Am 27.01.2014 20:44, schrieb Bernd Oppolzer:
Am 27.01.2014 20:27, schrieb Paul Gilmartin:
On Mon, 27 Jan 2014 19:33:25 +0100, Bernd Oppolzer wrote:
... (if you manage to LINK between LE modules
with the help of MVS LINK or similar mechanisms, which is possible -
for example: we do all linkage between our LE modules, no matter
which language,
with a home grown mechanism which relies on MVS LOAD macro and BASR,
basically).
You mean IBM provides no support for LINKing to an LE module!? That
seems
outrageous in the year 2014. How about ATTACH? How does initiator
do it?
-- gil
No, I meant that normally you use LE functions to do the linkage that use
MVS LINK under the cover. But if you really want to do it, you can manage
to do the linkage between LE modules without LE functions, but with
functions
of your own, that use pure MVS LINK.
Maybe it is necessary to be more precise:
if the (dynamically) called LE function or module has some of the newer
properties (like DLL, RENT, LONGNAME), you will need CEEFETCH or similar
mechanisms to do the linkage.
But if the called LE function is a "classical" load module involving only
NORENT functions (written in C or PL/1 for example, without the options
mentioned above), you will be able to do LOAD/BALR, LINK, ATTACH etc.
without the LE specific actions (for example, WSA initialization).
This is NOT related to enclave initialization; it's all true for
subprograms, too.
Kind regards
Bernd
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
