Dear Lisp hackers,
I am pleased to announce the release of ASDF 2.26. Since previous release 2.25, the changes are as follows:
* Run-program much improved, with a slight backward incompatibility. See the new documentation about it. * Portability enhanced, with more robust Windows support, plus patches for CCL, CLISP, LispWorks, XCL. * UIOP improved, with a README, more docstrings, a few new functions (println, writeln, stripln, ensure-gethash, vomit-output-stream), and bugfixes (to ensure-function, with-temporary-file, split-string). * Debian installations that don't export XDG_DATA_DIRS will benefit from the fix to split-string, as the bug was basically disabling the default source-registry entry for /usr/share/common-lisp/source// Closes: #723977 * Minor tweaks to the defsystem, mostly to print more information in some situations, less in other situations, but also to better handle systems with secondary names like foo/bar, a fix for user functions in output-translation, and to :version (:read-file-...) * Documentation somewhat updated.
This is a very substantial bug fix release, notably on Windows, but has a stable API wrt the previous version. For this reason, CL implementations are strongly encouraged to replace previous bundled versions with 3.0.3.
As always, thanks to Faré for what seems like a countless number of fixes. Thanks to Dave Cooper for many hours of testing on Windows, without whom the improvements on that platform would have been impossible.
Work will soon begin on the 3.1.0 series, with some additional API features, so please make sure to pull a copy of 3.0.3 instead of the bleeding edge when you are bundling.
Best, Robert
"Robert" == Robert Goldman rpgoldman@sift.net writes:
Robert> Dear Lisp hackers, Robert> I am pleased to announce the release of ASDF 2.26. Robert> Since previous release 2.25, the changes are as follows:
I assume you really meant 3.0.3 and not 2.26. :-)
I ran the tests and they all passed. However, I noticed there were some warnins such as the one appended below. I don't remember if these happened before or not.
Ray
Testing: test-utilities.script cmulisp -noinit -batch -eval "'(#.(load(string'|script-support.lisp|))#.(asdf-test::load-asdf)#.(asdf-test::frob-packages)#.(asdf-test::with-test()(load(string'|test-utilities.script|))))" ; Loading #P"/Volumes/share2/src/clnet/asdf/test/script-support.lisp". ; Loading #P"/Volumes/share2/src/clnet/asdf/build/fasls/cmucl/asdf.sse2f". Configuring ASDF Enabling debugging Being a bit verbose Comparing directories ASDF-TEST:*TEST-DIRECTORY* and ASDF-TEST::X both evaluate to same path: #P"/Volumes/share2/src/clnet/asdf/test/" Frob packages ; Loading #P"/Volumes/share2/src/clnet/asdf/test/test-utilities.script". Warning: These two expressions yield pathname-equal yet not equal path the first expression (NTH 0 (SPLIT-NATIVE-PATHNAMES-STRING (JOIN-NAMESTRINGS '("foo" "bar")))) yields this: #P"foo" (:HOST "" :DEVICE NIL :DIRECTORY NIL :NAME "foo" :TYPE :UNSPECIFIC :VERSION NIL)
the other expression (NTH 0 '(#P"foo" #P"bar")) yields that: #P"foo" (:HOST "" :DEVICE NIL :DIRECTORY NIL :NAME "foo" :TYPE NIL :VERSION NIL)
Raymond Toy wrote:
Robert> Dear Lisp hackers, Robert> I am pleased to announce the release of ASDF 2.26. Robert> Since previous release 2.25, the changes are as follows:I assume you really meant 3.0.3 and not 2.26. :-)
Yes, I saw that I had failed to edit that out about 2 minutes after I sent the email (and before I made the posting to c.l.l).
asdf-announce@common-lisp.net is dead, btw, so for the nonce you'll have to rely on asdf-devel and c.l.l for release announcements.
I see Faré has already answered about the warnings.
I hope you will like the new version of ASDF. It doesn't have a lot of new features, but what's there should be significantly more reliable.
Cheers, r
Warnings during tests are OK.
These particular warnings are relatively recent (definitely something from the asdf3 days), but are not damning. ASDF 3 has learned to work around a whole lot of weirdness in the way implementations handle pathnames (or fail to handle them, as it all too often happens).
—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org Your conscience never stops you from doing anything. It just stops you from enjoying it.
On Wed, Oct 23, 2013 at 10:53 PM, Raymond Toy toy.raymond@gmail.com wrote:
"Robert" == Robert Goldman rpgoldman@sift.net writes:
Robert> Dear Lisp hackers, Robert> I am pleased to announce the release of ASDF 2.26. Robert> Since previous release 2.25, the changes are as follows:I assume you really meant 3.0.3 and not 2.26. :-)
I ran the tests and they all passed. However, I noticed there were some warnins such as the one appended below. I don't remember if these happened before or not.
Ray
Testing: test-utilities.script cmulisp -noinit -batch -eval "'(#.(load(string'|script-support.lisp|))#.(asdf-test::load-asdf)#.(asdf-test::frob-packages)#.(asdf-test::with-test()(load(string'|test-utilities.script|))))" ; Loading #P"/Volumes/share2/src/clnet/asdf/test/script-support.lisp". ; Loading #P"/Volumes/share2/src/clnet/asdf/build/fasls/cmucl/asdf.sse2f". Configuring ASDF Enabling debugging Being a bit verbose Comparing directories ASDF-TEST:*TEST-DIRECTORY* and ASDF-TEST::X both evaluate to same path: #P"/Volumes/share2/src/clnet/asdf/test/" Frob packages ; Loading #P"/Volumes/share2/src/clnet/asdf/test/test-utilities.script". Warning: These two expressions yield pathname-equal yet not equal path the first expression (NTH 0 (SPLIT-NATIVE-PATHNAMES-STRING (JOIN-NAMESTRINGS '("foo" "bar")))) yields this: #P"foo" (:HOST "" :DEVICE NIL :DIRECTORY NIL :NAME "foo" :TYPE :UNSPECIFIC :VERSION NIL)
the other expression (NTH 0 '(#P"foo" #P"bar")) yields that:#P"foo" (:HOST "" :DEVICE NIL :DIRECTORY NIL :NAME "foo" :TYPE NIL :VERSION NIL)
-
Faré -
Raymond Toy -
Robert Goldman -
Robert P. Goldman