1
Fork 0
mirror of https://git.savannah.gnu.org/git/guile.git synced 2025-06-29 06:20:30 +02:00
guile/test-suite
Andy Wingo b839233282 peval uses effort counters, propagates lambdas more effectively
* module/language/tree-il/optimize.scm (code-contains-calls?): Remove
  this helper, we will deal with recursion when it happens, not after
  the fact.
  (peval): Add keyword args for various size and effort limits.  Instead
  of keeping a call stack, keep a chain of <counter> records, each with
  an abort continuation.  If ever an inlining attempt is taking too
  long, measured in terms of number of trips through the main loop, the
  counter will abort.  Add new contexts, `operator' and `operand'.  They
  have different default size limits.  In the future we should actually
  use the size counter, instead of these heuristics.

  The <lexical-ref> case is smarter now, and tries to avoid propagating
  too much data.  Perhaps it should be dumber though, and use a
  counter.  That would require changes to the environment structure.

  Inline <lambda> applications to <let>, so that we allow residual
  lexical references to have bindings.  Add a `for-operand' helper, and
  use it for the RHS of `let' expressions.  A `let' is an inlined
  `lambda'.

  `Let' and company no longer elide bindings if the result is a
  constant, as the arguments could have effects.  Peval will still do as
  much as it can, though.

* test-suite/tests/tree-il.test ("partial evaluation"): Update the tests
  for the new expectations.  They are uniformly awesomer, with the
  exception of two cases in which pure but not constant data is not
  propagated.
2011-09-25 02:49:02 +02:00
..
lalr Add Boucher's lalr-scm' as the (system base lalr)' module. 2010-03-31 00:41:59 +02:00
standalone fix scm_to_latin1_stringn for substrings 2011-09-10 11:38:25 -07:00
tests peval uses effort counters, propagates lambdas more effectively 2011-09-25 02:49:02 +02:00
vm Update (ice-9 match) from Chibi-Scheme. 2011-09-03 22:18:02 +02:00
ChangeLog-2008 Rename ChangeLog' files to ChangeLog-2008'. 2008-09-12 21:49:58 +02:00
guile-test make guile-test work without configuration 2010-12-07 13:21:00 +01:00
lib.scm Misc textual editing 2011-02-13 22:13:33 +00:00
Makefile.am Update (ice-9 match) from Chibi-Scheme. 2011-09-03 22:18:02 +02:00
README Revert "Note need for subscription to bug-guile@gnu.org." 2008-12-10 19:07:14 +00:00

This directory contains some tests for Guile, and some generic test
support code.

To run these tests, you will need a version of Guile more recent than
15 Feb 1999 --- the tests use the (ice-9 and-let*) and (ice-9
getopt-long) modules, which were added to Guile around then.

For information about how to run the test suite, read the usage
instructions in the comments at the top of the guile-test script.

You can reference the file `lib.scm' from your own code as the module
(test-suite lib); it also has comments at the top and before each
function explaining what's going on.

Please write more Guile tests, and send them to bug-guile@gnu.org.
We'll merge them into the distribution.  All test suites must be
licensed for our use under the GPL, but I don't think I'm going to
collect assignment papers for them.



Some test suite philosophy:

GDB has an extensive test suite --- around 6300 tests.  Every time the
test suite catches a bug, it's great.

GDB is so complicated that folks are often unable to get a solid
understanding of the code before making a change --- we just don't
have time.  You'll see people say things like, "Here's a fix for X; it
doesn't cause any regressions."  The subtext is, I made a change that
looks reasonable, and the test suite didn't complain, so it must be
okay.

I think this is terrible, because it suggests that the writer is using
the test suite as a substitute for having a rock-solid explanation of
why their changes are correct.  The problem is that any test suite is
woefully incomplete.  Diligent reasoning about code can catch corner
conditions or limitations that no test suite will ever find.



Jim's rule for test suites:

Every test suite failure should be a complete, mysterious surprise,
never a possibility you were prepared for.  Any other attitude
indicates that you're using the test suite as a crutch, which you need
only because your understanding is weak.