Hi all,
I don't know much about JPL, but I was faced with this problem a while
back with my
Inline::Java project. I managed to overcome the problem by copying the
environment
myself before calling perl_parse:
char *args[] = {"inline-java", "-e1"} ;
int envl = 0 ;
int i = 0 ;
char **envdup = NULL ;
/* This will leak, but it's a one shot... */
for (i = 0 ; environ[i] != NULL ; i++){
envl++ ;
}
envdup = (char **)calloc(envl + 1, sizeof(char *)) ;
for (i = 0 ; i < envl ; i++){
envdup[i] = strdup(environ[i]) ;
}
interp = perl_alloc() ;
perl_construct(interp) ;
perl_parse(interp, xs_init, 2, args, envdup) ;
perl_run(interp) ;
It seemed to work for me without having to patch or
recompile perl (but maybe my perl was already patched? I use
just a standard RedHat 9.0 perl).
Maybe this soultion be applied to JPL?
Patrick
----- Original Message -----
>
> Is there any chance of Jarkko's core patch for JPL (Larry's
> Java Perl Lingo) making it back into core..?
>
> History: Java (jdk 1.2+) stopped tolerating Perl's environment
> diddling so Jarkko wrote a patch to just dup the environment so
> Perl can diddle its private copy. Java was happy because the
> original env was remained pristine, JPL kept working, and all
> was right with the world.
>
> The original patch made into 5.6.1 but not in subsequent
> versions that I can detect since I don't see the JPL or
> NEED_DUP_ENVIRON_FOR_MODIFY defines. Jarkko updated the
> patch for 5.8.1 and it's working there. However, I wonder
> if its not making the cut after 5.6.1 was oversight or
> deliberate euthanasia...
>
> We're worried that Jarkko's 5.8.1 patch will eventually stop
> working altogether as Perl marches on. JPL's heavily used
> here in several large legacy Java applications so its loss
> would cause major pain. A couple of us would gladly act as
> custodians for the code and try to pitch in in whatever way
> we're able.
>
> Rgds,
> --
> Charles DeRykus
>
>
I don't know much about JPL, but I was faced with this problem a while
back with my
Inline::Java project. I managed to overcome the problem by copying the
environment
myself before calling perl_parse:
char *args[] = {"inline-java", "-e1"} ;
int envl = 0 ;
int i = 0 ;
char **envdup = NULL ;
/* This will leak, but it's a one shot... */
for (i = 0 ; environ[i] != NULL ; i++){
envl++ ;
}
envdup = (char **)calloc(envl + 1, sizeof(char *)) ;
for (i = 0 ; i < envl ; i++){
envdup[i] = strdup(environ[i]) ;
}
interp = perl_alloc() ;
perl_construct(interp) ;
perl_parse(interp, xs_init, 2, args, envdup) ;
perl_run(interp) ;
It seemed to work for me without having to patch or
recompile perl (but maybe my perl was already patched? I use
just a standard RedHat 9.0 perl).
Maybe this soultion be applied to JPL?
Patrick
----- Original Message -----
>
> Is there any chance of Jarkko's core patch for JPL (Larry's
> Java Perl Lingo) making it back into core..?
>
> History: Java (jdk 1.2+) stopped tolerating Perl's environment
> diddling so Jarkko wrote a patch to just dup the environment so
> Perl can diddle its private copy. Java was happy because the
> original env was remained pristine, JPL kept working, and all
> was right with the world.
>
> The original patch made into 5.6.1 but not in subsequent
> versions that I can detect since I don't see the JPL or
> NEED_DUP_ENVIRON_FOR_MODIFY defines. Jarkko updated the
> patch for 5.8.1 and it's working there. However, I wonder
> if its not making the cut after 5.6.1 was oversight or
> deliberate euthanasia...
>
> We're worried that Jarkko's 5.8.1 patch will eventually stop
> working altogether as Perl marches on. JPL's heavily used
> here in several large legacy Java applications so its loss
> would cause major pain. A couple of us would gladly act as
> custodians for the code and try to pitch in in whatever way
> we're able.
>
> Rgds,
> --
> Charles DeRykus
>
>