Commit 0d9883a9 authored by Francois-Rene Rideau's avatar Francois-Rene Rideau Committed by Robert Goldman
Browse files

Update list of versions to test upgrade from

parent 40816022
......@@ -434,33 +434,36 @@ extract_all_tagged_asdf () {
}
valid_upgrade_test_p () {
case "${1}:${2}:${3}" in
# It's damn slow. Also, for some reason, we punt on anything earlier than 2.25,
# and only need to test it once, below for 2.24.
abcl:1.*|abcl:2.00[0-9]:*|abcl:201[0-9]:*|abcl:2.2[0-3]:*) : ;;
# ccl fasl numbering broke loading of old asdf 2.0
ccl:2.0[01]*|ccl:2.2[0-6]*) : ;;
# It's damn slow. Also, we punt on anything 2.26 or earlier.
abcl:1.*|abcl:2.[01]*|abcl:2.2[0-5]:*) : ;;
# Allegro ships with versions 3*, so give up testing 2
# Also, unpatched Allegro 10 has bug updating from 2.26 and before
allegro*:[12].*) : ;;
# My old ubuntu 10.04LTS clisp 2.44.1 came wired in
# ccl fasl numbering broke loading of old asdf 2.0. 2.27 has trouble with deferred-warnings.
ccl:2.[01]*|ccl:2.2[0-7]*) : ;;
# clasp only since 3.1.4.3
clasp:2.*|clasp:3.0*|clasp:3.1.[0-4]*) : ;;
# CLISP: My old ubuntu 10.04LTS clisp 2.44.1 came wired in
# with an antique ASDF 1.374 from CLC that can't be downgraded.
# More recent CLISPs work.
# 2.00[0-7] use UID, which fails on some old CLISPs.
# Note that for the longest time, CLISP has included 2.011 in its distribution.
# Now its hg repository includes 3.0.2.29, but clisp hasn't released in many years(!)
# Note that for the longest time, CLISP has included 2.011 in its source repository.
# Now its hg repository includes 3.x, but clisp hasn't released in many years(!)
# We don't punt on upgrade anymore, so we can go at it!
#clisp:2.00[0-7]:*|clisp:1.*|clisp:2.0[01]*|clisp:2.2[0-5]:*) : ;;
# CMUCL has problems with 2.32 and earlier because of
# CMUCL has problems with 3.2.1 and earlier because of
# the redefinition of system's superclass component.
cmucl:1.*|cmucl:2.[012]*|cmucl:2.3[012]*) : ;;
cmucl:1.*|cmucl:2.*|cmucl:3.[012]*) : ;;
# Skip many ECL tests, for various ASDF issues
ecl*:1.*|ecl*:2.0[01]*|ecl*:2.20:*) : ;;
# GCL 2.7.0 from late November 2013 is required, with ASDF 3.1.2
gcl:REQUIRE:*|gcl:1.*|gcl:2.*|gcl:3.0*) : ;;
gcl:REQUIRE:*|gcl:1.*|gcl:2.*|gcl:3.0*|gcl:3.1.[0-3]*) : ;;
# LispWorks is broken at ASDF 3.0.3, but can upgrade from earlier and later ASDFs.
lispworks:3.0.3:*) : ;;
lispworks:1*|lispworks:2.[0-2]*|lispworks:2.3[0-2]*|lispworks:3.0.3:*) : ;;
# MKCL is only supported starting with specific versions 2.24, 2.26.x, 3.0.3.0.x, so skip.
mkcl:[12]*|mkcl:3.0*) : ;;
# SBCL won't run ASDF 1 anymore
sbcl:1*) : ;;
# XCL support starts with ASDF 2.014.2
# — It also dies during upgrade trying to show the backtrace.
xcl:1.*|xcl:2.00*|xcl:2.01[0-4]:*|xcl:*) : ;;
......
......@@ -16,7 +16,10 @@
;; The 3.3 series provides the asdf3.3 feature, meaning users can rely on
;; all its new features (proper phase separation) as well as earlier features.
