Let's stir up some discussion...

When you're working with your system, you'll undoubtedly hit some
point where things become difficult, and you say to yourself "I just
wish I could do ...this". I hit that point every day, whether with MV
environments, .NET, PHP, or any of the tools or products I work with.

I know I can find or write tools to get over the obstacles, or maybe I
just need to read more documentation. But I'd rather not have that
bump in the road in the first place. Why is that obstacle there? Why
can't I just do ...this?

Do I take the time to find or create a solution?
Do I avoid the problem and make do with some other approach?
How long do I keep avoiding the same problems?
Does it cost less (in time or money) to (a) live with the problem, (b)
research a DIY solution, or (c) just buy someone else's solution.

So what problems stump you?

What do your customers and prospects ask you for that you never quite
get to because you're not really sure how to make it work? You might
think some solution is rocket science when it's not, or that it's
costly or unattainable when it's free and easy to download.

Do you have people coding for you who keep hitting some common
stumbling block?

Is there some procedure that needs to be fixed "one of these days"
with some kind of automation? Is there some device that you want to
get rid of, or make better use of, in conjunction with your app?

Is there something you saw that's really cool and you want to hook
into it, but you just don't know how? Maybe something your customers
would pay you for, if only you could make it work?

Ultimately where I'm going with this is that a lot of you solve
business problems with the tools we have, but I and others here spend
time solving more technical or procedural problems with the tools
themselves. There might be solutions to the things that bug you.
Someone here might have something they use every day to fix problems
you face every day. I might have written some utility and discarded
it years ago that someone here would love to use. I or others might
write something new if we perceive adequate demand.

Let's solve some problems here ... I have to believe there's at least
one out there.

Tony Gravagno
Nebula Research and Development
TG@ remove.pleaseNebula-RnD.com
Nebula R&D sells mv.NET and other Pick/MultiValue products
worldwide, and provides related development services
remove.pleaseNebula-RnD.com/blog
Visit PickWiki.com! Contribute!
http://Twitter.com/TonyGravagno

"Tony Gravagno" <nospam (AT) nospam (DOT) invalid> wrote


A seamless connection from Ruby on Rails to Pick (like)
environment via Active Record would be nice. =`:^>

--
frosty

On 2010-12-18 18:38:02 -0500, Tony Gravagno <nospam (AT) nospam (DOT) invalid> said:

I knew there was going to be a pitch in there somewhere.

--
Kevin Powick

On Dec 19, 4:55*am, "frosty" <fros... (AT) bogus (DOT) invalid> wrote:
If we are going to solve problems effectively, it would be best to
describe the problem rather than give a specific solution. For
example, a while back we had a QM user who said they needed ODBC which
would not have been cost effective. Once they described the actual
problem, we pointed them at a solution that already existed.

Please let's see this thread develop into a wish list of the issues
that cause developers the headaches that Tony describes.

Martin Phillips, Ladybridge Systems

I wish the d3/nt platform had the system id available
at the tcl prompt like the other platforms do. I mean
yes, you can manually create an item for the SYSID verb to
see, but that only helps my shop, not support clients.
{And how can I be sure it isn't an old one that hitchiked
in when they migrated their database from an older
platform?} << Oooh! When you hand-edit that item,
just put remarks on additional attributes!>>
Now I'm curious. I'll have to see if it is visible in the
window registry, and if any windows command-line tools
can reveal it, if so...

I wish doing a shelled-out command in d3/nt
wasn't such a landmine -- if the command tries
to wait for user input, your line is hung. Maybe
it would take a sabotaged version of the windows
executable cmd.exe to achieve this {do !foo.exe stuff
at tcl prompt, where foo.exe is that sabotaged exe
that is sortof like cmd.exe that always returns
end of file on input, or whatever...}

Remember you can create foo.exe commands
using the compilers that come free with cygwin package...
Just put any supporting .dll files in same windows
folder, or somewhere on the PATH environ. variable.
If you leave out a .dll, at run time it stops and tells you
which one it needs.

Or if I could get Logoff nnn to logoff user on
pick port nnn even though that user is shelled out,
that would make me happy, too. {Or give me
some clue what process id it is so I can go kill off
the shelled out subprocess myself, then I can do
regular Logoff verb to 'tidy up' if that helps any...}

On 2010-12-19 07:36:31 -0500, "Frank Winans" <fwinans (AT) sbcglobal (DOT) net> said:

Have you not tried the "k" option?

logoff nn (k

--
Kevin Powick

On Dec 20, 12:50*am, Kevin Powick <nos... (AT) spamless (DOT) com> wrote:
Are you trying to tell me that we can ALREADY do everything we were
ever likely to want to do from Pick already? Whodathoughthat?!?

Merry Christmas

"Frank Winans" wrote:

In the rest of the world, once software is available, whether
commercial or FOSS, it's understood that everyone must go get tools
that they want. In the absence of something like CPAN for D3 (a
variant of which is available for QM), loading software to a client
system is currently a manual process. Sure, to satisfy your specific
request, someone can create a little code to retrieve the sysid, and
then you can load it wherever you want, depending on cost, licensing,
etc. I think that's a simple request and someone here may do it.

But let's expand your request a bit. You want some functionality and
you don't want to have to install it to all of your clients manually
once you get it. I've been advertising tools that do this kind of
thing for years and no one seemed interested. So maybe it's time to
revisit that.

Thoughts?

I know I've done this from D3 before, but if I were in a hurry I would
do something like this:

EXECUTE "!foo " CAPTURING OUT

The "foo" command will be a command-line directive to dump a specific
registry key.

HTH
T

"Frank Winans" wrote:

What you're talking about has nothing to do with D3NT really. This is
a common situation known as synchronous blocking that can occur with
processes or network connections. If you have a web client that
connects to a web server and that server tries to do something that
evades a timeout, the web client is simply hung until the user clicks
away. Timeouts are set at the browser, web server, and other tiers to
prevent this. But at the OS level if you fork or invoke some other
process and that process just sits there, the calling process isn't
going to release it unless there is special logic in place to do that.

So... you're looking for a way to timeout an OS request. The way to
do that is to start a utility that you can control, that has an
asynchronous timer on it in another thread. That utility executes the
thing that you want to do. When the timeout occurs your process
should wrapup, regardless of what the underlying process is doing.
"Someone" needs to do that. Since this functionality isn't built into
any DBMS, it's up to app developers to do this.


That's what I was describing above. It's not a sabotaged cmd.exe but
just a normal exe that can release resources as soon as it perceives a
hang or timeout.


That's severe overkill. You can do this with just about any
programming language in Windows. You don't need a virtual OS like
cygwin.


Kevin commented on that. I think the best solution is to avoid the
problem, rather than to brutally reset processes.

HTH
T