| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
a system/superuser binary. At the same time, move the source code from its
current lib/libssl/src/apps location to a more appropriate home under
usr.bin/openssl.
ok deraadt@ miod@
|
| |
|
|
|
|
|
|
| |
load_config() once when we start.
ok miod@
|
| |
|
|
|
|
|
|
| |
are the current range checks. Help from millert and lteo. Please test
now that it is deployed and let us know if any numbers are off..
ok lteo
|
| |
|
| |
|
|
|
|
| |
ok miod
|
|
|
|
|
|
|
|
|
| |
potential integer overflows easily changed into an allocation return
of NULL, with errno nicely set if need be. checks for an allocations
returning NULL are commonplace, or if the object is dereferenced
(quite normal) will result in a nice fault which can be detected &
repaired properly.
ok tedu
|
|
|
|
|
|
| |
this again in each app.
ok miod@
|
|
|
|
| |
before calling free - of some course parts of the code already did this.
|
|
|
|
|
|
|
|
| |
Currently "apps.h" needs to be included before any of the openssl headers
and there are several cases where the code depends on headers that are
included by something included by apps.h...
More untangling saved for later on.
|
|
|
|
|
|
|
|
|
|
|
|
| |
in each application (since there is no longer a non-monolithic mode).
In typical OpenSSL fashion, the code is inconsistent and there are multiple
ways that bio_err was initialised - none of them actually checked to see if
the initialisation actually succeeded. Additionally, it is worth noting
that in at least two cases bio_err was also being used before it would have
been initialised.
ok miod@
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
than once.
ok beck@
|
|
|
|
|
|
|
| |
signal(SIGPIPE, SIG_IGN) and apps_shutdown is a no-op. So just do that
instead.
ok beck@
|
|
|
|
|
|
| |
do not use lint.
ok miod@
|
|
|
|
| |
ok deraadt@
|
| |
|
|
|
|
| |
ok tedu@
|
| |
|
|
|
|
| |
ok miod@
|
| |
|
|
|
|
|
|
|
|
| |
USE_SOCKETS is unrelated to using sockets, but just pulls in .h files. It
makes every file buy a kitchen sink, because 11 files forgot to.
EXIT() is really exit(), a gentle surprise
but... OPENSSL_EXIT() is really just return(), because noone compiles the
openssl command non-monolithic anymore
|
|
|
|
|
|
|
|
| |
OPENSSL_foo wrappers. This changes:
OPENSSL_malloc->malloc
OPENSSL_free->free
OPENSSL_relloc->realloc
OPENSSL_freeFunc->free
|
|
|
|
| |
a maze of conditional #define's
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
so that its lovecraftian horror is not forever lost, i reproduce below
a comment from the deleted code.
/* 2011-03-22 SMS.
* If we have 32-bit pointers everywhere, then we're safe, and
* we bypass this mess, as on non-VMS systems. (See ARGV,
* above.)
* Problem 1: Compaq/HP C before V7.3 always used 32-bit
* pointers for argv[].
* Fix 1: For a 32-bit argv[], when we're using 64-bit pointers
* everywhere else, we always allocate and use a 64-bit
* duplicate of argv[].
* Problem 2: Compaq/HP C V7.3 (Alpha, IA64) before ECO1 failed
* to NULL-terminate a 64-bit argv[]. (As this was written, the
* compiler ECO was available only on IA64.)
* Fix 2: Unless advised not to (VMS_TRUST_ARGV), we test a
* 64-bit argv[argc] for NULL, and, if necessary, use a
* (properly) NULL-terminated (64-bit) duplicate of argv[].
* The same code is used in either case to duplicate argv[].
* Some of these decisions could be handled in preprocessing,
* but the code tends to get even uglier, and the penalty for
* deciding at compile- or run-time is tiny.
*/
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
correctly autogenerate obj_mac.h
|
| |
|
|
|
|
|
| |
Note that this is a maintenence release, API's appear *not* to have changed.
As such, I have only increased the minor number on these libraries
|
| |
|
| |
|
| |
|