summaryrefslogtreecommitdiffstats
path: root/c/src/ada-tests/sptests/sp19/README
diff options
context:
space:
mode:
Diffstat (limited to 'c/src/ada-tests/sptests/sp19/README')
-rw-r--r--c/src/ada-tests/sptests/sp19/README23
1 files changed, 23 insertions, 0 deletions
diff --git a/c/src/ada-tests/sptests/sp19/README b/c/src/ada-tests/sptests/sp19/README
new file mode 100644
index 0000000000..07d60c5bc4
--- /dev/null
+++ b/c/src/ada-tests/sptests/sp19/README
@@ -0,0 +1,23 @@
+#
+# $Id$
+#
+
+This is a "problem" test in Ada versus the C equivalent. In C,
+we used macros to :
+
+ + declare a LOT of local integer and floating point variables
+ + load them with values
+ + check them at certain times
+ + to avoid floating point operations on CPUs w/o hardware support
+
+The macros are not equivalent to subprograms doing the same thing.
+We are trying to insure that the register set is fully utilized and
+checking that it survives context switches. If the routines are
+subprograms, then calling conventions and scoping rules apply. Thus
+the variables do not exist throughout the entire life of the task
+as in the C test.
+
+It seems to point out that although the conditional compilation
+and macro capabilities of C are often abused, they do provide
+capabilities which are difficult if not impossible to mimic entirely
+with subprograms.