openssl/crypto/dso
Geoff Thorpe 75a382bd90 None of the DSO_METHOD's were handling anything except generic messages.
These are now processed inside DSO_ctrl() itself.
2000-10-08 22:36:49 +00:00
..
.cvsignore Ignore lib and Makefile.save. 2000-04-14 23:37:44 +00:00
dso_dl.c None of the DSO_METHOD's were handling anything except generic messages. 2000-10-08 22:36:49 +00:00
dso_dlfcn.c None of the DSO_METHOD's were handling anything except generic messages. 2000-10-08 22:36:49 +00:00
dso_err.c A DSO method for VMS was missing, and I had the code lying around... 2000-09-15 21:22:50 +00:00
dso_lib.c Time to get rid of some rather silly code duplication - some DSO_ctrl() 2000-10-08 22:32:57 +00:00
dso_null.c Currently the DSO_METHOD interface has one entry point to bind all 2000-06-16 10:45:36 +00:00
dso_openssl.c A DSO method for VMS was missing, and I had the code lying around... 2000-09-15 21:22:50 +00:00
dso_vms.c None of the DSO_METHOD's were handling anything except generic messages. 2000-10-08 22:36:49 +00:00
dso_win32.c None of the DSO_METHOD's were handling anything except generic messages. 2000-10-08 22:36:49 +00:00
dso.h A DSO method for VMS was missing, and I had the code lying around... 2000-09-15 21:22:50 +00:00
Makefile.ssl 'ranlib' doesn't always run on some systems. That's actually 2000-09-25 08:53:15 +00:00
README Currently the DSO_METHOD interface has one entry point to bind all 2000-06-16 10:45:36 +00:00

TODO
----

Find a way where name-translation can be done in a way that is
sensitive to particular methods (ie. generic code could still do
different path/filename substitutions on win32 to what it does on
*nix) but doesn't assume some canonical form. Already one case
exists where the "blah -> (libblah.so,blah.dll)" mapping doesn't
suffice. I suspect a callback with an enumerated (or string?)
parameter could be the way to go here ... DSO_ctrl the callback
into place and it can be invoked to handle name translation with
some clue to the calling code as to what kind of system it is.

NOTES
-----

I've checked out HPUX (well, version 11 at least) and shl_t is
a pointer type so it's safe to use in the way it has been in
dso_dl.c. On the other hand, HPUX11 support dlfcn too and
according to their man page, prefer developers to move to that.
I'll leave Richard's changes there as I guess dso_dl is needed
for HPUX10.20.