Re: Script locals in library stack script

[J. Landman Gay via use-livecode]

On 2/10/17 2:22 PM, J. Landman Gay via use-livecode wrote:

I have set the LC preference to preserve variables. In the past I have
rarely seen a script local lose its value, and never if there has been
no script compile. I don't know exactly when it started but I was in
8.1.3 rc 1 when I noticed it.


In case anyone has been losing sleep over this, it was indeed me. Two 
weeks later while revising my scripts I found the culprit line, 
sandwiched between two large blocks of commented-out code, lurking with 
its head down and its eyes closed. But there was a snicker on its face.


I hit it over the head, put it in a gunny sack, and threw it in the 
river. Our river is frozen right now so I hope it hurt.


--
Jacqueline Landman Gay | jac...@hyperactivesw.com
HyperActive Software   | http://www.hyperactivesw.com

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Bob Sneidar via use-livecode]

I mised that this was on a web server. I should pay more attention. 

Bob S


> On Feb 10, 2017, at 20:03 , J. Landman Gay via use-livecode 
>  wrote:
> 
> I know, that's why I didn't report it. I scoured the script looking for a 
> difference but I couldn't find anything, the two queries were executed by the 
> same handler. I looked for anything that would empty the variable and that 
> wasn't there either.


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[J. Landman Gay via use-livecode]

I know, that's why I didn't report it. I scoured the script looking for 
a difference but I couldn't find anything, the two queries were executed 
by the same handler. I looked for anything that would empty the variable 
and that wasn't there either.


My current theory is creeping stack corruption. The stack ran with only 
this one oddity, but when I tried to reload it from disk later it was 
reported as corrupted and wouldn't open. I had to open it in BBEdit and 
use a days-old backup in LC to paste in all the changes. Apparently it 
had corrupted several days before, but because I'd left the Mac on with 
the stack open I didn't know.


I've sent the working stack and the corrupted one to bugzilla though I 
don't know if the team can make anything of it. The missing variable 
appeared in LC 8.1.2 so I switched over to LC 9 to see if that helped. 
That's when the stack gave up. I think I'll stay away from 9 until it's 
out of dp, but there's something going on with LC 8 too in some unique 
circumstance. I haven't had any trouble with 8 before now.


On 2/10/17 8:47 PM, Bob Sneidar via use-livecode wrote:

I don't mean to oversimplify, but when customers call complaing that only one 
of several scan to SMB registrations is failing, I have to get them to see that 
if all the others are not failing ever, it cannot be a problem with the copier.

If only one script local is getting reset, then it cannot be that LC is 
resetting script locals because they would ALL reset, not just one.

It may be possible to set a breakpoint on the script local to trigger when it 
becomes empty.

Bob S



On Feb 10, 2017, at 12:22 , J. Landman Gay via use-livecode 
 wrote:

The extremely odd thing was that the script had three script locals, and only 
one of them lost its value. The other two were fine. It was impossible to track 
down.



___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode




--
Jacqueline Landman Gay | jac...@hyperactivesw.com
HyperActive Software   | http://www.hyperactivesw.com

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Bob Sneidar via use-livecode]

I don't mean to oversimplify, but when customers call complaing that only one 
of several scan to SMB registrations is failing, I have to get them to see that 
if all the others are not failing ever, it cannot be a problem with the copier. 

If only one script local is getting reset, then it cannot be that LC is 
resetting script locals because they would ALL reset, not just one. 

It may be possible to set a breakpoint on the script local to trigger when it 
becomes empty. 

Bob S


> On Feb 10, 2017, at 12:22 , J. Landman Gay via use-livecode 
>  wrote:
> 
> The extremely odd thing was that the script had three script locals, and only 
> one of them lost its value. The other two were fine. It was impossible to 
> track down.


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Bob Hall via use-livecode]

 —>There used to be a way to do that but I can't remember how, or else I forgot 
it on purpose because it didn't work.

I think this is how you would do that…

in IDE, debug menu->Add variable watch

select the library stack and sVarA in the pop up.

What I haven’t tested is the Condition. Maybe "is empty"

You could also add "if sVarA is empty then breakpoint” at interesting points in 
the script. 

