|
Re: Mobile offline programming
I understood that you were having fun...
The way it works, is:
Main Program Login (Rich client non-interactive) called by the execution.properties Calls Enter login (Regular rich client) subtask If all OK, calls Menu (Offline rich client)
menu waits for someone to tap.
When someone taps, Menu then raises an event (Sync from Server) in the Main Program
That is when I get -197. From what you are implying, the problem is Login?
Regards
Lawrence Author of the award winning book, "Kill Me Now!", amusing stories from the world of dating
From: Todd Baremore <tbaremor@...>
To: main@magicu-l.groups.io
Sent: Thursday, March 30, 2017 3:33 PM
Subject: Re: [magicu-l] Mobile offline programming
Lawrence,
I understood the book reference. Just having a bit of fun with it.
No offense intended.
In my app, the online tasks are batch tasks for updating data in
either direction. All my interactive tasks are offline. If an
online task is open when the timeout occurs, I do not think you will
be able to reconnect. All my online batch tasks are launched by
raising an events that are handled in the main program.
To get a better of idea of what is happening try the following:
In your startup program evaluate
ClientSessionSet('EnableCommunicationDialogs','FALSE'LOG). This
will turn off the error messages.
In the Main program create an event handler for Internal Event
"Unavailable Server" and set the scope to SubTree. In the handler
put the following :
Verify Warning expression = 'Unable to establish a server
connection. Status code : '&Str(ServerLastAccessStatus(),'1')
Launch your app , wait for the timeout and then call a program that
accesses the server.
If the returned code is 4 then –> The context is no longer
available (such as due to ContextActivityTimeout) and the client
could not recreate a new context in the Magic xpa server since there
were non-Offline tasks in the execution call tree.
You can also check the BrokerActivity.log for more info on what is
happening.
Todd
On 3/30/2017 2:58 AM, Lawrence Fisher
via Groups.Io wrote:
I am doing
something wrong then and I cannot figure out what. The Help
and the other documentation are not helpful and neither is
support
Regards
Lawrence
Author of the award winning book, "Kill Me Now!", amusing
stories from the world of dating
Lawrence,
It is
working for me. I have single user RIA runtime
license for XPA 3.2. Using two tablets, I am
able to push data to the server from both tablets
without having to restart the Android app,
provided I wait 30 seconds (see below). If I try
to do both at the same time, one of them gets an
error which is to be expected.
In my
magic.ini I have ContextInactivityTimeout=300
(30 seconds)
from my
mgrb.ini :
[APPLICATIONS_LIST]
My_App = MgxpaRuntime.exe
/INI=\\server1\folder1\control\Magic_XPA3_RIA.ini,,,,0,1
Now please
step down from the ledge or drop the gun/any other
applicable object.
Todd
On 3/29/2017
11:11 AM, Lawrence Fisher via Groups.Io wrote:
I have had no success at all
Kill me now
Regards
Lawrence
Author of the award winning book, "Kill
Me Now!", amusing stories from the world of
dating
Lawrence,
What happens if you restart the
broker after the context timeout?
I'll be visiting a customer
tomorrow to test my upgrades.
I'll let you know if I have any
success.
Todd
On
3/28/2017 8:49 AM, Lawrence Fisher
via Groups.Io wrote:
It
is "available"
This
is a regular runtime license
Regards
Lawrence
Author
of the award winning book,
"Kill Me Now!", amusing
stories from the world of
dating
Lawrence,
In the broker
monitor, what is
Server Application
Status after the
context has timed
out?
Todd
On
3/28/2017 2:19 AM,
Lawrence Fisher
via Groups.Io
wrote:
For me that
does not work.
My
context does
timeout but
when I try the
sync to
server, I get
the -197 which
is context not
found. This is
3.2a
I even
tried doing
this:
Updated
a Main Program
logical
variable,
"Must Sync"
with TRUE
Then
I have a Main
Program timer
event that
runs every 5
minutes and if
the "Must
Sync" is true,
runs the sync
to server.
-197.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
Tim,
I'm
experimenting
with setting
Context
Inactivity
Timeout in
lieu of
deleting a
context. I
don't have a
runtime
license
available, so
I'm
experimenting
with Studio.
Setting the
timeout to 600
( 1 minute )
does indeed
delete the
inactive
context, but
does not allow
any new
contexts to be
created. I
did notice
that Server
Application
Status
switches from
"Available" to
"Studio Mode"
after the
context times
out. On a
customer
server without
any open
contexts the
Server
Application
Status remains
"Available".
I'll visit
this customer
later this
week to run
some tests.
Todd
On
3/27/2017 5:37
PM, Tim Downie
wrote:
Ive had
inconsistent
results with
this and this
was due to a
bug if using
https which
was fixed
in3.1b. The
only thing is
you can never
kill this
context from
the app - or
at least i
cant but i
think this is
logical as in
a program can
not terminate
its own
context.
Id like to try
this process
in 3.2 but
havnt had the
time yet. I
will be back
to this soon
as we have
another client
wanting mobile
Tim and
Lawrence,
After
some guidance
from MSE and
pulling my
foot out of my
mouth, I need
to make some
corrections.
First,
only offline
programs are
copied to the
mobile device.
This only
happens when
the app is
started.
Second,
you can open a
new context by
calling a
non-offline
program. Here
is my mistake:
The
ServerLastAccessStatus()
does not check
if the server
is available.
It returns the
status of the
LAST request,
which could
have been an
hour ago.
To test this I
started my
mobile app
with a
connection to
the server.
With my mobile
app still
running, I
shutdown the
server(s)/broker
and restarted
the broker.
>From the
mobile app I
then tried to
push data to
the server,
which was
successful.
I still need
to work out
server error
handling, but
this is great
news for
mobile offline
development.
Anyone up for
a
Go-To-Meeting
to put our
heads together
on this (and
share the
pain?) I'm
thinking
Thursday
afternoon New
York time.
Todd
On
3/26/2017
10:11 PM, Todd
Baremore
wrote:
TIm,
I agree.
Ideally a
server will
have a
floating pool
of mobile
contexts.
This would
allow a
customer to to
put your app
on 20 devices
while only
purchasing a 5
user
licenses. At
startup each
device would
synch the
source xml
files, launch
your app
(which would
check for any
data
synchronization)
and then drop
the context if
applicable for
the app to do
so. For
server
dependent apps
the device
would stay
connected to
the server and
the licensing
scheme would
be 1:1. For
mobile offline
apps, having
an N:1
licensing
scheme will
give us a much
more
affordable
solution to
offer our
customers.
If a mobile
app is started
without a
server
connection and
the
application
ecf has been
updated on the
server you
could have
problems with:
DataviewToDataSource() if column names have changed
Callling
server
programs if
the number of
parameters or
their data
types have
changed
In order for
this
diabolical
plan to work,
the developer
would be
responsible
for
determining a
version
difference
between the
server and
mobile app (we
can already do
this) and
maintaining
some form of
change list
that could be
checked after
the mobile app
creates a new
context. The
mobile app
would then
need to
determine if
it is safe to
synchronize
data or call a
server
program. I'm
doubtful that
updating the
app xml files
will ever be
achieved with
the app open
and program(s)
running.
I'm sure there
is more to it
than I have
outlined here.
If MSE is
listening I
would like to
hear your
thoughts on
the matter.
Todd
On
3/26/2017 7:14
PM, Tim Downie
wrote:
This is
where in my
belief magic
mobile is
flawed....you
should be able
to open a
connection
(context )
when you like
and then
immediately
terminate it,
and then
create another
one if/when
you need to
interact with
the server -
this reduces
connections to
the server
etc.
if you
start without
a context it
will then
create one,
but it should
also be able
to create a
new context if
for some
reason the
context is not
found.
Lawrence,
This is
my
understanding
and experience
of how
contexts
works. If
anyone knows
better please
correct me as
needed.
A context
is added to
the server
upon startup
of your mobile
device
application,
assuming your
device has
network access
to the
server. With
the Context
Inactivity
Timeout set to
2 minutes, if
your device
has not made
any requests
of the server
in 2 minutes,
the context is
closed and
your app is
now offline.
The only way
to reestablish
a context is
to close and
restart the
app. The
default
Context
Inactivity
Timeout is 24
hours which
means you
could start
your app in
Seattle in the
morning, catch
a flight to
New York and
still have the
same context
available when
you land in
New York as
long as you
leave the app
running.
Open the
broker monitor
on your server
before
starting your
mobile app.
You will see
the traffic
coming from
you mobile
device and the
creation of a
context for
your device.
The first time
you run your
app, the
source xml is
sent to the
device which
is why it
takes a bit
longer to
start the
first time.
Assuming
the Context
Inactivity
Timeout is set
to 24 hours,
if you want to
test what
happens to
your app when
it loses the
connection to
the server,
start your
mobile app,
turn off the
mobile/WiFi
data
connections on
your device.
Work with your
app for a few
minutes and
then turn on
your
mobile/WiFi
connections on
and try to
send data to
the server. In
the Broker
Monitor you
will notice
that the
context is
maintained
even though
the device has
lost its
network
connections.
This
could prove
helpful in
your next
book, "Kill Me
Again",
amusing
anecdotes from
the world of
Magic mobile
development
Todd
On
3/26/2017 8:51
AM, Lawrence
Fisher via
Groups.Io
wrote:
Hi all
I am
tried to
define a proof
of concept
working with
offline
mobile.
I
defined a
Context
Timeout of 2
mins.
I managed
to get the
sync from
server to
client working
with no
problem. Once
there is the
timeout, the
client
continues to
work.
The
problem is
when I need to
update the
server,
meaning
"client to
server"
From what
I understand
from the
documentation,
when you make
the first
server access,
such as
accessing a
server table,
there is a
connection to
the server.
What I get is
an error -197
which is
context not
found.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
|
|
|
Re: Get.dll crashing unipaas 1.9 all of a sudden
Florian Groothuis
We had the same problem a few years ago. In our case it depended on the number of printers a user had assigned to his
account. Try raising the buffer size with “GET_SET_PRN_BUFFSIZE”. We set it to 10000.
Van: main@magicu-l.groups.io [mailto:main@magicu-l.groups.io]
Namens John Knickerbocker
Verzonden: donderdag 30 maart 2017 14:15
Aan: main@magicu-l.groups.io
Onderwerp: [magicu-l] Get.dll crashing unipaas 1.9 all of a sudden
Hello group,
We have a client that we use get.dll with in order to retrieve the list of windows printers. It has been working for years now, but this week it began crashing right when the dll is called. Not every time, sometimes it works without crashing.
Has anyone else seen this behavior?
Met vriendelijke groet - With kind regards,
Florian Groothuis
Analist/programmeur
+31 (0)6 21927914 |
|
|
|
|
Re: Get.dll crashing unipaas 1.9 all of a sudden
Hi John,
I had crashes of this caused by printerdriverupdates. Ignoring the update caused the crash. I think there should be changed something with your printerdriverlist that GET cannot handle.
Best regards,
André
|
|
|
Re: Mobile offline programming
Lawrence,
I understood the book reference. Just having a bit of fun with it.
No offense intended.
In my app, the online tasks are batch tasks for updating data in
either direction. All my interactive tasks are offline. If an
online task is open when the timeout occurs, I do not think you will
be able to reconnect. All my online batch tasks are launched by
raising an events that are handled in the main program.
To get a better of idea of what is happening try the following:
In your startup program evaluate
ClientSessionSet('EnableCommunicationDialogs','FALSE'LOG). This
will turn off the error messages.
In the Main program create an event handler for Internal Event
"Unavailable Server" and set the scope to SubTree. In the handler
put the following :
Verify Warning expression = 'Unable to establish a server
connection. Status code : '&Str(ServerLastAccessStatus(),'1')
Launch your app , wait for the timeout and then call a program that
accesses the server.
If the returned code is 4 then –> The context is no longer
available (such as due to ContextActivityTimeout) and the client
could not recreate a new context in the Magic xpa server since there
were non-Offline tasks in the execution call tree.
You can also check the BrokerActivity.log for more info on what is
happening.
Todd
toggle quoted messageShow quoted text
On 3/30/2017 2:58 AM, Lawrence Fisher
via Groups.Io wrote:
I am doing
something wrong then and I cannot figure out what. The Help
and the other documentation are not helpful and neither is
support
Regards
Lawrence
Author of the award winning book, "Kill Me Now!", amusing
stories from the world of dating
Lawrence,
It is
working for me. I have single user RIA runtime
license for XPA 3.2. Using two tablets, I am
able to push data to the server from both tablets
without having to restart the Android app,
provided I wait 30 seconds (see below). If I try
to do both at the same time, one of them gets an
error which is to be expected.
In my
magic.ini I have ContextInactivityTimeout=300
(30 seconds)
from my
mgrb.ini :
[APPLICATIONS_LIST]
My_App = MgxpaRuntime.exe
/INI=\\server1\folder1\control\Magic_XPA3_RIA.ini,,,,0,1
Now please
step down from the ledge or drop the gun/any other
applicable object.
Todd
On 3/29/2017
11:11 AM, Lawrence Fisher via Groups.Io wrote:
I have had no success at all
Kill me now
Regards
Lawrence
Author of the award winning book, "Kill
Me Now!", amusing stories from the world of
dating
Lawrence,
What happens if you restart the
broker after the context timeout?
I'll be visiting a customer
tomorrow to test my upgrades.
I'll let you know if I have any
success.
Todd
On
3/28/2017 8:49 AM, Lawrence Fisher
via Groups.Io wrote:
It
is "available"
This
is a regular runtime license
Regards
Lawrence
Author
of the award winning book,
"Kill Me Now!", amusing
stories from the world of
dating
Lawrence,
In the broker
monitor, what is
Server Application
Status after the
context has timed
out?
Todd
On
3/28/2017 2:19 AM,
Lawrence Fisher
via Groups.Io
wrote:
For me that
does not work.
My
context does
timeout but
when I try the
sync to
server, I get
the -197 which
is context not
found. This is
3.2a
I even
tried doing
this:
Updated
a Main Program
logical
variable,
"Must Sync"
with TRUE
Then
I have a Main
Program timer
event that
runs every 5
minutes and if
the "Must
Sync" is true,
runs the sync
to server.
-197.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
Tim,
I'm
experimenting
with setting
Context
Inactivity
Timeout in
lieu of
deleting a
context. I
don't have a
runtime
license
available, so
I'm
experimenting
with Studio.
Setting the
timeout to 600
( 1 minute )
does indeed
delete the
inactive
context, but
does not allow
any new
contexts to be
created. I
did notice
that Server
Application
Status
switches from
"Available" to
"Studio Mode"
after the
context times
out. On a
customer
server without
any open
contexts the
Server
Application
Status remains
"Available".
I'll visit
this customer
later this
week to run
some tests.
Todd
On
3/27/2017 5:37
PM, Tim Downie
wrote:
Ive had
inconsistent
results with
this and this
was due to a
bug if using
https which
was fixed
in3.1b. The
only thing is
you can never
kill this
context from
the app - or
at least i
cant but i
think this is
logical as in
a program can
not terminate
its own
context.
Id like to try
this process
in 3.2 but
havnt had the
time yet. I
will be back
to this soon
as we have
another client
wanting mobile
Tim and
Lawrence,
After
some guidance
from MSE and
pulling my
foot out of my
mouth, I need
to make some
corrections.
First,
only offline
programs are
copied to the
mobile device.
This only
happens when
the app is
started.
Second,
you can open a
new context by
calling a
non-offline
program. Here
is my mistake:
The
ServerLastAccessStatus()
does not check
if the server
is available.
It returns the
status of the
LAST request,
which could
have been an
hour ago.
To test this I
started my
mobile app
with a
connection to
the server.
