<br><tt><font size=2>Antoine Martin <antoine@nagafix.co.uk> wrote
on 07/04/2010 22:40:00:<br>
<br>
[snip]<br>
</font></tt>
<br><tt><font size=2>> > To emulate synchronous launching, a delay
must be introduced.<br>
> Well spotted! I had hit this bug too.<br>
> Generally happened when I was doing a well rehearsed demo.. ;)<br>
> <br>
> The solution that I have used successfully (and is also used in neatx)<br>
> is to watch the output of the nxagent process: once the line saying<br>
> "Waiting for connection from ..." shows up, you can be certain
that the<br>
> socket is ready.<br>
> I have also previously used a tcp-connect loop (with a 1 second<br>
> timeout), but the log parsing is both quicker and more proper.<br>
> <br>
> The "sleep 4" that you have added may work in most cases,</font></tt>
<br>
<br><tt><font size=2>It will work unless the machine is maxed out . . .</font></tt>
<br>
<br><tt><font size=2>> but you can't<br>
> guarantee that it is the correct delay.</font></tt>
<br>
<br><tt><font size=2>No ! it will only do for about 99.99% of first logins
after a reboot.</font></tt>
<br>
<br><tt><font size=2>> Too short and it causes the<br>
> aforementioned bug, too long is also problematic as it slows down<br>
> session startup unnecessarily. (4 extra seconds is pretty long when
you<br>
> don't actually need it in most cases)</font></tt>
<br>
<br><tt><font size=2>No one will notice !!</font></tt>
<br><tt><font size=2><br>
</font></tt>
<br><tt><font size=2><Snip></font></tt>
<br>
<br>
<br><tt><font size=2>> <br>
> My 2p<br>
> Cheers<br>
> Antoine<br>
> <br>
</font></tt>
<br>
<br><tt><font size=2>So we differ.</font></tt>
<br>
<br><tt><font size=2>Here is my view.</font></tt>
<br>
<br><tt><font size=2>I am concerned with live systems staying live.</font></tt>
<br>
<br><tt><font size=2>I am unconcerned with log in times of 8 seconds (
which</font></tt>
<br><tt><font size=2>is the time my machines take ).</font></tt>
<br>
<br><tt><font size=2>I want a fix that can be understood by people who
can't solve</font></tt>
<br><tt><font size=2>the problem for themselves, that is to say ordinary
users.</font></tt>
<br>
<br><tt><font size=2>I want something that they can type in easily without
making</font></tt>
<br><tt><font size=2>mistakes.</font></tt>
<br>
<br><tt><font size=2>I want something which will will be benign in the
event a </font></tt>
<br><tt><font size=2>mistake is make ie. no unterminated loops or "</font></tt>
<br>
<br>
<br><tt><font size=2>I want minimum intervention patching, following the
coding </font></tt>
<br><tt><font size=2>style used by the original developer, right or wrong
(which</font></tt>
<br><tt><font size=2>is of course standard practice throughout the industry).</font></tt>
<br>
<br><tt><font size=2>More elegant and quicker solutions are for the maintainers</font></tt>
<br><tt><font size=2>to include for thorough testing in the next version,
after a</font></tt>
<br><tt><font size=2>complete review of the surrounding code and associated</font></tt>
<br><tt><font size=2>algorithms.</font></tt>
<br>
<br>
<br><tt><font size=2>Your solution is to write a new product, which is
perfectly</font></tt>
<br><tt><font size=2>valid and of course I wish you good luck.</font></tt>
<br>
<br><tt><font size=2>But that won't help an overworked and underpaid systems
</font></tt>
<br><tt><font size=2>manager supporting for a charity, who is getting flak
about a </font></tt>
<br><tt><font size=2>system which appears cronky or badly configured and
is </font></tt>
<br><tt><font size=2>giving a shaky "user experience" to that
senior manager </font></tt>
<br><tt><font size=2>who is always first in the office on the monday after
reboot</font></tt>
<br><tt><font size=2>sunday, and views things like this as a test of competence</font></tt>
<br><tt><font size=2>or feels he needs to push people along . . .</font></tt>
<br><tt><font size=2><br>
</font></tt>
<br>
<br><tt><font size=2>Antoine, did you send me an out of band email a bit
back</font></tt>
<br><tt><font size=2>asking about printing ??</font></tt>
<br>
<br><tt><font size=2>There was a thread going about cups re-direction for
a</font></tt>
