Re: Odd GLOBALS behaviour?

Hi

I had the same problem, my solution make one big globals.4gl merging all
the variables used in all executables.

To avoid confusion and changes in make files I created symbolic links in
each subdirectory to the 'unique merged one'

[cte@adela /]$ ll /u/*/src/*/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/cont/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/cta/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/cte/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/flow/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl
-rw-rw-rw- 1 cte corona 1584 Apr  7 15:05 /u/globals/src/lib/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/mrp/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl
lrwxrwxrwx 1 cte corona   30 Oct  1  2014 /u/nomi/src/lib/globals.4gl ->
/u/globals/src/lib/globals.4gl

I need all the global variables to be pre initialized ....

Regards,

On Wed, Apr 15, 2015 at 9:02 AM, Malcolm Perrior <
[email protected]> wrote:

>  Mike, sorry for the long delay – there’s been other things going on here
> and so on – but yes, this works fine. Our system is a pain because we have
> several hundred executables all in their own subdirectory and all with
> their own globals file; and all of these reference all our 15 or so shared
> libraries which all have their own globals file as well…whoever wrote this
> (and I could name names – they were pretty big in the bespoke insurance
> tools world), burning in hell’s too good for them!
>
>
>
> *From:* [email protected] [mailto:[email protected]] *On Behalf
> Of *Mike Aubury
> *Sent:* 25 March 2015 18:03
> *To:* Malcolm Perrior
> *Cc:* [email protected]
> *Subject:* Re: [Aubit4gl-discuss] Odd GLOBALS behaviour?
>
>
>
> The chances are that its referring to a different globals somewhere
> somehow.
>
> I think i've mentioned before that we try to auto initialize global
> variables to NULL but - this isn't 100% accurate when you define the same
> variable in multiple places.
>
>
>
> Try setting :
>
>
>
>    export NOGLOBALINIT=Y
>
>
>
> and then **recompile** all your programs/libraries
>
>
>
>
>
> The comment for that variable is  :
>
> "if GLOBALS are specified in more than one file - they can overwrite each
> other when being initialized. This shouldn't happen if you specify the
> second (or subsequent) globals via the GLOBALS "filename" - just if they
> are repeated in "GLOBALS .. END GLOBALS" in two separate files"
>
>
>
>
>
> Note - this is a compile time setting - it changes whats actually put in
> the generated .c files
>
>
>
>
>
> On 25 March 2015 at 16:59, Malcolm Perrior <
> [email protected]> wrote:
>
> Or have I not set some variable? I recall something about global
> persistence across links but can’t remember the context.
>
> NB this is under Ubuntu using an informix db and standard 4GL
>
>
>
> What we have is:
>
> A “.a” library built from several 4GLs in path
> “<path1>/<path2>/<path3>/base_lib/”. All the 4GLs in here refer to
> ../../globals.4gl. The makefile also refers to that globals.4gl
>
> A program built from a single 4GL in
> “<path1>/<path4>/<path5>/program.4gl”. The 4GL points to the same globals
> as the library, using “../../<path2>/globals.4gl” (as does its makefile).
>
>
>
> The globals file defines a global variable “prog_name”. When the program
> runs, it calls a get_general_variables() function in one of the library
> 4GLs which uses arg_val(0) to get the program name and populate the global.
>
> (Set font to COURIER to see the below diagram properly)
>
>
>
> <path1>
>
> |
>
> |------/<path2>/
>
> |      |       |--globals.4gl
>
> |      |
>
> |      |-------/<path3>/
>
> |                      |
>
> |                      /<base_lib>/
>
> |                                 |--lib_file_1.4gl
>
> |                                 |--lib_file_2.4gl
>
> |
>
> |------<path4>
>
> |       |
>
> |       |------/<path5>/
>
> |                      |--program_source.4gl
>
>
>
> The program 4GL also contains a “WHENEVER ERROR CALL get_error_action”;
> this refers to a function in a second library 4gl (in the same library
> path), which is called if anything untoward happens (NB it refers to the
> same ../../globals.4gl)
>
>
>
> This works on our current HP-UX host running informix 4gl.
>
>
>
> Under the linux proof-of-concept I’m running, the prog_name variable
> appears to be unset in the case of an error.
>
>
>
> I’ve proved this by inserting “DISPLAY prog_name” statements in each of
> the modules at appropriate times; it’s there in the library 4GL that sets
> it when the get_general_variable() function is called, it’s there in the
> program 4GL after that call, but if I force an error it is NULL in the
> error handler code.
>
>
>
> Do I dimly recall that there’s something I have to do to make global
> variables persistent in a4gl?
>
>
>
> *Malcolm Perrior*
>
> Database Administrator
>
> *T *020 8628 3618
>
>
>
> *Foresters*
>
> Foresters House,
>
> 2 Cromwell Avenue
>
> Bromley BR2 9BF
>
> foresters.com
>
>
>
>
>

------------------------------------------------------------------------------
BPM Camp - Free Virtual Workshop May 6th at 10am PDT/1PM EDT
Develop your own process in accordance with the BPMN 2 standard
Learn Process modeling best practices with Bonita BPM through live exercises
http://www.bonitasoft.com/be-part-of-it/events/bpm-camp-virtual- event?utm_
source=Sourceforge_BPM_Camp_5_6_15&utm_medium=email&utm_campaign=VA_SF

_______________________________________________
Aubit4gl-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/aubit4gl-discuss