duplicate variables in CL, msg CPD0858 "Variable or data area &1 declared previously."

FWiW, TWiMC:

Seems that at least since v6r1, what was previously diagnosed with CPD0858 "Variable or data area &1 declared previously.", may no longer get diagnosed. Seems the intention, since some compiler change [probably in support of the Include CL Source (INCLUDE) CL command as compiler directive available starting with IBM i 6.1], is to operate quite literally, according to the implication of the Recovery text, whereby either action of "Remove one of the conflicting declarations or change the attributes on both declarations so they are the same" should suffice to allow the compile. Thus the Control Language compiler [for any of: Create Bound CL Program (CRTBNDCL), Create CL Module (CRTCLMOD), Create CL Program (CRTCLPGM)] now just ignores a duplicate-named declarative, without any indication(s) being logged, given each of the same-named variable has [exactly] matching attributes.

As late as v5r3, the CL compiler was [still; ¿and AFaIK, had always been, since the S/38?] diagnosing the condition of a duplicate CL Variable Name (VAR) of a Declare Variable (DCL) with the sev-30 failure msgCPD0858, despite matching attributes across the multiply-declared variable name; i.e. any duplicate-named declarative was being diagnosed, *irrespective the attributes*. Given the recovery text has not changed between those releases, I suppose that the error was being diagnosed on v5r3 could have been a defect on that release; and any other [earlier] releases exhibiting the same behavior.?

I felt the need to investigate after I stupidly had coded an Initial Value (VALUE) on the Declare Variable (DCL) for a Parameter (PARM), but having left the correct\proper DCL in the source, quite conspicuously, on the next line of source. The incorrectly coded declarative with the initialization was properly diagnosed with the sev-20 msg CPD0702 "Initial value not allowed for variable passed into program." Because the VALUE() specification was disallowed and essentially ignored by the compiler [though oddly, the recovery text for that non-terminating error suggests deleting and re-creating the program? As if!], that conspicuous _mismatch_ of initializations between the two declares, was undetected and thus [not surprisingly] also not diagnosed with CPD0858 :-)

Only thing left to do, is find that darned Declare Data Area (DCLDTAARA) command ;-) FWiW, the command is for the S38 environment and thus only [in\available-from QSYS38] and thus Source Type (SRCTYPE) of CLP38 is all that can use the feature for which there are also both of Receive Data Area (RCVDTAARA) and Send Data Area (SNDDTAARA). The implementation\effects for the feature, are much like what occurs for the Declare File (DCLF), Send File (SNDF) and Receive File (RCVF) whereby the variable generated for the declare can be read and written. The message probably should have been modified long ago, to remove the "or data area" from the first level text; leaving reference in the second-level text and late in the recovery text.