T.R | Title | User | Personal Name | Date | Lines |
---|
1108.1 | KP7, etc. | POOL::HALLYB | The Smart Money was on Goliath | Fri Jul 14 1989 16:56 | 4 |
| See 4TRACK::DECWINDOWS_DOCUMENTATION note 196. I don't get it either.
John
|
1108.2 | ... | ORACLE::WATERS | | Sun Jul 16 1989 20:07 | 6 |
| >> See 4TRACK::DECWINDOWS_DOCUMENTATION note 196. I don't get it either.
Btw, note 196.* does not answer the question. Just points out the
lack of explanatory documentation. So can someone in this conference
describe the purpose and benefits of using multiple application_contexts?
|
1108.3 | Is this a start? | EPIK::BUEHLER | An education teaches you why you need one | Mon Jul 17 1989 12:22 | 6 |
| I don't know why you'd create multiple application contexts, but you
would create one application context if you wanted display connections
to more than one machine and you also wanted to use the toolkit.
John
|
1108.4 | | KYOA::MIANO | O.K. so who cares about the METS? | Mon Jul 17 1989 13:23 | 21 |
| RE: <<< Note 1108.0 by ORACLE::WATERS >>>
> The VMS DECw docs don't mention "application context" objects.
> The related routines (XtCreateApplicationContext, XtAppAddInput,
> etc.) don't describe the purpose of the application_context argument
> either.
According to the MIT X documentation there are several "obsolete"
routines that are replaced by routines that use an application_context
argument. I have not found this mentioned anywhere in the DECwindow
documentation and most of the examples in the self-paced instruction
course use the obsolete routines.
I found out about the obsolete routines because I could not get
XtAddInput to work. I looked in the MIT documentation and saw that it
was listed a obsolete and had been replaced by XtApplAddInput. I then
changed the program to use the XtAppl routines and the program worked on
the first try.
John
|
1108.5 | | SDSVAX::SWEENEY | Honey, I iconified the kids | Mon Jul 17 1989 13:47 | 12 |
| XtAppAddInput is described on page 2-39 of VMS DECwindows Toolkit
Routines Reference Manual AA-MG23A-TE.
The matter of using non-App'ed intrinsic routines in VMS DECwindows
Guide to Application Programming probably has more to do with the
timing of X11R3 and DECwindows V1 field tests than a conscious decision
to use the obsolete routines. Would you have liked to see DECwindows
V1 delayed?
Is it too late to have the examples for the V2 documentation updated?
I don't know.
|
1108.6 | Maybe a little help? | EXLIB::DLUGOSZ | Never try to prove what nobody doubts | Mon Jul 17 1989 14:55 | 13 |
| The only thing that I've been able to scrounge up about application
contexts (other than the 'canned' line about them allowing multiple
logical aplications in a single address space and that every 'program'
needs at least one) is that you only need more than one if you are
dealing with multiple threads in one process (does this mean that
the server also has to be mutithreaded?).
Also the app context maintains a list of of the displays opened
by XtOpenDisplay.
Ron
|
1108.7 | | LEOVAX::TREGGIARI | | Mon Jul 24 1989 09:34 | 31 |
| > The only thing that I've been able to scrounge up about application
> contexts (other than the 'canned' line about them allowing multiple
> logical aplications in a single address space and that every 'program'
> needs at least one) is that you only need more than one if you are
> dealing with multiple threads in one process (does this mean that
> the server also has to be mutithreaded?).
The R4 Intrinsics specification is removing the wording around "allowing
multiple logical applications in a single address space". It seems that
no one (meaning the various Intrinsics vendors) is convinced that it works
nor is willing to commit to it working. If anyone has used multiple
application contexts successfully, I'd like to hear about it.
The following paragraph is my first pass at a definition for application
contexts. Comments are welcome:
An application context is an Intrinsics data structure which contains
much of the "global" state of the Intrinsics. For example, it contains
the list of open displays and the list of registered work procedures.
Each application using the Intrinsics has a application context.
It can be created explicitly by the application by calling the routine
XtCreateApplicationContext. The Intrinsics routine XtInitialize
creates a default application context. Many Intrinsics routines have
two interfaces; one that takes an application context as its first
argument, and one that does not. An example is XtAddTimeOut and
XtAppAddTimeOut. The routines without the application context argument
use the default application context. The Intrinsics standard suggests
the use of the routines with the application context argument. The
other routines have been placed in an appendix of the standard.
However, either set of routines can be used by a DECtoolkit application.
|
1108.8 | | EPIK::BUEHLER | The IRS believes in effort: tax yourself | Mon Jul 24 1989 10:14 | 14 |
| >two interfaces; one that takes an application context as its first
>argument, and one that does not. An example is XtAddTimeOut and
>XtAppAddTimeOut. The routines without the application context argument
A piddly change. I suggest that you swap the ordering of XtAddTimeOut
and XtAppAddTimeOut so that they match the wording of the previous
sentence.
I didn't have any problem with understanding the definition, but that's
largely because it matches my prior understanding of application
contexts.
John
|
1108.9 | More primitive question | POOL::HALLYB | The Smart Money was on Goliath | Tue Jul 25 1989 09:57 | 10 |
| .7> Each application using the Intrinsics has a application context.
What's an "application" in this context? No pun intended.
Is it a pseudonym for "process" in the VMS sense?
I'm working on an "application" that currently involves two processes,
may grow to more, so the question is nontrivial.
John
|
1108.10 | | LEOVAX::TREGGIARI | | Tue Jul 25 1989 11:09 | 6 |
| > Is it a pseudonym for "process" in the VMS sense?
Yes, as far as I understand VMS processes (but I'm not an expert...), it is.
Leo
|