"3.3.1" ;; (2017-11-14) bug fixes, second and latest in 3.3 series
"3.3.4" ;; (2020-02-14) bug fixes, fifth and latest in 3.3 series
"3.3.3" ;; (2019-03-27) bug fixes, fourth in 3.3 series
"3.3.2" ;; (2018-05-04) bug fixes, third in 3.3 series
"3.3.1" ;; (2017-11-14) bug fixes, second in 3.3 series, SBCL has been stuck there for years
"3.3.0" ;; (2017-10-06) first in 3.3 series
;; The 3.2 series provides the asdf3.2 feature, meaning users can rely on
......@@ -133,9 +136,10 @@ Use at a given tag, put it under build/asdf-${tag}.lisp"
:allegro_64_s :allegromodern_64_s :allegro8_64_s :allegromodern8_64_s)
(version<= "2.27" tag))
;; CCL fasl numbering broke loading of old asdf 2.0, and the punting for 2.26 fails,
;; but who cares since CCL has always been shipping recent versions of ASDF.
((:ccl) (version<= "2.27" tag))
;; CCL fasl numbering broke loading of old asdf 2.0, the punting for 2.26 fails, and
;; 2.27 enables deferred-warnings then the new UIOP causes confusion with its changed format.
;; But who cares since CCL has always been shipping recent versions of ASDF.
((:ccl) (version<= "2.32" tag))
;; CLASP is only supported as of 3.1.4.3
((:clasp) (version<= "3.1.4.3" tag))
......@@ -149,9 +153,9 @@ Use at a given tag, put it under build/asdf-${tag}.lisp"
;; However, whether we punt or don't punt, these should all work.
((:clisp) t)
;; CMUCL has problems with 2.32 and earlier because of
;; the redefinition of system's superclass component.
((:cmucl) (version<= "2.33" tag))
;; CMUCL has problems with 3.2.1 and earlier because of the redefinition of
;; some classes that causes its subtype to answer NIL when it shouldn't.
((:cmucl) (version<= "3.3.0" tag))
;; Skip many ECL tests, for various ASDF issues
((:ecl :ecl_bytecodes) (version<= "2.21" tag))
......@@ -161,14 +165,20 @@ Use at a given tag, put it under build/asdf-${tag}.lisp"
;; against anything but the latest release.
((:gcl) (version<= "3.1.4" tag))
;; LispWorks is fine, but ASDF 3.0.3 has a bug and can't be loaded.
((:lispworks) (not (equal "3.0.3" tag)))
;; LispWorks is fine, but ASDF 3.0.3 can't be loaded due to a bug it has a bug,
;; while 2.27 to 2.32 use system:pid-exit-status that doesn't exist anymore in LispWorks 7.
((:lispworks) (and (version<= "2.33" tag) (not (equal "3.0.3" tag))))
;; MKCL is only supported starting with specific versions 2.24, 2.26.x, 3.0.3.0.x, so skip.
((:mkcl) (version<= "3.1.2" tag))
;; The bundle support also changed a lot in 3.2.0 and this breaks upgrading from 3.1.
((:mkcl) (version<= "3.2.0" tag))
;; all clear on these implementations
((:sbcl :scl) t)
;; SBCL fails to load ASDF as a system 1.369 because ASDF 1.369's builtin
;; asdf-binary-locations gets confused when SBCL_HOME is not set by SBCL anymore.
((:sbcl) (version<= "2.000" tag))
;; SCL is all good.
((:scl) (version<= "2.000" tag))
;; XCL support starts with ASDF 2.014.2
;; — It also dies during upgrade trying to show the backtrace.
......@@ -196,6 +206,7 @@ Use the preferred lisp implementation"
(extract-tagged-asdf tag)
(run-test-lisp description
`((load "test/script-support.lisp")
(asdf-test::debug-asdf)
(asdf-test::test-upgrade ,@method ,tag))
:lisp lisp :log log))
description))
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment