On 21 Nov 2018, at 17:21, CV Bruce wrote:Enrico, I’m a little concerned that installing anything in ~/… will make it user specific, and not system wide. Since only one copy of rxapi can run at a time, then only one user can use ooRexx. In your explorations can you test for thi
Hi Bruce, please add any additional info to the existing ticket.
CV Bruce schrieb am Mi., 21. Nov. 2018 22:08:
> Eric,
>
> I’ve been playing with rxqueue from the current repository. It doesn’t
> work. I notice that there are already open tickets for rxqueue(), and the
> root cause for both is
Eric,
I’ve been playing with rxqueue from the current repository. It doesn’t work.
I notice that there are already open tickets for rxqueue(), and the root cause
for both is probably the same. Do you want me to open a new ticket or to add
info to the existing tickets.
Yours,
Bruce
___
could we, for the moment, just have an environment variable that overrides the
default? Meaning that when it is not there, we write the PID to standard; when
it is there, we write the PID where it indicates we want it to be. That should
satisfy worries over changing this?
René.
> On 21 Nov 201
my understanding was that all instances connect to a running daemon and have
separated user spaces in it. Running separate processes would be possible only
when making the interpreter instances connect to different ports.
When testing the s390x port at a an IBM (Marist) Linux VM that I could not
I may not have been entirely accurate. I believe that two rxapi process can run
at the same time as long as they have different PID files. The problem is that
only one of them can bind the socket. Do we know if rxapi issues an error
message if it can not bind the socket? If so where does it s
I agree with this. What I’ve seen before is at install time the user is asked
whether they want to install for just the user or for all users. Selecting all
users then requests authentication for the install.
Bruce
> On Nov 21, 2018, at 8:39 AM, René Jansen wrote:
>
> User specific is fine
Greetings Rick,
Rick McGuire wrote:
It's not really that simple. I believe writing the pid file at that location is part of the requirements for running as a daemon on other unix-based systems. I wish David Ashley were still around, he
did all of that work and would know the issues.
Ah yes, t
User specific is fine and is what will suffice for most users. It will be the
default for a portable install. Multiuser installs should be an option and
there is no issue in requiring admin for those, I’d rather they do.
René.
> On 21 Nov 2018, at 12:21, CV Bruce wrote:
>
> Enrico,
>
> I’
> On 21 Nov 2018, at 17:11, Rick McGuire wrote:
>
> It might not be true for the Mac, but I believe it is true for Linux, which
> means just changing that line is not the correct behavior.
As I said It was just a VERY quick and dirty test…
The ooRexx.pid name/location could be a build opti
Enrico,
I’m a little concerned that installing anything in ~/… will make it user
specific, and not system wide. Since only one copy of rxapi can run at a time,
then only one user can use ooRexx. In your explorations can you test for this?
"The main difference [on Mac OS] is that an agent
It might not be true for the Mac, but I believe it is true for Linux, which
means just changing that line is not the correct behavior.
I just discovered that the source tree has a Mac version of APIServer.cpp
that uses a tmp file, but the build is unconditionally using the Linux
version. Not sure
> On 21 Nov 2018, at 15:23, Rick McGuire wrote:
>
> I believe writing the pid file at that location is part of the requirements
> for running as a daemon on other unix-based systems.
Does not seem true for Mac OS
I have been running and ooRexx-ing for a while with /tmp/ooRexx.pid
and ever
Noted, if we got it in all package managers we would be set.
I thought that Suse is waiting for the green light (GA release).
We did not have stellar results from previous attempts to get a certificate,
but I agree we need it and will reatart the effort.
best regards,
René.
> On 21 Nov 2018, a
I’m more of a MacPorts kind of guy, and I’ve used fink in the past. It would
be nice if we could support more of the Mac OS specific package managers.
> On Nov 21, 2018, at 7:07 AM, René Jansen wrote:
>
> I agree.
>
> Want to add to the discussion: homebrew (brew) has a ‘brew services’ comma
I agree.
Want to add to the discussion: homebrew (brew) has a ‘brew services’ command
that can install a service on macOS without requiring sudo.
A portable ‘installer’ (drag an application icon from a .dmg to ~/Applications)
would be a good thing. A ‘brew install oorexx’ would be a good thing -
Well….
I agree with Enrico, mostly.
Installation: Yes you need sudo privileges to write to /Library/LaunchDaemons.
You also need sudo to write to /usr/local/… which is as far as I know, the
current place to put system wide user programs. I could be wrong as I haven’t
been keeping up.
Execu
There are some minor changes to the code that would allow ooRexx to run, on Mac
OS, from any folder it is copied to without recompiling. There are some
special library load paths that inform Mac OS of the relative location of
shared libraries. Because Mac OS by default searches some libraries
It's not really that simple. I believe writing the pid file at that
location is part of the requirements for running as a daemon on other
unix-based systems. I wish David Ashley were still around, he did all of
that work and would know the issues.
Rick
On Wed, Nov 21, 2018 at 9:14 AM Enrico Soric
Grrr, didn't get to the undo button in time...
On Wed, Nov 21, 2018 at 9:07 AM Rick McGuire wrote:
> I have started doing a little research on this, and I see how to implement
> this on Windows, but I'm not sure the existing client/server set up will
> work with *ix named pipes. The rxapi proce
The only reason for having sudo privileges is to install the plist into
/Library/LaunchDaemons
The only reason for having sudo privileges when starting by hand the rxapi
thing is to write /var/run/ooRexx.pid
I just run a VERY quick and dirty test changing
rexxapi/server/platform/unix/linux/AP
I have started doing a little research on this, and I see how to implement
this on Windows, but I'm not sure the existing client/server set up will
work with *ix named pipes. The rxapi process handles requests like this:
Main thread binds to a port
Main thread listens for an inbound connection, wh
Rather discussing general programming questions in bugs where the discussion
results evaporate once
the item gets closed, I think it makes more sense to discuss them here in the
developer list.
In the past weeks/months P.O. has started and has been doing an immensive
amount of work in learning
23 matches
Mail list logo