- 27 Aug, 2016 2 commits
-
-
Francois-Rene Rideau authored
This notably avoids confusion on Windows, where `make` might run .\make.bat instead of looking for make.exe in the %PATH%. Also update usage documentation in Makefile-lisp-scripting after renaming.
-
Francois-Rene Rideau authored
Note that the latest release is 3.1.7. Also note that since CLISP adopted it in May 2016, all active Common Lisp implementations provide ASDF 3.1. Yay. Also import some information that used to be in the README.md but removed in 81452d5d. That README also had information on building and testing with the asdf-tools, and on the directory structure of the git repo.
-
- 19 Aug, 2016 1 commit
-
-
Robert Goldman authored
-
- 18 Aug, 2016 4 commits
-
-
Robert Goldman authored
Merge branch 'master' into run-program-bug-fixes.
-
Robert Goldman authored
-
Robert Goldman authored
Pull Makefile fix into the branch.
-
Robert Goldman authored
Under Cygwin, the CL_SOURCE_REGISTRY will be a set of *cygwin* pathnames. But when this gets passed to the Windows lisps (and through them to other programs), that caused failures. So now we check for cygwin, and if we find it, we use the Cygwin "cygpath" command to translate the CL_SOURCE_REGISTRY into Windows native paths. Possible problem: originally the CL_SOURCE_REGISTRY in the Makefile was terminated with ":" which at least sometimes was treated as an implicit ".". But then I found that this caused failures on some Windows tests, because that was a relative pathname, which the path translator didn't like.
-
- 16 Aug, 2016 4 commits
-
-
Robert Goldman authored
* master: Make test more verbose.
-
Robert Goldman authored
-
Robert Goldman authored
* master: Fix treatment of CCL:PROCESS-RESET. Cut the Markdown file to simply point to common-lisp.net project page.
-
Robert Goldman authored
This condition, incorrectly made a subclass of SERIOUS-CONDITION, was breaking some tests, because a normal termination using (EXIT-LISP 0) could cause this to be signaled.
-
- 15 Aug, 2016 2 commits
-
-
Robert Goldman authored
Needed for clean compile.
-
Robert Goldman authored
Previously, if the tests failed because of unexpected warnings, the WITH-ASDF-CONDITIONS macro would swallow the warning you needed to fix.
-
- 11 Aug, 2016 8 commits
-
-
Elias Pipping authored
If a process is killed by a signal, sys:reap-os-subprocess will return (0, pid, signal) so that checking the exit code alone does not suffice.
-
Elias Pipping authored
The code for Clasp, Clozure CL, and ECL is there, only the list of implementations in #-(...) was not up-to-date.
-
Elias Pipping authored
Use the public function ccl:external-process-id instead the private ccl::external-process-pid external-process-id has been exported from the ccl packages since at least 2008.
-
Elias Pipping authored
:wait t and :*put :stream cannot be mixed on some platforms but not others.
-
Elias Pipping authored
-
Elias Pipping authored
The system package (si for short) is internal and "never to be used" according to https://common-lisp.net/project/ecl/manual/ch11.html This change does not require testing because (presumably unintentionally) the process-info-pid function generates an error for ECL anyway (to be fixed in a follow-up commit)
-
Elias Pipping authored
Prior to this, %run-command would generate errors such as File #P"/dev/null" already exists.: File exists [errno=17]. Tested with acl-10.0-linux-x86.
-
Elias Pipping authored
-
- 03 Aug, 2016 1 commit
-
-
Robert Goldman authored
With apologies, I do not have time to maintain two different front pages for this project. Since my efforts have so far gone to the HTML page, I am making that the official portal and removing the content here. See ASDF-devel email for discussion of this decision.
-
- 01 Aug, 2016 2 commits
-
-
Robert Goldman authored
-
Robert Goldman authored
Previous method was over-complicated. Now we simply appropriately define the type UIOP:FATAL-CONDITION, and we can do without both UIOP:*FATAL-CONDITIONS* and UIOP::*FATAL-CONDITION-EXCEPTIONS*.
-
- 25 Jul, 2016 2 commits
-
-
Robert Goldman authored
-
Robert Goldman authored
See https://bugs.launchpad.net/asdf/+bug/1605650 According to the spec, SERIOUS-CONDITIONs are "All conditions serious enough to require interactive intervention if not handled should inherit from the type serious-condition." This is almost exactly the definition of *FATAL-CONDITIONS*, so I substituted SERIOUS-CONDITION for ERROR here. Added exceptions to the set of *FATAL-CONDITIONS* in order to fix test condition handling for CCL. The exceptions permit us to have a class of conditions (e.g., SERIOUS-CONDITION) that has a subclass that is not a FATAL-CONDITION, but that does not have an obvious more specific alternative. The specific example that caused me to add this: On shutdown, CCL signals CCL:PROCESS-RESET, a subclass of SERIOUS-CONDITION. So recent test harness modifications to handle SERIOUS-CONDITIONs caused the test scripts to mistakenly think there were failures because of calls to EXIT-LISP.
-
- 24 Jul, 2016 3 commits
-
-
Robert Goldman authored
-
Robert Goldman authored
-
Robert Goldman authored
There are some serious conditions (e.g., STORAGE-CONDITION) that are not ERRORs. Previously, our test scripts only caught ERRORs, meaning that if a lisp implementation encountered a non-ERROR SERIOUS-CONDITION in the course of running the tests, odd behavior could result. E.g., running in ECL we would drop into the debugger instead of exiting with a non-zero exit status.
-
- 27 Jun, 2016 3 commits
-
-
Robert Goldman authored
ECL version 16+ seems to have working encodings support. So we check the ECL version early and either abort the test (successfully) if ECL is too old, or push on if ECL is version 16+
-
Robert Goldman authored
Looks like this didn't get updated after we installed the systems in the ext/ subdirectory of the asdf-directory.
-
Robert Goldman authored
Mistakenly committed after refilling a comment and emacs refilled it wrong, so that comment was incorrectly not all prefixed by comment characters.
-
- 24 Jun, 2016 3 commits
-
-
Robert Goldman authored
-
Robert Goldman authored
Version numbers end with a number, so are: (digit-block .)+ digit-block Old regexp didn't have final digit-block.
-
Robert Goldman authored
-
- 11 Jun, 2016 2 commits
-
-
Francois-Rene Rideau authored
-
Francois-Rene Rideau authored
Bug found by @axionic on github: on very large directories, directory-files was slow, because it was using remove-duplicates with non-hashable test function #'pathname-equal. But this functionality was only really needed for the purpose of dealing with logical-pathnames, and having a few duplicates in case of wanton symlinks isn't much of an issue.
-
- 25 May, 2016 3 commits
-
-
Robert Goldman authored
Catch up with Fare's doc and submodule modifications.
-
Robert Goldman authored
Key functions used in the special-purpose COLLECT-SUB*DIRECTORIES call were anonymous lambda functions. That meant they were impossible to trace for debugging. Replaced them with FLET functions that can be traced.
-
Robert Goldman authored
In some cases a subdirectory pathname returned from this function was not a DIRECTORY-PATHNAME (i.e., leaf directory name was in the PATHNAME-NAME, not PATHNAME-DIRECTORY). I used ENSURE-DIRECTORY-PATHNAME to avoid this happening.
-