From: "Richard Levitte via RT" <r...@openssl.org>

> Stephen Henson has fixed the issues with speed.c and adding cts128 to mks=
> hared.com.  I=20
> have just now committed a fix for the problem with [.crypto]install.com.

   It's a start.  Not very timely, though.

> As for the BN test, I have to wade through older mail to see what was sug=
> gested.  I'll be=20
> back about it.

   That'd be around 6-APR-2010.

   Also, I doubt that the tests are all being run.  It's hard to tell,
because nothing in [.test]tests.com announces a test when it begins. 
(That was a suggestion for 0.9.8j, around 4-FEB-2009.)  I get more
complete results doing this:

--- tests.com_orig      2010-01-24 18:20:32 -0600
+++ tests.com   2010-11-18 15:10:41 -0600
@@ -76,6 +76,7 @@
 $      tests_e = f$element(tests_i,",",tests)
 $      tests_i = tests_i + 1
 $      if tests_e .eqs. "," then goto exit
+$      write sys$output "---> ''tests_e'"
 $      gosub 'tests_e'
 $      goto loop_tests
 $
@@ -246,7 +247,9 @@
 $ test_ssl:
 $      write sys$output "test SSL protocol"
 $      gosub maybe_test_ss
+$      set noon
 $      @testssl.com keyU.ss certU.ss certCA.ss
+$      set on
 $      return
 $ test_ca:
 $      set noon
@@ -276,8 +279,11 @@
 $      then
 $          write sys$output "skipping testtsa.com test -- requires RSA"
 $      else
+$          set noon
 $          @testtsa.com
+$          set on
 $      endif
+$      set default '__here'
 $      return
 $ test_ige: 
 $      write sys$output "Test IGE mode"

I haven't tried to determine if the problems involve expected or
unexpected test failures (or just bad DCL).  Losing the default
directory, as when testtsa.com drops down to [.tsa] and never gets back,
could be helped if each test were run in a spawned sub-process. 
Otherwise, something like that "set default '__here'" could be added
after every dangerous "@".  (Doing it everywhere would be simpler than
trying to determine where it's actually needed, of course.)

   The IGE test seems to work better if the program gets built
(25-JAN-2010):

--- maketests.com_orig  2010-04-13 09:38:39 -0500
+++ maketests.com       2010-11-17 16:35:54 -0600
@@ -105,10 +105,12 @@
               "MDC2TEST,RMDTEST,"+ -
               "RANDTEST,DHTEST,ENGINETEST,"+ -
               "BFTEST,CASTTEST,SSLTEST,EXPTEST,DSATEST,RSA_TEST,"+ -
-              "EVP_TEST,JPAKETEST"
+              "EVP_TEST,IGETEST,JPAKETEST,"+ -
+              "ASN1TEST"
 $! Should we add MTTEST,PQ_TEST,LH_TEST,DIVTEST,TABTEST as well?
 $!
 $! Additional directory information.
+$ T_D_ASN1TEST   := [-.test]
 $ T_D_BNTEST     := [-.crypto.bn]
 $ T_D_ECTEST     := [-.crypto.ec]
 $ T_D_ECDSATEST  := [-.crypto.ecdsa]

   These changes let me get to the test/cms-test.pl problems
(27-APR-2010, 3-MAY-2010).  I didn't actually add a "test_asn1" to
tests.com, either, but ASN1TEST.EXE can be built.

------------------------------------------------------------------------

   Steven M. Schweda               s...@antinode-info
   382 South Warwick Street        (+1) 651-699-9818
   Saint Paul  MN  55105-2547
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       openssl-dev@openssl.org
Automated List Manager                           majord...@openssl.org

Reply via email to