With my mobile
app still
running, I
shutdown the
server(s)/broker
and restarted
the broker.
>From the
mobile app I
then tried to
push data to
the server,
which was
successful.
I still need
to work out
server error
handling, but
this is great
news for
mobile offline
development.
Anyone up for
a
Go-To-Meeting
to put our
heads together
on this (and
share the
pain?) I'm
thinking
Thursday
afternoon New
York time.
Todd
On
3/26/2017
10:11 PM, Todd
Baremore
wrote:
TIm,
I agree.
Ideally a
server will
have a
floating pool
of mobile
contexts.
This would
allow a
customer to to
put your app
on 20 devices
while only
purchasing a 5
user
licenses. At
startup each
device would
synch the
source xml
files, launch
your app
(which would
check for any
data
synchronization)
and then drop
the context if
applicable for
the app to do
so. For
server
dependent apps
the device
would stay
connected to
the server and
the licensing
scheme would
be 1:1. For
mobile offline
apps, having
an N:1
licensing
scheme will
give us a much
more
affordable
solution to
offer our
customers.
If a mobile
app is started
without a
server
connection and
the
application
ecf has been
updated on the
server you
could have
problems with:
DataviewToDataSource() if column names have changed
Callling
server
programs if
the number of
parameters or
their data
types have
changed
In order for
this
diabolical
plan to work,
the developer
would be
responsible
for
determining a
version
difference
between the
server and
mobile app (we
can already do
this) and
maintaining
some form of
change list
that could be
checked after
the mobile app
creates a new
context. The
mobile app
would then
need to
determine if
it is safe to
synchronize
data or call a
server
program. I'm
doubtful that
updating the
app xml files
will ever be
achieved with
the app open
and program(s)
running.
I'm sure there
is more to it
than I have
outlined here.
If MSE is
listening I
would like to
hear your
thoughts on
the matter.
Todd
On
3/26/2017 7:14
PM, Tim Downie
wrote:
This is
where in my
belief magic
mobile is
flawed....you
should be able
to open a
connection
(context )
when you like
and then
immediately
terminate it,
and then
create another
one if/when
you need to
interact with
the server -
this reduces
connections to
the server
etc.
if you
start without
a context it
will then
create one,
but it should
also be able
to create a
new context if
for some
reason the
context is not
found.
Lawrence,
This is
my
understanding
and experience
of how
contexts
works. If
anyone knows
better please
correct me as
needed.
A context
is added to
the server
upon startup
of your mobile
device
application,
assuming your
device has
network access
to the
server. With
the Context
Inactivity
Timeout set to
2 minutes, if
your device
has not made
any requests
of the server
in 2 minutes,
the context is
closed and
your app is
now offline.
The only way
to reestablish
a context is
to close and
restart the
app. The
default
Context
Inactivity
Timeout is 24
hours which
means you
could start
your app in
Seattle in the
morning, catch
a flight to
New York and
still have the
same context
available when
you land in
New York as
long as you
leave the app
running.
Open the
broker monitor
on your server
before
starting your
mobile app.
You will see
the traffic
coming from
you mobile
device and the
creation of a
context for
your device.
The first time
you run your
app, the
source xml is
sent to the
device which
is why it
takes a bit
longer to
start the
first time.
Assuming
the Context
Inactivity
Timeout is set
to 24 hours,
if you want to
test what
happens to
your app when
it loses the
connection to
the server,
start your
mobile app,
turn off the
mobile/WiFi
data
connections on
your device.
Work with your
app for a few
minutes and
then turn on
your
mobile/WiFi
connections on
and try to
send data to
the server. In
the Broker
Monitor you
will notice
that the
context is
maintained
even though
the device has
lost its
network
connections.
This
could prove
helpful in
your next
book, "Kill Me
Again",
amusing
anecdotes from
the world of
Magic mobile
development
Todd
On
3/26/2017 8:51
AM, Lawrence
Fisher via
Groups.Io
wrote:
Hi all
I am
tried to
define a proof
of concept
working with
offline
mobile.
I
defined a
Context
Timeout of 2
mins.
I managed
to get the
sync from
server to
client working
with no
problem. Once
there is the
timeout, the
client
continues to
work.
The
problem is
when I need to
update the
server,
meaning
"client to
server"
From what
I understand
from the
documentation,
when you make
the first
server access,
such as
accessing a
server table,
there is a
connection to
the server.
What I get is
an error -197
which is
context not
found.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
|
|
|
Get.dll crashing unipaas 1.9 all of a sudden
Hello group,
We have a client that we use get.dll with in order to retrieve the list of windows printers. It has been working for years now, but this week it began crashing right when the dll is called. Not every time, sometimes it works without crashing. Has anyone else seen this behavior?
|
|
|
Re: Behavior XPA vs Unipaas
Hello Harry, that's exactly what I needed. Thanks a lot. Best Regards Koen
|
|
|
Re: Behavior XPA vs Unipaas
Hi Koen, From the release notes:
SpecialExpReturnNull – This flag maintains backward compatibility when comparing a variable that has a Null value to an empty value. Best regards, Harry Kleinsmit.
|
|
|
Hello All, I have started to see what we have to do to transfer our application from Unipaas to XPA. But I have noticed other behavior in XPA. If you create the following program in Unipaas and in XPA: Variable A = Alpha U20, NULL Allowed Record Suffix Update A with NULL() IF TRIM(A)<>'' THEN 'OK' END IF In Unipaas the OK is not shown, but in XPA it will show the OK.
Is there something changed at the TRIM-function in XPA? I have tried to add the special setting SpecialTrimTrailingNulls in the INI-file, but nothing has changed.
Any ideas? Thanks in advance Best regards Koen
|
|
|
Re: Mobile offline programming
Kill Me Now is the name of my book (http://killmenow.org)
I am doing something wrong then and I cannot figure out what. The Help and the other documentation are not helpful and neither is support Regards
Lawrence Author of the award winning book, "Kill Me Now!", amusing stories from the world of dating
From: Todd Baremore <tbaremor@...>
To: main@magicu-l.groups.io
Sent: Wednesday, March 29, 2017 9:15 PM
Subject: Re: [magicu-l] Mobile offline programming
Lawrence,
It is working for me. I have single user RIA runtime license for
XPA 3.2. Using two tablets, I am able to push data to the server
from both tablets without having to restart the Android app,
provided I wait 30 seconds (see below). If I try to do both at
the same time, one of them gets an error which is to be expected.
In my magic.ini I have ContextInactivityTimeout=300 (30
seconds)
from my mgrb.ini :
[APPLICATIONS_LIST]
My_App = MgxpaRuntime.exe
/INI=\\server1\folder1\control\Magic_XPA3_RIA.ini,,,,0,1
Now please step down from the ledge or drop the gun/any other
applicable object.
Todd
On 3/29/2017 11:11 AM, Lawrence Fisher
via Groups.Io wrote:
I have had no success at all
Kill me now
Regards
Lawrence
Author of the award winning book, "Kill Me Now!", amusing
stories from the world of dating
Lawrence,
What happens if you restart the broker after
the context timeout?
I'll be visiting a customer tomorrow to test my
upgrades. I'll let you know if I have any
success.
Todd
On
3/28/2017 8:49 AM, Lawrence Fisher via Groups.Io
wrote:
It
is "available"
This
is a regular runtime license
Regards
Lawrence
Author
of the award winning book, "Kill Me Now!",
amusing stories from the world of dating
Lawrence,
In the broker monitor, what is
Server Application Status after
the context has timed out?
Todd
On
3/28/2017 2:19 AM, Lawrence Fisher
via Groups.Io wrote:
For me that does not work.
My
context does timeout but when
I try the sync to server, I
get the -197 which is context
not found. This is 3.2a
I even tried doing this:
Updated a Main
Program logical variable,
"Must Sync" with TRUE
Then I have a
Main Program timer event that
runs every 5 minutes and if
the "Must Sync" is true, runs
the sync to server. -197.
