- Feb 22, 2010
-
-
Francois-Rene Rideau authored
-
Robert P. Goldman authored
COMPONENT-NAME-TO-PATHNAME-COMPONENTS.
-
Robert P. Goldman authored
-
- Feb 21, 2010
-
-
Robert P. Goldman authored
-
- Feb 19, 2010
-
-
Robert P. Goldman authored
-
Francois-Rene Rideau authored
(Thanks to janderson for the bug report and patch.)
-
- Feb 18, 2010
-
-
Francois-Rene Rideau authored
Reorder functions to reduce SBCL style-warnings.
-
- Feb 16, 2010
-
-
Francois-Rene Rideau authored
* rename ASDF-EXTENSIONS into ASDF-UTILITIES (keep the old as nickname) and update it to contain all the modern conveniences. * deprecate PATHNAME-SANS-NAME+TYPE, use PATHNAME-DIRECTORY-PATHNAME. * INITIALIZE-SOURCE-REGISTRY now takes an optional parameter that allows the user to provide his override. * Refactor PROCESS-SOURCE-REGISTRY and friends to make things more XCVB-ready. * Change optimize speed setting to 3 to prevent a weird compiler hang in SBCL at (speed 2) (debug 2) (safety 2).
-
- Feb 13, 2010
-
-
Francois-Rene Rideau authored
Also, add :execute to his eval-when clauses, add some documentation to asdf-version, and change the default optimize level.
-
Francois-Rene Rideau authored
Thanks to tcr for suggestions.
- Feb 12, 2010
-
-
Robert P. Goldman authored
-
Robert P. Goldman authored
This solution, unlike the earlier candidate, does not make /systems/ re<operate> when their dependencies are triggered, but only /modules/ that are not systems. The intended result is to trigger the behavior only for INTRA-system dependencies.
-
Robert P. Goldman authored
Some introspection (and test engineering) revealed that in some circumstances the value of the *forcing* dynamic variable could reach either too far (upstream into dependencies) or not far enough (not into nested modules). This patch fixes that.
-
Robert P. Goldman authored
Was loading wrong system once.
-
Robert P. Goldman authored
Examining the new TRAVERSE method definition, it became clear to me that the current method would not propagate the value of the *forcing* special variable properly. The modification to the test-module-depend.{script,asd} verifies this bug.
-
Robert P. Goldman authored
-
Robert P. Goldman authored
-
Robert P. Goldman authored
-
Robert P. Goldman authored
I believe that this patch, using a dynamic variable, fixes the TRAVERSE method so that the components of a module will be operated on if any of the module's dependencies are activated in TRAVERSE. However, this proposed patch has ramifications. Because SYSTEM isa MODULE, this means that downstream systems will be compiled if upstream ones change, which ASDF has never done before.
-
Robert P. Goldman authored
My module-dependency test had strayed from the other tests.
-
Robert P. Goldman authored
This test checks to see if a file in a module will be recompiled if a file it depends on changes. In "classic" asdf this does not happen --- it's launchpad bug https://bugs.launchpad.net/asdf/+bug/502946 This is just a test for the condition. The test fails in the current state of ASDF.
-
Francois-Rene Rideau authored
Started as an attempt to "enhance" it, until thinking about why I was breaking tests told me I couldn't.
-
- Feb 10, 2010
-
-
Francois-Rene Rideau authored
or else upgrades from old SBCL asdf will cause an "interesting" inversion of priority between contrib-sysdef-search and sysdef-source-registry-search.
-
Francois-Rene Rideau authored
-
- Feb 08, 2010
-
-
Francois-Rene Rideau authored
-
- Feb 07, 2010
-
-
Robert P. Goldman authored
There is only one minor fix (substitution of string-equal for string= in a component name comparison to avoid problems with mlisp). The rest of this commit is all commenting and docstrings. Trying to get a handle on the bug about module dependencies.
- Feb 06, 2010
-
-
Francois-Rene Rideau authored
Simplify test scripts.
-
- Feb 05, 2010
-
-
Francois-Rene Rideau authored
-
Robert P. Goldman authored
The eval command line argument was not correct for Allegro CL.
-
Robert P. Goldman authored
There were bugs in the tests having to do with repeatedly calling quit-on-error in a single test script. Only one of those calls will be evaluated, and then the script will exit. That means that multiple calls to quit-on-error masked failures in the tests. Also fixed a package-related error in test1.script that I conjecture led to the mistaken use of multiple calls to quit-on-error.
-
Robert P. Goldman authored
-
Francois-Rene Rideau authored
Better test debugging support by Robert Goldman.
-