<br><tt><font size=2>bit, but I've been a bit too busy since December to
bother</font></tt>
<br><tt><font size=2>with anything else, and NX printing is a bit of a
convoluted</font></tt>
<br><tt><font size=2>pass the parcel.</font></tt>
<br>
<br><tt><font size=2>If you are writing a new application with a new CLIENT</font></tt>
<br><tt><font size=2>then you can do what you want and don't have to follow</font></tt>
<br><tt><font size=2>the nomachine client.</font></tt>
<br>
<br><tt><font size=2>Looking back at the thread though , my advice is still</font></tt>
<br><tt><font size=2>to run the userland cupsd on the NX server, so all
sessions</font></tt>
<br><tt><font size=2>are the same regardless, and to spool postcript as
far</font></tt>
<br><tt><font size=2>along the line as you can if you are using a non postscript</font></tt>
<br><tt><font size=2>printer rather than having the .ppd on the NX server</font></tt>
<br><tt><font size=2>and spooling raster down a remote link.</font></tt>
<br>
<br>
<br><tt><font size=2>Whatever you do . . .</font></tt>
<br>
<br><tt><font size=2>. . . the main userbase will stay with FreeNX, because</font></tt>
<br><tt><font size=2>it works OK, and almost no organisations will gain
</font></tt>
<br><tt><font size=2>overall benefit from the cost of a project to move</font></tt>
<br><tt><font size=2>to one of the other new nxagent launchers </font></tt>
<br><tt><font size=2>( whichis what FreeNX is, though it does also aim</font></tt>
<br><tt><font size=2> to pass ipp and samba printing and samba drive</font></tt>
<br><tt><font size=2>mapping over the link).</font></tt>
<br>
<br>
<br><tt><font size=2>I don't see there is much demand for remote printing</font></tt>
<br><tt><font size=2>especially to windows work stations, though I use
it </font></tt>
<br><tt><font size=2>quite a bit to put ooo-writer on remote XP with local</font></tt>
<br><tt><font size=2>USB printers and mapped drives.</font></tt>
<br>
<br>
<br><tt><font size=2>The truisms are :-</font></tt>
<br>
<br><tt><font size=2>99% of the work on any of these projects has been</font></tt>
<br><tt><font size=2>done by Nomachine, building and maintaining the </font></tt>
<br><tt><font size=2>libraries, nxagent</font></tt>
<br>
<br><tt><font size=2>and </font></tt>
<br>
<br><tt><font size=2>the client software, which deals with all the caching</font></tt>
<br><tt><font size=2>on the workstation and provides an X server for </font></tt>
<br><tt><font size=2>windows.</font></tt>
<br>
<br><tt><font size=2>In adition there is ssh/sshd doing the encrypted link.</font></tt>
<br>
<br><tt><font size=2>All the other stuff runs once at login, again at logout</font></tt>
<br><tt><font size=2>and is irrelevant during the operation session.</font></tt>
<br>
<br><tt><font size=2>cb<br>
<br>
> > <br>
__<br>
> -----BEGIN PGP SIGNATURE-----<br>
> Version: GnuPG v1.4.9 (GNU/Linux)<br>
> Comment: Using GnuPG with Mozilla - </font></tt><a href=http://enigmail.mozdev.org/><tt><font size=2>http://enigmail.mozdev.org</font></tt></a><tt><font size=2><br>
> <br>
> iEYEAREKAAYFAku8+68ACgkQGK2zHPGK1rtW9wCfezKY/yCoTMpOtQ2d55DJZx0l<br>
> 758An3on5jn6B4IRGitCPDVAdDKl4j4N<br>
> =U1L/<br>
> -----END PGP SIGNATURE-----<br>
> ________________________________________________________________<br>
> Were you helped on this list with your FreeNX
problem?<br>
> Then please write up the solution in the FreeNX Wiki/FAQ:<br>
> <br>
> </font></tt><a href="http://openfacts2.berlios.de/wikien/index.php/BerliosProject:FreeNX_-_FAQ"><tt><font size=2>http://openfacts2.berlios.de/wikien/index.php/BerliosProject:FreeNX_-_FAQ</font></tt></a><tt><font size=2><br>
> <br>
> Don't forget to check the NX Knowledge
Base:<br>
> </font></tt><a href=http://www.nomachine.com/kb/><tt><font size=2>http://www.nomachine.com/kb/</font></tt></a><tt><font size=2>
<br>
> <br>
> ________________________________________________________________<br>
> FreeNX-kNX mailing list --- FreeNX-kNX@kde.org<br>
> </font></tt><a href="https://mail.kde.org/mailman/listinfo/freenx-knx"><tt><font size=2>https://mail.kde.org/mailman/listinfo/freenx-knx</font></tt></a><tt><font size=2><br>
> ________________________________________________________________<br>
</font></tt>