What am I doing wrong?
Regards
Lawrence
Author
of the award winning book,
"Kill Me Now!", amusing
stories from the world of
dating
Tim,
I'm
experimenting with
setting Context
Inactivity Timeout
in lieu of
deleting a
context. I don't
have a runtime
license available,
so I'm
experimenting with
Studio. Setting
the timeout to 600
( 1 minute ) does
indeed delete the
inactive context,
but does not allow
any new contexts
to be created. I
did notice that
Server Application
Status switches
from "Available"
to "Studio Mode"
after the context
times out. On a
customer server
without any open
contexts the
Server Application
Status remains
"Available".
I'll visit this
customer later
this week to run
some tests.
Todd
On
3/27/2017 5:37 PM,
Tim Downie wrote:
Ive had
inconsistent results
with this and this
was due to a bug if
using https which
was fixed in3.1b.
The only thing is
you can never kill
this context from
the app - or at
least i cant but i
think this is
logical as in a
program can not
terminate its own
context.
Id like to try this
process in 3.2 but
havnt had the time
yet. I will be back
to this soon as we
have another client
wanting mobile
Tim and
Lawrence,
After some
guidance from
MSE and pulling
my foot out of
my mouth, I need
to make some
corrections.
First, only
offline programs
are copied to
the mobile
device. This
only happens
when the app is
started.
Second, you
can open a new
context by
calling a
non-offline
program. Here is
my mistake:
The
ServerLastAccessStatus()
does not check
if the server is
available. It
returns the
status of the
LAST request,
which could have
been an hour
ago.
To test this I
started my mobile
app with a
connection to the
server. With my
mobile app still
running, I
shutdown the
server(s)/broker
and restarted the
broker. From the
mobile app I then
tried to push data
to the server,
which was
successful. I
still need to work
out server error
handling, but this
is great news for
mobile offline
development.
Anyone up for a
Go-To-Meeting to
put our heads
together on this
(and share the
pain?) I'm
thinking Thursday
afternoon New York
time.
Todd
On
3/26/2017 10:11
PM, Todd
Baremore wrote:
TIm,
I agree.
Ideally a server
will have a
floating pool of
mobile
contexts. This
would allow a
customer to to
put your app on
20 devices while
only purchasing
a 5 user
licenses. At
startup each
device would
synch the source
xml files,
launch your app
(which would
check for any
data
synchronization)
and then drop
the context if
applicable for
the app to do
so. For server
dependent apps
the device would
stay connected
to the server
and the
licensing scheme
would be 1:1.
For mobile
offline apps,
having an N:1
licensing scheme
will give us a
much more
affordable
solution to
offer our
customers.
If a mobile app
is started
without a server
connection and
the application
ecf has been
updated on the
server you could
have problems
with:
DataviewToDataSource() if column names have changed
Callling server
programs if the
number of
parameters or
their data types
have changed
In order for
this diabolical
plan to work,
the developer
would be
responsible for
determining a
version
difference
between the
server and
mobile app (we
can already do
this) and
maintaining some
form of change
list that could
be checked after
the mobile app
creates a new
context. The
mobile app would
then need to
determine if it
is safe to
synchronize data
or call a server
program. I'm
doubtful that
updating the app
xml files will
ever be achieved
with the app
open and
program(s)
running.
I'm sure there
is more to it
than I have
outlined here.
If MSE is
listening I
would like to
hear your
thoughts on the
matter.
Todd
On
3/26/2017 7:14
PM, Tim Downie
wrote:
This is
where in my
belief magic
mobile is
flawed....you
should be able
to open a
connection
(context )
when you like
and then
immediately
terminate it,
and then
create another
one if/when
you need to
interact with
the server -
this reduces
connections to
the server
etc.
if you
start without
a context it
will then
create one,
but it should
also be able
to create a
new context if
for some
reason the
context is not
found.
Lawrence,
This is
my
understanding
and experience
of how
contexts
works. If
anyone knows
better please
correct me as
needed.
A context
is added to
the server
upon startup
of your mobile
device
application,
assuming your
device has
network access
to the
server. With
the Context
Inactivity
Timeout set to
2 minutes, if
your device
has not made
any requests
of the server
in 2 minutes,
the context is
closed and
your app is
now offline.
The only way
to reestablish
a context is
to close and
restart the
app. The
default
Context
Inactivity
Timeout is 24
hours which
means you
could start
your app in
Seattle in the
morning, catch
a flight to
New York and
still have the
same context
available when
you land in
New York as
long as you
leave the app
running.
Open the
broker monitor
on your server
before
starting your
mobile app.
You will see
the traffic
coming from
you mobile
device and the
creation of a
context for
your device.
The first time
you run your
app, the
source xml is
sent to the
device which
is why it
takes a bit
longer to
start the
first time.
Assuming
the Context
Inactivity
Timeout is set
to 24 hours,
if you want to
test what
happens to
your app when
it loses the
connection to
the server,
start your
mobile app,
turn off the
mobile/WiFi
data
connections on
your device.
Work with your
app for a few
minutes and
then turn on
your
mobile/WiFi
connections on
and try to
send data to
the server. In
the Broker
Monitor you
will notice
that the
context is
maintained
even though
the device has
lost its
network
connections.
This
could prove
helpful in
your next
book, "Kill Me
Again",
amusing
anecdotes from
the world of
Magic mobile
development
Todd
On
3/26/2017 8:51
AM, Lawrence
Fisher via
Groups.Io
wrote:
Hi all
I am
tried to
define a proof
of concept
working with
offline
mobile.
I
defined a
Context
Timeout of 2
mins.
I managed
to get the
sync from
server to
client working
with no
problem. Once
there is the
timeout, the
client
continues to
work.
The
problem is
when I need to
update the
server,
meaning
"client to
server"
From what
I understand
from the
documentation,
when you make
the first
server access,
such as
accessing a
server table,
there is a
connection to
the server.
What I get is
an error -197
which is
context not
found.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
|
|
|
Re: Mobile offline programming
Lawrence,
It is working for me. I have single user RIA runtime license for
XPA 3.2. Using two tablets, I am able to push data to the server
from both tablets without having to restart the Android app,
provided I wait 30 seconds (see below). If I try to do both at
the same time, one of them gets an error which is to be expected.
In my magic.ini I have ContextInactivityTimeout=300 (30
seconds)
from my mgrb.ini :
[APPLICATIONS_LIST]
My_App = MgxpaRuntime.exe
/INI=\\server1\folder1\control\Magic_XPA3_RIA.ini,,,,0,1
Now please step down from the ledge or drop the gun/any other
applicable object.
Todd
toggle quoted messageShow quoted text
On 3/29/2017 11:11 AM, Lawrence Fisher
via Groups.Io wrote:
I have had no success at all
Kill me now
Regards
Lawrence
Author of the award winning book, "Kill Me Now!", amusing
stories from the world of dating
Lawrence,
What happens if you restart the broker after
the context timeout?
I'll be visiting a customer tomorrow to test my
upgrades. I'll let you know if I have any
success.
Todd
On
3/28/2017 8:49 AM, Lawrence Fisher via Groups.Io
wrote:
It
is "available"
This
is a regular runtime license
Regards
Lawrence
Author
of the award winning book, "Kill Me Now!",
amusing stories from the world of dating
Lawrence,
In the broker monitor, what is
Server Application Status after
the context has timed out?
Todd
On
3/28/2017 2:19 AM, Lawrence Fisher
via Groups.Io wrote:
For me that does not work.
My
context does timeout but when
I try the sync to server, I
get the -197 which is context
not found. This is 3.2a
I even tried doing this:
Updated a Main
Program logical variable,
"Must Sync" with TRUE
Then I have a
Main Program timer event that
runs every 5 minutes and if
the "Must Sync" is true, runs
the sync to server. -197.
What am I doing wrong?
Regards
Lawrence
Author
of the award winning book,
"Kill Me Now!", amusing
stories from the world of
dating
Tim,
I'm
experimenting with
setting Context
Inactivity Timeout
in lieu of
deleting a
context. I don't
have a runtime
license available,
so I'm
experimenting with
Studio. Setting
the timeout to 600
( 1 minute ) does
indeed delete the
inactive context,
but does not allow
any new contexts
to be created. I
did notice that
Server Application
Status switches
from "Available"
to "Studio Mode"
after the context
times out. On a
customer server
without any open
contexts the
Server Application
Status remains
"Available".
I'll visit this
customer later
this week to run
some tests.
Todd
On
3/27/2017 5:37 PM,
Tim Downie wrote:
Ive had
inconsistent results
with this and this
was due to a bug if
using https which
was fixed in3.1b.
The only thing is
you can never kill
this context from
the app - or at
least i cant but i
think this is
logical as in a
program can not
terminate its own
context.
Id like to try this
process in 3.2 but
havnt had the time
yet. I will be back
to this soon as we
have another client
wanting mobile
Tim and
Lawrence,
After some
guidance from
MSE and pulling
my foot out of
my mouth, I need
to make some
corrections.
First, only
offline programs
are copied to
the mobile
device. This
only happens
when the app is
started.
Second, you
can open a new
context by
calling a
non-offline
program. Here is
my mistake:
The
ServerLastAccessStatus()
does not check
if the server is
available. It
returns the
status of the
LAST request,
which could have
been an hour
ago.
To test this I
started my mobile
app with a
connection to the
server. With my
mobile app still
running, I
shutdown the
server(s)/broker
and restarted the
broker. From the
mobile app I then
tried to push data
to the server,
which was
successful. I
still need to work
out server error
handling, but this
is great news for
mobile offline
development.
Anyone up for a
Go-To-Meeting to
put our heads
together on this
(and share the
pain?) I'm
thinking Thursday
afternoon New York
time.
Todd
On
3/26/2017 10:11
PM, Todd
Baremore wrote:
TIm,
I agree.
Ideally a server
will have a
floating pool of
mobile
contexts. This
would allow a
customer to to
put your app on
20 devices while
only purchasing
a 5 user
licenses. At
startup each
device would
synch the source
xml files,
launch your app
(which would
check for any
data
synchronization)
and then drop
the context if
applicable for
the app to do
so. For server
dependent apps
the device would
stay connected
to the server
and the
licensing scheme
would be 1:1.
For mobile
offline apps,
having an N:1
licensing scheme
will give us a
much more
affordable
solution to
offer our
customers.
If a mobile app
is started
without a server
connection and
the application
ecf has been
updated on the
server you could
have problems
with:
DataviewToDataSource() if column names have changed
Callling server
programs if the
number of
parameters or
their data types
have changed
In order for
this diabolical
plan to work,
the developer
would be
responsible for
determining a
version
difference
between the
server and
mobile app (we
can already do
this) and
maintaining some
form of change
list that could
be checked after
the mobile app
creates a new
context. The
mobile app would
then need to
determine if it
is safe to
synchronize data
or call a server
program. I'm
doubtful that
updating the app
xml files will
ever be achieved
with the app
open and
program(s)
running.
I'm sure there
is more to it
than I have
outlined here.
If MSE is
listening I
would like to
hear your
thoughts on the
matter.
Todd
On
3/26/2017 7:14
PM, Tim Downie
wrote:
This is
where in my
belief magic
mobile is
flawed....you
should be able
to open a
connection
(context )
when you like
and then
immediately
terminate it,
and then
create another
one if/when
you need to
interact with
the server -
this reduces
connections to
the server
etc.
if you
start without
a context it
will then
create one,
but it should
also be able
to create a
new context if
for some
reason the
context is not
found.
Lawrence,
This is
my
understanding
and experience
of how
contexts
works. If
anyone knows
better please
correct me as
needed.
A context
is added to
the server
upon startup
of your mobile
device
application,
assuming your
device has
network access
to the
server. With
the Context
Inactivity
Timeout set to
2 minutes, if
your device
has not made
any requests
of the server
in 2 minutes,
the context is
closed and
your app is
now offline.
The only way
to reestablish
a context is
to close and
restart the
app. The
default
Context
Inactivity
Timeout is 24
hours which
means you
could start
your app in
Seattle in the
morning, catch
a flight to
New York and
still have the
same context
available when
you land in
New York as
long as you
leave the app
running.
Open the
broker monitor
on your server
before
starting your
mobile app.
You will see
the traffic
coming from
you mobile
device and the
creation of a
context for
your device.
The first time
you run your
app, the
source xml is
sent to the
device which
is why it
takes a bit
longer to
start the
first time.
Assuming
the Context
Inactivity
Timeout is set
to 24 hours,
if you want to
test what
happens to
your app when
it loses the
connection to
the server,
start your
mobile app,
turn off the
mobile/WiFi
data
connections on
your device.
Work with your
app for a few
minutes and
then turn on
your
mobile/WiFi
connections on
and try to
send data to
the server. In
the Broker
Monitor you
will notice
that the
context is
maintained
even though
the device has
lost its
network
connections.
This
could prove
helpful in
your next
book, "Kill Me
Again",
amusing
anecdotes from
the world of
Magic mobile
development
Todd
On
3/26/2017 8:51
AM, Lawrence
Fisher via
Groups.Io
wrote:
Hi all
I am
tried to
define a proof
of concept
working with
offline
mobile.
I
defined a
Context
Timeout of 2
mins.
I managed
to get the
sync from
server to
client working
with no
problem. Once
there is the
timeout, the
client
continues to
work.
The
problem is
when I need to
update the
server,
meaning
"client to
server"
From what
I understand
from the
documentation,
when you make
the first
server access,
such as
accessing a
server table,
there is a
connection to
the server.
What I get is
an error -197
which is
context not
found.
What am I
doing wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me
Now!", amusing
stories from
the world of
dating
|
|
|
Re: Mobile offline programming
I have had no success at all
Kill me now Regards
Lawrence Author of the award winning book, "Kill Me Now!", amusing stories from the world of dating
From: Todd Baremore <tbaremor@...>
To: main@magicu-l.groups.io
Sent: Tuesday, March 28, 2017 4:39 PM
Subject: Re: [magicu-l] Mobile offline programming
Lawrence,
What happens if you restart the broker after the context timeout?
I'll be visiting a customer tomorrow to test my upgrades. I'll
let you know if I have any success.
Todd
On 3/28/2017 8:49 AM, Lawrence Fisher
via Groups.Io wrote:
It is "available"
This is a regular
runtime license
Regards
Lawrence
Author of the award
winning book, "Kill Me Now!", amusing stories from the world
of dating
Lawrence,
In the broker monitor, what is Server
Application Status after the context has timed
out?
Todd
On
3/28/2017 2:19 AM, Lawrence Fisher via Groups.Io
wrote:
For me that does not work.
My
context does timeout but when I try the sync
to server, I get the -197 which is context not
found. This is 3.2a
I even tried doing this:
Updated a Main Program logical
variable, "Must Sync" with TRUE
Then I have a Main Program timer
event that runs every 5 minutes and if the
"Must Sync" is true, runs the sync to server.
-197.
What am I doing wrong?
Regards
Lawrence
Author
of the award winning book, "Kill Me Now!",
amusing stories from the world of dating
Tim,
I'm
experimenting with setting Context
Inactivity Timeout in lieu of
deleting a context. I don't have
a runtime license available, so
I'm experimenting with Studio.
Setting the timeout to 600 ( 1
minute ) does indeed delete the
inactive context, but does not
allow any new contexts to be
created. I did notice that Server
Application Status switches from
"Available" to "Studio Mode" after
the context times out. On a
customer server without any open
contexts the Server Application
Status remains "Available". I'll
visit this customer later this
week to run some tests.
Todd
On
3/27/2017 5:37 PM, Tim Downie
wrote:
Ive had inconsistent results
with this and this was due to a bug
if using https which was fixed
in3.1b. The only thing is you can
never kill this context from the app
- or at least i cant but i think
this is logical as in a program can
not terminate its own context.
