two systems then definitions are only fond in the header where they
exist on the first system. For instance, the mpi.h header exists in
/usr/lpp/ppe.poe/include on AIX and /opt/ibmhpc/ppe.poe/include on Linux.
If I create the project on AIX, then I seem to be able to find the
definitions for types like MPI_Comm. If I move to the Linux system, I
don't find the definition. Is indexing information for a project cached
and updated only when files are created, deleted or changed?
Dave



Chris Recoskie <recoskie@xxxxxxxxxx>
Sent by: ptp-user-bounces@xxxxxxxxxxx
01/22/2009 02:25 PM
Please respond to
PTP User list <ptp-user@xxxxxxxxxxx>


To
ptp-user@xxxxxxxxxxx
cc

Subject
[ptp-user] re: Changing connection name for a remote project






Sorry for any tardy replies I owe you Dave, I only just subscribed to this
list this morning.

> Is it possible and a good idea to provide the ability to change the
> connection name for a remote project after the remote project has been
> created? I have the source code in a directory that is shared between
> multiple systems (by NFS). Our lab development environment is also set
up
> with directories shared by multiple systems.

Right now RSE doesn't give us notification when a connection is renamed,
which means RDT can't react to it.

> The reason I'm wondering if this is worthwhile is that this source code
> could be common source code for AIX and/or multiple Linux versions where

> depending what I'm doing, I may want to compile that code on AIX or
Linux
> systems.

We have thought of this... which is why the service model allows for
multiple service configurations per project.  However there is no UI yet
to allow you to do this.

>
> Also, if I'm running in a cluster environment like this, the initial
> system I specified when creating the project may be down or I want to
use
> a different system for performance reasons.
>

Ok... then change the machine the connection points to, but leave the
connection name as it is, and it should work.

> Is this possible for a closed project or one where the defined
connection
> is not connected? Is there sufficient info in the local Eclipse dtaa to
> make this work, or do you depend on what's on the remote system that you

> can't get to?
> Dave

As long as you keep the connection name in RSE the same, you should be
able to muck with the actual connection parameters, regardless of the
open/closed state of the project, and it should work.

===========================
Chris Recoskie
Team Lead, IBM CDT and RDT
IBM Toronto_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user


_______________________________________________
ptp-user mailing list
ptp-user@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ptp-user

GIF image

GIF image

GIF image


Back to the top