Some quick notes on status of 0.19 for release where I see things.
1. [r12550][1] should probably be backported if anything.
[1]: http://trac.common-lisp.net/armedbear/changeset/12550
2. [r12549][2] is probably mostly just cosmetic, but brings up questions of how well the Pathname use of :UNSPECIFIC should be.
[2]: http://trac.common-lisp.net/armedbear/changeset/12549
3. I'm still unhappy about the state of the Pathname support, but its a goal that should be postponed until 0.20, as the current implementation is no worse that what shipped in 0.18.1. Most notably, I think I really need to a) support URLs as Pathnames and b) allow all Pathnames to be arguments of OPEN which probably means subclassing Stream.java. And this is a whole lot of work, so postpone.
4. Disturbingly, I uncovered a [compiler stack inconsistency error][3] today in working on ASDF. I checked that this was *NOT* induced by the recent PROGV fix, and exists at least as far back at 0.18.1. So, it should not affect the decision to release unless we feel we should fix it before release.
[3]: http://trac.common-lisp.net/armedbear/ticket/89
So, if we're in a "lets get this thing out the door", just release what's on 0.19 (after updating the README).
If you want to backport [r12550], I think it should not affect any of the existing test results, but they should be re-run just to make sure. I probably won't have such time until tomorrow (Wednesday).
On Mar 16, 2010, at 5:13 PM, Mark Evenson wrote: […]
If you want to backport [r12550], I think it should not affect any of the existing test results, but they should be re-run just to make sure. I probably won't have such time until tomorrow (Wednesday).
[…]
I have now backported the changes necessary to make abcl-0.19 run ASDF2.
Then I re-ran the both the compiled and interpreted ANSI tests as well as the ABCL-TEST-LISP internal suite on three platforms: 'dada' OSX, 'jupiter' OpenSolaris, and 'xp1' WindowsXP all running a recent version of Java6. [The details are in the trunk version of the ANSI test result database][1], for which the reports (at the bottom) show that there are no new test failures except for a PRINT.BACKQUOTE.RANDOM.14 under XP which probably points to an undiagnosed problem with random backquoted expressions which unfortunately can't be duplicated.
Additionally, there seems to be a problem with ASDF-INSTALL under OpenSolaris having to do with not properly switching the HTTP download stream from character encoding to a binary stream, but I guess we let that slide for now.
I would recommend releasing 0.19.x as of r12555.
[1]: http://trac.common-lisp.net/armedbear/changeset/12556
ANSI> (report 'doit '0.18.1 '0.19.x) 0.18.1[ALQAEDA] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.18.1[ALQAEDA] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.18.1[ALQAEDA] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 MAKE-BROADCAST-STREAM.8 FRESH-LINE.5) 0.19.X[XP1] --> 0.18.1[ALQAEDA] additional failures: (PRINT.BACKQUOTE.RANDOM.14) 0.18.1[DADA] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 PRINT.BACKQUOTE.RANDOM.14 MAKE-BROADCAST-STREAM.8) 0.18.1[DADA] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 PRINT.BACKQUOTE.RANDOM.14 MAKE-BROADCAST-STREAM.8) 0.18.1[DADA] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 MAKE-BROADCAST-STREAM.8 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.19.X[DADA] --> 0.18.1[XP1] additional failures: (FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.19.X[JUPITER] --> 0.18.1[XP1] additional failures: (FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.19.X[XP1] --> 0.18.1[XP1] additional failures: (PRINT.BACKQUOTE.RANDOM.14) NIL ANSI> (report 'compileit '0.18.1 '0.19.x) 0.18.1[ALQAEDA] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.18.1[ALQAEDA] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.18.1[ALQAEDA] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 MAKE-BROADCAST-STREAM.8 FRESH-LINE.5) 0.18.1[DADA] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 PRINT.BACKQUOTE.RANDOM.14 MAKE-BROADCAST-STREAM.8) 0.18.1[DADA] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 PRINT.BACKQUOTE.RANDOM.14 MAKE-BROADCAST-STREAM.8) 0.18.1[DADA] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 PRINT.BACKQUOTE.RANDOM.14 MAKE-BROADCAST-STREAM.8 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[DADA] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.19.X[DADA] --> 0.18.1[XP1] additional failures: (FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[JUPITER] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8) 0.19.X[JUPITER] --> 0.18.1[XP1] additional failures: (FORMAT.JUSTIFY.32 FORMAT.JUSTIFY.30 FRESH-LINE.5) 0.18.1[XP1] --> 0.19.X[XP1] additional failures: (WITH-STANDARD-IO-SYNTAX.23 MAKE-BROADCAST-STREAM.8)
--
"A screaming comes across the sky. It has happened before, but there is nothing to compare to it now."
armedbear-devel@common-lisp.net
-
Mark Evenson