On Wed, 10 Jan 2018 15:12:17 -0500, John Eells wrote:

>Whether differently-sized modules are inequal cannot be determined from
>size alone when measured on disk (at least) for the reasons I posted
>earlier.  The same is true of record-based, LOAD-based, and even Binder
>API-based binary compares.
>
>Comparing LOADed modules is also not reliable in the general case for
>somewhat different reasons that I mentioned in the same post.  In
>addition to those reasons, it occured to me that different ordering can
>cause different amounts of storage to be consumed to maintain alignment
>boundaries (and the owner of the Binder confirms that).
>
Relink one with ORDER statements to match the other, then compare?
Oops!  What about anonymous CSECTs?

>(This is *way* harder than it looks at first glance.)
>
Yup.  And even IEBCOPY modifies the FAMS timestamp.  As if that mattered.
(Am I supposed to know about FAMS timestamps?)

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to