Under Windows, there seems to be a problem relinking fips_premain_dso
because that file is locked.  Changing from backtick op to using
system() with redirection and reading the hash from the output file
seems to fix the problem.

In an ideal world, there should be no difference, as a command in a
backtick op should terminate before the backtick returns, same as it
does with system().  We suspect, though, that the loaded binary is
cached by Windows for a little while, and that reading the output from
a file provides enough delay for the lock to drop before we try to
relink.

alert.

Submitted by: Rob Austein <sra@hactrn.net>
Reviewed by: steve

Fix corner cases in RFC3779 code.

must escape the escape character itself (backslash).

Submitted by: Rob Austein <sra@hactrn.net>
Reviewed by: steve

Use OPENSSL_assert() instead of assert().

Submitted by: Mark Phalan <mark.phalan@oracle.com>
Reviewed by: steve

Use L suffix in version number.

[CVS head and later branches have this since revision 1.7 of this file.]

Submitted by: Chandler Carruth (Google)

Submitted by: Stefan Birrer <stefan.birrer@adnovum.ch>
Reviewed by: steve

Correct SKM_ASN1_SET_OF_d2i macro.

Reviewed by: steve

If store is NULL set flags correctly.

Submitted by: Alexei Khlebnikov <alexei.khlebnikov@opera.com>
Reviewed by: steve

OOM checking. Leak in OOM fix. Fall-through comment. Duplicate code
elimination.

Submitted by: Mounir IDRASSI <mounir.idrassi@idrix.net>
Reviewed by: steve

Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939

are read correctly for non-blocking I/O.

this means that some implementations will be used automatically, e.g. aesni,
we do this for cryptodev anyway.

Setup cpuid in ENGINE_load_builtin_engines() too as some ENGINEs use it.

Submitted by: Emilia Kapser (Google)

Submitted by: Antony, Benoy <bantony@ebay.com>
Approved by: steve@openssl.org

Fix bug in AES wrap code when t > 0xff.