Re: thread CPU usage monitor?

2015-03-26 Thread René J . V . Bertin
On Wednesday March 25 2015 20:53:17 Jason Mitchell wrote:

Instruments  Select Application  Multicore  Tread States ?
/Applications/Xcode.app/Contents/Applications/Instruments.app

Thanks, that might work. For the moment it doesn't allow me to attach to a 
running application, though, which makes it a bit cumbersome when not hunting 
down some issue that occurs during startup but rather something with the idle 
behaviour after doing a substantial amount of work (in the app).

R.
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
On Thursday March 26 2015 11:22:44 Brandon Allbery wrote:

 It shouldn't be using :0.0 but instead the launchd socket. Or have you
 improved that? In which case you get to debug your own breakage.

I have done nothing to improve that, this is what I get by launching X11 by 
hand. Period.

So according to you, I shouldn't be able to connect to foo.local:0.0 because 
that's not using a launchd socket either?

R
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread Chris Jones

On 26/03/15 17:46, René J.V. Bertin wrote:

On Thursday March 26 2015 11:22:44 Brandon Allbery wrote:


It shouldn't be using :0.0 but instead the launchd socket. Or have you
improved that? In which case you get to debug your own breakage.


I have done nothing to improve that, this is what I get by launching X11 by 
hand. Period.


No, wrong. Even if you choose to launh X11/Xquartz by hand, unless you 
have explicitly turned off launchd support your DISPLAY connection still 
would take the form


 echo $DISPLAY
/private/tmp/com.apple.launchd.PjdrVEAjIV/org.macports:0



So according to you, I shouldn't be able to connect to foo.local:0.0 because 
that's not using a launchd socket either?

R
  ___
Do not post admin requests to the list. They will be ignored.
X11-users mailing list  (x11-us...@lists.apple.com)
Help/Unsubscribe/Update your Subscription: 
https://lists.apple.com/mailman/options/x11-users/jonesc%40hep.phy.cam.ac.uk

This email sent to jon...@hep.phy.cam.ac.uk



___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread Brandon Allbery
On Thu, Mar 26, 2015 at 1:46 PM, René J.V. rjvber...@gmail.com wrote:

 So according to you, I shouldn't be able to connect to foo.local:0.0
 because that's not using a launchd socket either?


That's using TCP which is also disabled by default for security reasons,
although there's a checkbox to enable it.

-- 
brandon s allbery kf8nh   sine nomine associates
allber...@gmail.com  ballb...@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonadhttp://sinenomine.net
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
Well I'll be ...

I tried once more, launched XQuartz (2.7.4) without my own .xinitrc in place, 
noted it had taken :1.0 (which sometimes happens when a previous instance 
didn't exit cleanly), and then tried to use a fully specified DISPLAY 
(${HOST}.local:1.0) and port:xterm. Whatever happened with the :0.0 address 
(which resisted a reboot) apparently doesn't concern :1.0 because I could even 
move back my .xinitrc and get my full environment. (Which is nothing but 2 
xterms, a clock and a legacy ctwm window manager.)

R.
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread Brandon Allbery
On Thu, Mar 26, 2015 at 2:23 PM, René J.V. rjvber...@gmail.com wrote:

   So according to you, I shouldn't be able to connect to foo.local:0.0
   because that's not using a launchd socket either?

  That's using TCP which is also disabled by default for security reasons,
  although there's a checkbox to enable it.

 Erm, I was talking about outgoing connections. The server itself had no
 issues accepting


What does that have to do with it? You asked about a connect string and I
answered that the server will only accept connections based on it if it's
listening on TCP. (It does not recognize the local hostname. As an ancient
hack, clients (not servers) *may* recognize the special hostname unix and
use a local socket.)

-- 
brandon s allbery kf8nh   sine nomine associates
allber...@gmail.com  ballb...@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonadhttp://sinenomine.net
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
On Thursday March 26 2015 13:49:54 Brandon Allbery wrote:
 On Thu, Mar 26, 2015 at 1:46 PM, René J.V. rjvber...@gmail.com wrote:

Yes, I did remove the launchd plist, because I don't want that functionality. 
ANd that's always been perfectly supported in XQuartz at least. I did not yet 
do anything of the sort for port:xorg-server though.

  So according to you, I shouldn't be able to connect to foo.local:0.0
  because that's not using a launchd socket either?

 That's using TCP which is also disabled by default for security reasons,
 although there's a checkbox to enable it.

Erm, I was talking about outgoing connections. The server itself had no issues 
accepting connections from clients that were NOT using X libraries from 
MacPorts (i.e. /opt/X11/bin/xterm), providing more evidence that there is no 
requirement to use weird, untypable and unthinkable DISPLAY strings :)

R
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: thread CPU usage monitor?

2015-03-26 Thread James Linder

 On 27 Mar 2015, at 3:00 am, macports-users-requ...@lists.macosforge.org wrote:
 
 
 On Tuesday March 24 2015 15:30:19 Daniel J. Luke wrote:
 On Mar 24, 2015, at 1:59 PM, Mihai Moldovan io...@macports.org wrote:
 I don't know of any application for OS X to get this kind of information
 out of the box, sorry.
 ps -M ?
 
 Instruments  Select Application  Multicore  Tread States ?
 /Applications/Xcode.app/Contents/Applications/Instruments.app

God’s gift to Mac programming (sorry for the sarcasim) makes an app that gives 
all the detail. Since the motherboards are based on standard boards maybe he 
discovered how to read the IIC stream. In any event it is
TemperatureMonitor.dmg

James
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
On Thursday March 26 2015 13:10:31 Jeremy Huddleston Sequoia wrote:

 Well, if you remove the launchd.plist, then you're basically stuck with the 
 Tiger-era behavior of having to either launch your apps from X11's 
 Applications menu (or from an xterm launched from that menu) because 
 Terminal.app has no way of knowing what your DISPLAY environment variable 
 should be set to.

Yes, I launch X applications either
- through my .xinitrc
- from an xterm (strangely enough, that remains my favourite terminal emulator 
on OS X; something with the text selection and fast copy/paste)
- from xfce4-panel

 It looks like you just manually set it to :0.0 in your ~/.profile (or 
 similar), and that's a good first guess, but as you noticed, it can easily be 
 wrong.

No, I don't! I even go to some lengths to unset DISPLAY under certain 
conditions. So I'm not guessing what my DISPLAY variable is, esp. since I 
*know* that it sometimes is :1.0 or even :2.0 or higher. I really believe that 
XQuartz.app (or X11.app) or xinit set DISPLAY to something traditional unless 
told otherwise. I also think that there must be some resource-lock that ensures 
that if you launch a 2nd X server (or if a previous instance didn't terminate 
gracefully) it will adopt a higher display number.

In the situation that I reported, my $DISPLAY was really :0.0 as verified by 
starting /opt/X11/bin/xterm through the X11 Application menu. I just couldn't 
connect to it through the X11 client libraries in MacPorts, only through those 
in /opt/X11 .

R.
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
On Thursday March 26 2015 15:18:20 Jeremy Huddleston Sequoia wrote:

  In the situation that I reported, my $DISPLAY was really :0.0 as verified 
  by starting /opt/X11/bin/xterm through the X11 Application menu. I just 
  couldn't connect to it through the X11 client libraries in MacPorts, only 
  through those in /opt/X11 .
 
 That is quite bizarre.  I'm able to run 'DISPLAY=:0.0 /opt/local/bin/xterm' 
 without issue.

I take it you're not running 10.6 in a VM hosted by Parallels Desktop ?

Anyway, bizarre it was, and even more bizarre that the disappeared after 
another reboot *and* after I had to send a SIGKILL to a server instance (which 
gave me DISPLAY=:1.0). I'll report back if it starts happening again.

R
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: thread CPU usage monitor?

2015-03-26 Thread Jason Mitchell
On Thu, 26 Mar 2015 12:25:01 +0100, Ren? J.V. Bertin wrote:
  On Wednesday March 25 2015 20:53:17 Jason Mitchell wrote:
 Instruments  Select Application  Multicore  Tread States ?
 /Applications/Xcode.app/Contents/Applications/Instruments.app
 
 Thanks, that might work. For the moment it doesn't allow me to attach
 to a running application, though, which makes it a bit cumbersome when
 not hunting down some issue that occurs during startup but rather
 something with the idle behaviour after doing a substantial amount
 of work (in the app).

Perhaps you could use dtruss/dtrace, following children/threads, and
compute what you need from relative on-cpu timestamps?  Check '$ man -k
dtrace' to see if there's an 80% solution available to specialize.
It'll also probably take a little log parsing after collection too.
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


Re: XQuartz on 10.6 (in a VM)

2015-03-26 Thread Craig Treleaven

At 3:44 PM +0100 3/26/15, René J.V. Bertin wrote:
I've got a 10.6 VM which is more or less a clone 
of my 10.6.8 system of about 18 months ago. I'm 
pretty sure that it used to run my X11 
environment, and the same host (Parallels 
Desktop 10) also runs X11 applications just fine 
when I boot it into OS X 10.9 .


Not so (anymore) with 10.6.8 though. There 
appears to be some communications glitch in the 
Xt toolkit, which causes X applications built 
against the X11 libraries in MacPorts to fail 
with Xt toolkit error: can't connect to display 
:0.0.


I have XQuartz 2.7.4 installed, because for some 
(related?) reason the 2.7.7 installer remains 
stuck on Install for all users on this system 
with the proceed button greyed out. However, the 
latest port:xorg-server from MacPorts shows the 
exact same behaviour (the Xt error, not the 
installer issue :)).


Does any of this sound familiar?


Maybe it is a Parallels 10 thing?  I use 
Parallels 9 with a 10.6.8 Server VM, 10.6.8 host, 
and MacPorts xorg-server works fine for me.  This 
is recent, as I was working the last few days 
with an update to hdhomerun_gui, which runs under 
XQuartz.


Donno if this helps...

Craig
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users


XQuartz on 10.6 (in a VM)

2015-03-26 Thread René J . V . Bertin
Hello,

I've got a 10.6 VM which is more or less a clone of my 10.6.8 system of about 
18 months ago. I'm pretty sure that it used to run my X11 environment, and the 
same host (Parallels Desktop 10) also runs X11 applications just fine when I 
boot it into OS X 10.9 .

Not so (anymore) with 10.6.8 though. There appears to be some communications 
glitch in the Xt toolkit, which causes X applications built against the X11 
libraries in MacPorts to fail with Xt toolkit error: can't connect to display 
:0.0.

I have XQuartz 2.7.4 installed, because for some (related?) reason the 2.7.7 
installer remains stuck on Install for all users on this system with the 
proceed button greyed out. However, the latest port:xorg-server from MacPorts 
shows the exact same behaviour (the Xt error, not the installer issue :)).

Does any of this sound familiar?

Thanks,
René
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users