Bob Hall


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[J. Landman Gay via use-livecode]

On 2/10/17 5:56 AM, Bob Hall via use-livecode wrote:

I’ve been wondering this myself. I’m going to try to create a recipe
but it looks like the TTL of script local variables is different than
what I thought I knew it to be. I “think” I started see this behavior
about 8.1.0-ish timeframe but not sure.

In the past few months I started to put everything into properties as
I found that my understanding of how script local variables behaved
was different then how they do.


I have set the LC preference to preserve variables. In the past I have 
rarely seen a script local lose its value, and never if there has been 
no script compile. I don't know exactly when it started but I was in 
8.1.3 rc 1 when I noticed it. I spent two days trying to fix the 
problem, eventually gave up and worked around it. The extremely odd 
thing was that the script had three script locals, and only one of them 
lost its value. The other two were fine. It was impossible to track down.


Here is what I was doing:

Stack Main has a substack "InternetLib" that contains all handlers that 
deal with server connections. It is put in use when Main opens. 
InternetLib handles server queries and returns the retrieved data. That 
all works fine. It also has a "setter" handler that can be used to store 
data in a script local variable in its own script, and a "getter" 
handler to return data from the script local.


What happened:

1. Stack Main calls InternetLib to do a query.
2. Stack Main manipulates the returned data, parses out what it needs, 
and calls the "setter" handler to store it in the script local "sVarA" 
in InternetLib.
3. Immediately after the handler ends, Stack Main calls InternetLib to 
do a second query.
4. Stack Main uses the returned data and calls the "setter" handler to 
store the value in InternetLib script local "sVarB".

5. Later, stack Main tries to retrieve a value from sVarA. sVarA is empty.
6. Retrieving data from sVarB is always available.

The third script local, sVarC, always retains its value. It is not 
involved with the above process, it is set earlier when Main is opened.


When tracing through the handlers, it appears that sVarA loses its value 
some time between the first query and the second. What I need to do is 
find a way to track sVarA to see when it changes. There used to be a way 
to do that but I can't remember how, or else I forgot it on purpose 
because it didn't work.


--
Jacqueline Landman Gay | jac...@hyperactivesw.com
HyperActive Software   | http://www.hyperactivesw.com


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Richard Gaskin via use-livecode]

Bob Hall wrote:

> So I guess I need to understand what is the TTL definition of a
> script local variable?

In the engine, script-local variables will retain their values 
throughout the session.


If a script is edited, however, by default the recompilation of the 
script causes the script-local variables to be cleared.


This default can be overridden with the preserveVariables property. 
False by default, when set to true script-local variables will retain 
their values between compilations of the script they appear in.


I had previously thought that the preserveVariables was a global 
property, which would mean its value is preserved throughout the 
session, and that the "Variable Preservation" checkbox in the "Script 
Editor" section of Prefs reflected that value.


However, the Dictionary does not specify that the preserveVariables is a 
global property, and there appears to be no relationship between the 
seemingly-related Prefs checkbox and this property - recipe:


1. In the Message Box, run: put the preserveVariables
2. In the MB, set the preserveVariables to the opposite
3. Open the Prefs window, note the checkbox
4. Close the Prefs window
5. In the MB, set the preserveVariables to its opposite
6. Open the Prefs window, not that the checkbox retains the true/false 
state seen before, even though the property itself has changed.


The Dictionary does say:

   The preserveVariables property is provided as a background
   compatibility aid. It should not, in general, be used in
   user scripts as the IDE automatically handles preservation
   of variables via the Variable Preservation option in the
   preferences and Script Editor Script menu.

So while I once thought I'd understood both the property and the 
seemingly-related Prefs setting, apparently I do not.


Guidance from the IDE team would be welcome here.

--
 Richard Gaskin
 Fourth World Systems
 Software Design and Development for the Desktop, Mobile, and the Web
 
 ambassa...@fourthworld.comhttp://www.FourthWorld.com

___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Mike Kerner via use-livecode]

This might explain the bizarre bug I'm experiencing with one of Monte's
externals.  I manage the calls to the various routines for the external
from a library stack.  Periodically, I end up somewhere in the code that I
shouldn't be able to get to because of the values of various variables, and
the only way I could get there would be if the values were somehow wiped.
Hm.  Maybe it's not my code.  Time to go sleuthing...