Id like to try this process in 3.2
but havnt had the time yet. I will
be back to this soon as we have
another client wanting mobile
Tim and Lawrence,
After some guidance from MSE
and pulling my foot out of my
mouth, I need to make some
corrections.
First, only offline programs
are copied to the mobile device.
This only happens when the app
is started.
Second, you can open a new
context by calling a non-offline
program. Here is my mistake:
The ServerLastAccessStatus()
does not check if the server is
available. It returns the status
of the LAST request, which could
have been an hour ago.
To test this I started my mobile
app with a connection to the
server. With my mobile app still
running, I shutdown the
server(s)/broker and restarted the
broker. From the mobile app I
then tried to push data to the
server, which was successful. I
still need to work out server
error handling, but this is great
news for mobile offline
development. Anyone up for a
Go-To-Meeting to put our heads
together on this (and share the
pain?) I'm thinking Thursday
afternoon New York time.
Todd
On
3/26/2017 10:11 PM, Todd
Baremore wrote:
TIm,
I agree. Ideally a server will
have a floating pool of mobile
contexts. This would allow a
customer to to put your app on
20 devices while only purchasing
a 5 user licenses. At startup
each device would synch the
source xml files, launch your
app (which would check for any
data synchronization) and then
drop the context if applicable
for the app to do so. For
server dependent apps the device
would stay connected to the
server and the licensing scheme
would be 1:1. For mobile
offline apps, having an N:1
licensing scheme will give us a
much more affordable solution to
offer our customers.
If a mobile app is started
without a server connection and
the application ecf has been
updated on the server you could
have problems with:
DataviewToDataSource() if column
names have changed
Callling server programs if the
number of parameters or their
data types have changed
In order for this diabolical
plan to work, the developer
would be responsible for
determining a version difference
between the server and mobile
app (we can already do this) and
maintaining some form of change
list that could be checked after
the mobile app creates a new
context. The mobile app would
then need to determine if it is
safe to synchronize data or call
a server program. I'm doubtful
that updating the app xml files
will ever be achieved with the
app open and program(s) running.
I'm sure there is more to it
than I have outlined here. If
MSE is listening I would like to
hear your thoughts on the
matter.
Todd
On
3/26/2017 7:14 PM, Tim Downie
wrote:
This is where in my
belief magic mobile is
flawed....you should be
able to open a connection
(context ) when you like
and then immediately
terminate it, and then
create another one if/when
you need to interact with
the server - this reduces
connections to the server
etc.
if you start without a
context it will then
create one, but it should
also be able to create a
new context if for some
reason the context is not
found.
Lawrence,
This is my
understanding and
experience of how
contexts works. If
anyone knows better
please correct me as
needed.
A context is added
to the server upon
startup of your mobile
device application,
assuming your device
has network access to
the server. With the
Context Inactivity
Timeout set to 2
minutes, if your
device has not made
any requests of the
server in 2 minutes,
the context is closed
and your app is now
offline. The only
way to reestablish a
context is to close
and restart the app.
The default Context
Inactivity Timeout is
24 hours which means
you could start your
app in Seattle in the
morning, catch a
flight to New York and
still have the same
context available when
you land in New York
as long as you leave
the app running.
Open the broker
monitor on your server
before starting your
mobile app. You will
see the traffic coming
from you mobile device
and the creation of a
context for your
device. The first
time you run your app,
the source xml is sent
to the device which is
why it takes a bit
longer to start the
first time.
Assuming the
Context Inactivity
Timeout is set to 24
hours, if you want to
test what happens to
your app when it loses
the connection to the
server, start your
mobile app, turn off
the mobile/WiFi data
connections on your
device. Work with
your app for a few
minutes and then turn
on your mobile/WiFi
connections on and try
to send data to the
server. In the Broker
Monitor you will
notice that the
context is maintained
even though the device
has lost its network
connections.
This could prove
helpful in your next
book, "Kill Me Again",
amusing anecdotes from
the world of Magic
mobile development
Todd
On
3/26/2017 8:51 AM,
Lawrence Fisher via
Groups.Io wrote:
Hi all
I am tried to
define a proof of
concept working
with offline
mobile.
I
defined a Context
Timeout of 2 mins.
I managed to
get the sync from
server to client
working with no
problem. Once
there is the
timeout, the
client continues
to work.
The
problem is when I
need to update the
server, meaning
"client to server"
From what I
understand from
the documentation,
when you make the
first server
access, such as
accessing a server
table, there is a
connection to the
server. What I get
is an error -197
which is context
not found.
What am I doing
wrong?
Regards
Lawrence
Author
of the award
winning book,
"Kill Me Now!",
amusing stories
from the world
of dating
|
|
|
Re: How to read Structure of a Btrieve File
Thanks Govert, I am currently out but will let you know when I get back.
Regards, Lado
On Wed, Mar 29, 2017 at 4:28 AM, Govert Schipper <govertschipper@...> wrote:
Lado,
If you send me your export file from version 5, I will be happy to convert it to version 8 and than 9 for you.
Govert
Thanks Govert. I'll have to find version 8 first. I believe, I have ver 9 but not 8.
Regards,
Lado
|
|
|
Hi Graham,
I have never seen this error message before, maybe there is a problem with the smtp-server.
Do you get this message for every mail you send, or only for some? Does the mail have an attachment?
Can you try to send a message to the same server using your regular mail client?
Govert
Van: main@magicu-l.groups.io <main@magicu-l.groups.io> namens Graham White <graham.white@...>
Verzonden: woensdag 29 maart 2017 02:33
Aan: main@magicu-l.groups.io
Onderwerp: [magicu-l] Mail Send Error -83
Hi,
We are receiving a MailError -83 Failed to write to data source when doing a Mailsend, does anyone have any idea exactly what this means?
Graham White
|
Software Developer |
Graham.White@... |
EC Credit Control | www.eccreditcontrol.com |
 This
e-mail message and any attachments are confidential to EC Credit Control and subject to legal privilege. If you have received this e-mail in error, please advise the sender immediately and destroy the message and any attachments. If you are not the intended
recipient you are notified that any use, distribution, amendment, copying or any action taken or omitted to be taken in reliance of this message or attachments is prohibited.
|
|
|
Re: How to read Structure of a Btrieve File
Lado,
If you send me your export file from version 5, I will be happy to convert it to version 8 and than 9 for you.
Govert
Van: main@magicu-l.groups.io <main@magicu-l.groups.io> namens Lado Wali <ladowali@...>
Verzonden: dinsdag 28 maart 2017 17:38
Aan: main@magicu-l.groups.io
Onderwerp: Re: [magicu-l] How to read Structure of a Btrieve File
Thanks Govert. I'll have to find version 8 first. I believe, I have ver 9 but not 8.
Regards,
Lado
toggle quoted messageShow quoted text
On Tue, Mar 28, 2017 at 11:10 AM, Govert Schipper
<govertschipper@...> wrote:
Lado,
You don't have to migrate the complete application, just the data repository (and maybe the models if you like). Just use the Magic export function from version 5. Than import this into a new versio 8 application and repeat these steps. Than import into
version 9 and repeat again. Finally import in uniPaaS and use the AddOn utility DDFMaker.
Govert
We are currently using uniPaaS 1.9d and the dos base application is sitting in Magic 5.03c. The dos base magic application is a engineering application and it is already re-written in Labview but because dosbase application has many applications
with hundreds of tables, it require to read the data externally to move to the labview mysql database.
What is the method to convert Magic 5.03c application to uniPaaS 1.9d so I can create DDF files. Can some one please write the steps so I can try that.
Thanks
Lado
|
|
|
Re: Unipaas 1.9H connection to TFS 2015
Jakes du Preez <dupreez.jakes@...>
If you mean the version history... it will belost, yes. I came across these migration walk-through documents that might help if you wish to accomplish that.
On Wed, Mar 29, 2017 at 3:50 AM, katreen_naredo via Groups.Io <katreen_naredo@...> wrote: Hi Jakes,
We tried this option and it worked. However, if we pushed through with this, we will not be able to import the other versions stored in the VSS (for the project) to TFS, right?
|
|
|
Re: Unipaas 1.9H connection to TFS 2015
Hi Jakes,
We tried this option and it worked. However, if we pushed through with this, we will not be able to import the other versions stored in the VSS (for the project) to TFS, right?
|
|
|
Hi,
We are receiving a MailError -83 Failed to write to data source when doing a Mailsend, does anyone have any idea exactly what this means?
Graham White
|
Software Developer |
Graham.White@... |
EC Credit Control | www.eccreditcontrol.com |
 This
e-mail message and any attachments are confidential to EC Credit Control and subject to legal privilege. If you have received this e-mail in error, please advise the sender immediately and destroy the message and any attachments. If you are not the intended
recipient you are notified that any use, distribution, amendment, copying or any action taken or omitted to be taken in reliance of this message or attachments is prohibited.
|
|
|
Re: How to read Structure of a Btrieve File
Thanks Govert. I'll have to find version 8 first. I believe, I have ver 9 but not 8.
Regards, Lado
On Tue, Mar 28, 2017 at 11:10 AM, Govert Schipper <govertschipper@...> wrote:
Lado,
You don't have to migrate the complete application, just the data repository (and maybe the models if you like). Just use the Magic export function from version 5. Than import this into a new versio 8 application and repeat these steps. Than import into
version 9 and repeat again. Finally import in uniPaaS and use the AddOn utility DDFMaker.
Govert
We are currently using uniPaaS 1.9d and the dos base application is sitting in Magic 5.03c. The dos base magic application is a engineering application and it is already re-written in Labview but because dosbase application has many applications
with hundreds of tables, it require to read the data externally to move to the labview mysql database.
What is the method to convert Magic 5.03c application to uniPaaS 1.9d so I can create DDF files. Can some one please write the steps so I can try that.
Thanks
Lado
|
|
|
Re: How to read Structure of a Btrieve File
Thanks Keith, certainly this will help.
Best Regards, Lado
On Tue, Mar 28, 2017 at 11:12 AM, Keith Canniff <kcanniff@...> wrote: Lado, This should do it (actually takes longer to write this up than do it): - Export ‘Files’ from Magic 5.x
- Create a new application in Magic 8 (Settings/Applications)
- Repository Export/Import
- Import 5.xx 4.xx, which should give you a file repository in v8
- Go to the table repository, make sure your cursor is above the first table in the repository and press Ctrl+D (Options/DDF Make). This will create four DDF files in your current default folder (press Ctrl+Z and it will exit to a command line there).
- Put these files in the folder that contains your data files
- Go into PCC, create a new database (give a shorter name with no spaces just to be safe. I keep mine under 10 characters normally but that’s just me) pointing to this folder.
- In PCC go into that database, open up the tables folder and you should see all your tables. If you double click on any table that has data, you should see that data appear in the right frame in PCC. At least this will tell you that the DDFs and the data match up.
If you don’t need anything else, you can stop here. Any application that uses ODBC should now be able to access those tables. If you need this from inside of UP 1.9, continue with the steps below. Just using the DOS files in UP via SQL (relational engine) - Open UP 1.9
- Goto Options/Settings/Databases
- Create a new database where the “Database Name” is set to the database name you created in PCC
- Set the DBMS name to “Pervasive”. Click OK
- Open your application
- Goto the Data Repository and make sure your cursor is above the first entry in the table
- Select Options/Get Definition
- Select the database name you created in #3 above
- For “Tag data sources”, select either ‘All’ if you want all of them, or “Several” to pick from a list
- Once you press Ok, UP will generate the tables and add them to your data repository.
- Browser your tables to see if you can read them.
If you’re using traditional Btrieve: - From Magic v8, Export the data repository
- In Magic 9, create an application and import the above export (Options/Settings/Export Import
- Now do an Options/Settings/Export Import/Export for Version 10.x
- Open Up 1.9
- Press Help and search for “V9Converter” This will give you the instructions to run the command line converter to convert the v9 export you just created to something UP can read
- Once the above step is complete, just import the file created and you’ll have your new file definitions in UP.
Hope this helps. Keith Canniff
President and CTO
IntegrateOne LLC
www.integrateone.com
678-485-8665
From: main@magicu-l.groups.io [mailto:main@....io] On Behalf Of Lado Wali
Sent: Tuesday, March 28, 2017 7:25 AM
To: main@magicu-l.groups.io
Subject: Re: [magicu-l] How to read Structure of a Btrieve File We are currently using uniPaaS 1.9d and the dos base application is sitting in Magic 5.03c. The dos base magic application is a engineering application and it is already re-written in Labview but because dosbase application has many applications with hundreds of tables, it require to read the data externally to move to the labview mysql database. What is the method to convert Magic 5.03c application to uniPaaS 1.9d so I can create DDF files. Can some one please write the steps so I can try that. On Tue, Mar 28, 2017 at 9:16 AM, Rob Westland <magicrob@...> wrote: Hi Lado,
I think the easiest way is to export the data repository from that old dos program and import it in Magic 8, then go to Magic 9 and to XPA.
Rob
Op 28-3-2017 om 14:49 schreef Lado Wali: Govert, Ok, actually we have very old magic application in dos base and it is using bunch of .dat files. I found that I can see the definition in the old dos base magic. The software person wanted to read the data from all those files through the program. Please advise if it can be done through DDF and how to create DDF files out of the old dos base magic and then how to read DDF files? On Tue, Mar 28, 2017 at 7:10 AM, Govert Schipper <govertschipper@...> wrote: Lado, The only thing you know for sure is the total record length (1620 bytes) and that the first field is an integer of 2 bytes, probably just a sequence number. The rest of the record you can try to determine by looking at the contents of the remaining 1618 bytes. Was this file created/populate using Magic? Do you have DDF files? Than you can view the record structure using Pervasive Control Center. Govert
File Statistics for c:\downloads\aaa\aa1\gsfil053.dat Total Number of Records = 30 Available Linked Duplicate Keys = 0 Total Number of Segments = 1 Key Position Type Null Values* ACS Segment Length Flags Unique Values 0 1 1 2 Integer M -- 30 -- * = The values in this column are hexadecimal. From the above report, it looks like one field of integer and I tried putting one field with numeric from 1 to 10 digits and tried opening bit it does not open the file. Do you guess something else? On Mon, Mar 27, 2017 at 5:07 PM, Arch <Arch.Lineberger@...> wrote: Btrieve only knows the file length and key position/type. I'm afraid - short of some utility - you will have to parse the fields as described to specify the fields to Magic.
On Mon, Mar 27, 2017 at 5:00 PM, Lado Wali <ladowali@...> wrote: Arch, I do not have any file structure for ctrl+G but have many btrieve files that I need to open. I am trying right now with Pervasive utility if I can find some information to create the definition. On Mon, Mar 27, 2017 at 4:34 PM, Arch <Arch.Lineberger@...> wrote: From a long time ago by Steve Blank, I believe: Method 11 – Ctl+G in the data dictionary to find the file (Definition Mismatch) 2 - Uncheck "Check definition" and "Check index" and "Change tables in Studio" in the relevant Database Properties. 3 – Add a field at the end of the file in question. 4 – Exit the file and arrow-up, DO NOT CONVERT 5 – Remove the field just added 6 – Exit the file and arrow-up, DO CONVERT 7 – Database/Properties, Check Definition = Yes Chances are, at some point in the record, the fields start looking funky. That is where your mismatch is. Generally a field was added, but the data not converted. Usually the easiest fix is to delete the added field, so everything lines up. Then add the field again, and make sure Magic can reformat the file (or that you have a program to do the reformat). Method 2Use the Pervasive utility called DDF Builder. In addition to showing you the overall record length and index structure that the Maintenance utilities display, DDF Builder will also display actual data to help you figure out where one non-indexed column ends and the next begins. On the downside, DDF Builder is not terribly intuitive to use and is really a 95%-completed product. In the end however, Magic's Definition Mismatch error can only compare overall record length and index structure, so the aforementioned maintenance utilities will reveal enough information to at least solve the Definition Mismatch error. Method 31. Use the Btrieve utilities to get the record length and index information.
2. Add filler to the end of the record as needed so the length matches.
3. Uncheck "Check Definition" in the Database setup. Then use Ctrl-G to look at the data, and keep tweaking it til it looks right. This requires some knowledge of underlying data formats (something us old COBOL programmers do know!). You can set up "filler" fields that are alpha, which makes it easier to see the content. If the record contains "variable data" it's a bit more difficult. However, old files don't have that so much. And if it is there, it will be at the end of the record.
On Mon, Mar 27, 2017 at 4:16 PM, Lado Wali <ladowali@...> wrote: Can Some one help me read a structure of a Btrieve file from the attached report and advise the method of reading it so I can do for other files.
|
|
|
Re: How to read Structure of a Btrieve File
Keith Canniff
Lado, This should do it (actually takes longer to write this up than do it): - Export ‘Files’ from Magic 5.x
- Create a new application in Magic 8 (Settings/Applications)
- Repository Export/Import
- Import 5.xx 4.xx, which should give you a file repository in v8
- Go to the table repository, make sure your cursor is above the first table in the repository and press Ctrl+D (Options/DDF Make). This will create four DDF files in your current default folder (press Ctrl+Z and it will exit to a command line there).
- Put these files in the folder that contains your data files
- Go into PCC, create a new database (give a shorter name with no spaces just to be safe. I keep mine under 10 characters normally but that’s just me) pointing to this folder.
- In PCC go into that database, open up the tables folder and you should see all your tables. If you double click on any table that has data, you should see that data appear in the right frame in PCC. At least this will tell you that the DDFs and the data match up.
If you don’t need anything else, you can stop here. Any application that uses ODBC should now be able to access those tables. If you need this from inside of UP 1.9, continue with the steps below. Just using the DOS files in UP via SQL (relational engine) - Open UP 1.9
- Goto Options/Settings/Databases
- Create a new database where the “Database Name” is set to the database name you created in PCC
- Set the DBMS name to “Pervasive”. Click OK
- Open your application
- Goto the Data Repository and make sure your cursor is above the first entry in the table
- Select Options/Get Definition
- Select the database name you created in #3 above
- For “Tag data sources”, select either ‘All’ if you want all of them, or “Several” to pick from a list
- Once you press Ok, UP will generate the tables and add them to your data repository.
- Browser your tables to see if you can read them.
If you’re using traditional Btrieve: - From Magic v8, Export the data repository
- In Magic 9, create an application and import the above export (Options/Settings/Export Import
- Now do an Options/Settings/Export Import/Export for Version 10.x
- Open Up 1.9
- Press Help and search for “V9Converter” This will give you the instructions to run the command line converter to convert the v9 export you just created to something UP can read
- Once the above step is complete, just import the file created and you’ll have your new file definitions in UP.
Hope this helps. Keith Canniff
President and CTO
IntegrateOne LLC
www.integrateone.com
678-485-8665
From: main@magicu-l.groups.io [mailto:main@magicu-l.groups.io] On Behalf Of Lado Wali
Sent: Tuesday, March 28, 2017 7:25 AM
To: main@magicu-l.groups.io
Subject: Re: [magicu-l] How to read Structure of a Btrieve File We are currently using uniPaaS 1.9d and the dos base application is sitting in Magic 5.03c. The dos base magic application is a engineering application and it is already re-written in Labview but because dosbase application has many applications with hundreds of tables, it require to read the data externally to move to the labview mysql database. What is the method to convert Magic 5.03c application to uniPaaS 1.9d so I can create DDF files. Can some one please write the steps so I can try that. On Tue, Mar 28, 2017 at 9:16 AM, Rob Westland <magicrob@...> wrote: Hi Lado,
I think the easiest way is to export the data repository from that old dos program and import it in Magic 8, then go to Magic 9 and to XPA.
Rob
Op 28-3-2017 om 14:49 schreef Lado Wali: Govert, Ok, actually we have very old magic application in dos base and it is using bunch of .dat files. I found that I can see the definition in the old dos base magic. The software person wanted to read the data from all those files through the program. Please advise if it can be done through DDF and how to create DDF files out of the old dos base magic and then how to read DDF files? On Tue, Mar 28, 2017 at 7:10 AM, Govert Schipper <govertschipper@...> wrote: Lado, The only thing you know for sure is the total record length (1620 bytes) and that the first field is an integer of 2 bytes, probably just a sequence number. The rest of the record you can try to determine by looking at the contents of the remaining 1618 bytes. Was this file created/populate using Magic? Do you have DDF files? Than you can view the record structure using Pervasive Control Center. Govert
File Statistics for c:\downloads\aaa\aa1\gsfil053.dat Total Number of Records = 30 Available Linked Duplicate Keys = 0 Total Number of Segments = 1 Key Position Type Null Values* ACS Segment Length Flags Unique Values 0 1 1 2 Integer M -- 30 -- * = The values in this column are hexadecimal. From the above report, it looks like one field of integer and I tried putting one field with numeric from 1 to 10 digits and tried opening bit it does not open the file. Do you guess something else? On Mon, Mar 27, 2017 at 5:07 PM, Arch <Arch.Lineberger@...> wrote: Btrieve only knows the file length and key position/type. I'm afraid - short of some utility - you will have to parse the fields as described to specify the fields to Magic.
On Mon, Mar 27, 2017 at 5:00 PM, Lado Wali <ladowali@...> wrote: Arch, I do not have any file structure for ctrl+G but have many btrieve files that I need to open. I am trying right now with Pervasive utility if I can find some information to create the definition. On Mon, Mar 27, 2017 at 4:34 PM, Arch <Arch.Lineberger@...> wrote: From a long time ago by Steve Blank, I believe: Method 11 – Ctl+G in the data dictionary to find the file (Definition Mismatch) 2 - Uncheck "Check definition" and "Check index" and "Change tables in Studio" in the relevant Database Properties. 3 – Add a field at the end of the file in question. 4 – Exit the file and arrow-up, DO NOT CONVERT 5 – Remove the field just added 6 – Exit the file and arrow-up, DO CONVERT 7 – Database/Properties, Check Definition = Yes Chances are, at some point in the record, the fields start looking funky. That is where your mismatch is. Generally a field was added, but the data not converted. Usually the easiest fix is to delete the added field, so everything lines up. Then add the field again, and make sure Magic can reformat the file (or that you have a program to do the reformat). Method 2Use the Pervasive utility called DDF Builder. In addition to showing you the overall record length and index structure that the Maintenance utilities display, DDF Builder will also display actual data to help you figure out where one non-indexed column ends and the next begins. On the downside, DDF Builder is not terribly intuitive to use and is really a 95%-completed product. In the end however, Magic's Definition Mismatch error can only compare overall record length and index structure, so the aforementioned maintenance utilities will reveal enough information to at least solve the Definition Mismatch error. Method 31. Use the Btrieve utilities to get the record length and index information.
2. Add filler to the end of the record as needed so the length matches.
3. Uncheck "Check Definition" in the Database setup. Then use Ctrl-G to look at the data, and keep tweaking it til it looks right. This requires some knowledge of underlying data formats (something us old COBOL programmers do know!). You can set up "filler" fields that are alpha, which makes it easier to see the content. If the record contains "variable data" it's a bit more difficult. However, old files don't have that so much. And if it is there, it will be at the end of the record.
On Mon, Mar 27, 2017 at 4:16 PM, Lado Wali <ladowali@...> wrote: Can Some one help me read a structure of a Btrieve file from the attached report and advise the method of reading it so I can do for other files.
|
|
|