On Fri, Feb 10, 2017 at 6:56 AM, Bob Hall via use-livecode <
use-livecode@lists.runrev.com> wrote:

> I’ve been wondering this myself. I’m going to try to create a recipe but
> it looks like the TTL of script local variables is different than what I
> thought I knew it to be. I “think” I started see this behavior about
> 8.1.0-ish timeframe but not sure.
>
> In the past few months I started to put everything into properties as I
> found that my understanding of how script local variables behaved was
> different then how they do. I had chalked it up to getting old and just
> forgetting how things work.
>
> So I guess I need to understand what is the TTL definition of a script
> local variable? I was under the impression that if I have setters/getters
> for the variables and set their value, as long as that Script remains in
> memory, the script local variable retains it’s value. I do not see that for
> sure if I set the script local var in the libraryStack message in 8.1.2 or
> 8.1.3.
>
> Curious if you’ve figured out anything about this since the original note.
>
> Bob Hall
>
> >
> > Is this something others have seen? LC 8.1.3 rc 1.
> >
>
>
> ___
> use-livecode mailing list
> use-livecode@lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your
> subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-livecode
>



-- 
On the first day, God created the heavens and the Earth
On the second day, God created the oceans.
On the third day, God put the animals on hold for a few hours,
   and did a little diving.
And God said, "This is good."
___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Bob Hall via use-livecode]

I’ve been wondering this myself. I’m going to try to create a recipe but it 
looks like the TTL of script local variables is different than what I thought I 
knew it to be. I “think” I started see this behavior about 8.1.0-ish timeframe 
but not sure. 

In the past few months I started to put everything into properties as I found 
that my understanding of how script local variables behaved was different then 
how they do. I had chalked it up to getting old and just forgetting how things 
work.

So I guess I need to understand what is the TTL definition of a script local 
variable? I was under the impression that if I have setters/getters for the 
variables and set their value, as long as that Script remains in memory, the 
script local variable retains it’s value. I do not see that for sure if I set 
the script local var in the libraryStack message in 8.1.2 or 8.1.3.

Curious if you’ve figured out anything about this since the original note.

Bob Hall

> 
> Is this something others have seen? LC 8.1.3 rc 1.
> 


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[J. Landman Gay via use-livecode]

On 2/8/17 4:52 PM, Monte Goulding via use-livecode wrote:



On 9 Feb 2017, at 9:32 am, J. Landman Gay via use-livecode
 wrote:

I have a splash stack that opens a data stack from a server. The
data stack has a substack which is used as a library with "start
using".

The handlers in the library work okay. Some of them store data in
script local variables in the library script. The script locals do
not retain their values, they are almost always empty when the
library tries to access them later.

Is this something others have seen? LC 8.1.3 rc 1.


I’m guessing you don’t mean in a separate request on the server. I’m
not aware of anything special about the way server does stuff with
variables so someone would need to chase this down based on a bug
report I think.


Right, after downloading the data stack, for this stuff it's all local 
messaging. I'll try to track it down before submitting a bug, I just 
wondered if this was a known thing. I'm seeing other problems that 
indicate the message path isn't what I expect.


--
Jacqueline Landman Gay | jac...@hyperactivesw.com
HyperActive Software   | http://www.hyperactivesw.com


___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode

Re: Script locals in library stack script

[Monte Goulding via use-livecode]

> On 9 Feb 2017, at 9:32 am, J. Landman Gay via use-livecode 
>  wrote:
> 
> I have a splash stack that opens a data stack from a server. The data stack 
> has a substack which is used as a library with "start using".
> 
> The handlers in the library work okay. Some of them store data in script 
> local variables in the library script. The script locals do not retain their 
> values, they are almost always empty when the library tries to access them 
> later.
> 
> Is this something others have seen? LC 8.1.3 rc 1.

I’m guessing you don’t mean in a separate request on the server. I’m not aware 
of anything special about the way server does stuff with variables so someone 
would need to chase this down based on a bug report I think.

Cheers

Monte
___
use-livecode mailing list
use-livecode@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription 
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode