Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
On Sat, Oct 04, 2008 at 10:40:52AM +0200, Martin Zobel-Helas wrote: > Package: guile-1.6 > Version: 1.6.8-6.1 > Severity: serious > > There was an error while trying to autobuild your package: > > > Automatic build of guile-1.6_1.6.8-6.1 on lebrun by sbuild/sparc 98 > > Build started at 20081004-1001 > > [...] > > > ** Using build dependencies supplied by package: > > Build-Depends: libtool, debhelper (>= 4), libncurses5-dev, > > libreadline5-dev, quilt > > [...] > > > Running posix.test > > Running r4rs.test > > ERROR: Unbound variable: eot > > FAIL: check-guile FWIW the failing test seems to be test-cont in test-suite/tests/r4rs.test, so the bug might be in either leaf-eq? or next-leaf-generator filippo signature.asc Description: Digital signature
Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
Filippo Giunchedi (28/11/2008): > FWIW the failing test seems to be test-cont in test-suite/tests/r4rs.test, > so the bug might be in either leaf-eq? or next-leaf-generator Thanks for the hint, I'll try and have a look into it. Mraw, KiBi. signature.asc Description: Digital signature
Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
Cyril Brulebois (31/12/2008): > Filippo Giunchedi (28/11/2008): > > FWIW the failing test seems to be test-cont in test-suite/tests/r4rs.test, > > so the bug might be in either leaf-eq? or next-leaf-generator > > Thanks for the hint, > > I'll try and have a look into it. Thanks to Martin, got the B-D installed. With the attached diff, I'm getting: | Running r4rs.test | ERROR: Unbound variable: bar so leaf-eq? seems the buggy one. Still need to go deeper, but I thought I'd let the bug log know. I've also noticed there were other “bar” occurrences in the file, so I've replaced it with “fooish”, but that still breaks in the same way. And if there are other folks who want to play with it, once the build failed, the tests can be run using “make check-TESTS”. Happy everything. Mraw, KiBi. --- a/test-suite/tests/r4rs.test +++ b/test-suite/tests/r4rs.test @@ -874,11 +874,11 @@ ;;; r...@copper.ucs.indiana.edu (Raja Sooriamurthi) for fixing this ;;; code. The function leaf-eq? compares the leaves of 2 arbitrary ;;; trees constructed of conses. -(define (next-leaf-generator obj eot) +(define (next-leaf-generator obj baz) (letrec ((return #f) (cont (lambda (x) (recur obj) - (set! cont (lambda (x) (return eot))) + (set! cont (lambda (x) (return baz))) (cont #f))) (recur (lambda (obj) (if (pair? obj) @@ -890,12 +890,12 @@ (lambda () (call-with-current-continuation (lambda (ret) (set! return ret) (cont #f)) (define (leaf-eq? x y) - (let* ((eot (list 'eot)) - (xf (next-leaf-generator x eot)) - (yf (next-leaf-generator y eot))) + (let* ((bar (list 'bar)) + (xf (next-leaf-generator x bar)) + (yf (next-leaf-generator y bar))) (letrec ((loop (lambda (x y) (cond ((not (eq? x y)) #f) - ((eq? eot x) #t) + ((eq? bar x) #t) (else (loop (xf) (yf))) (loop (xf) (yf) (define (test-cont) signature.asc Description: Digital signature
Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
Cyril Brulebois (31/12/2008): > so leaf-eq? seems the buggy one. Still need to go deeper, but I > thought I'd let the bug log know. I've also noticed there were other > “bar” occurrences in the file, so I've replaced it with “fooish”, but > that still breaks in the same way. And if there are other folks who > want to play with it, once the build failed, the tests can be run > using “make check-TESTS”. Hrm. Maybe I'm wrong but that looks like it could be a miscompilation bug. Explanations: since upstream moved from CVS to git, I gave bisecting a try. I used “git merge-base” to find the branchpoint between the 1.6 and 1.8 branches, and try to build there. I had to tweak the environment a bit, but then stumbled upon FTBFS with gcc-4.3. Since I wasn't keen on patching every revision to be tried, I tried with gcc-4.2, and then I couldn't reproduce the test-suite failure anymore. I'm not sure how to debug that any further without having to fix every tried revision, which would mean a dozen or so. I think I'll be lacking either time or skills to fix every build failure, so I'll leave it up to the maintainer to see what to do from now. Mraw, KiBi. signature.asc Description: Digital signature
Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
On Thu, Jan 01, 2009 at 11:10:38PM +0100, Cyril Brulebois wrote: > Cyril Brulebois (31/12/2008): > > so leaf-eq? seems the buggy one. Still need to go deeper, but I > > thought I'd let the bug log know. I've also noticed there were other > > “bar” occurrences in the file, so I've replaced it with “fooish”, but > > that still breaks in the same way. And if there are other folks who > > want to play with it, once the build failed, the tests can be run > > using “make check-TESTS”. > > Hrm. Maybe I'm wrong but that looks like it could be a miscompilation > bug. Explanations: since upstream moved from CVS to git, I gave > bisecting a try. I used “git merge-base” to find the branchpoint between > the 1.6 and 1.8 branches, and try to build there. I had to tweak the > environment a bit, but then stumbled upon FTBFS with gcc-4.3. Since I > wasn't keen on patching every revision to be tried, I tried with > gcc-4.2, and then I couldn't reproduce the test-suite failure anymore. > > I'm not sure how to debug that any further without having to fix every > tried revision, which would mean a dozen or so. I think I'll be lacking > either time or skills to fix every build failure, so I'll leave it up to > the maintainer to see what to do from now. I just tried to build it with CFLAGS=-g -O0 and the test-suite didn't fail any more. -- Martin Zobel-Helas | Debian System Administrator Debian & GNU/Linux Developer | Debian Listmaster Public key http://zobel.ftbfs.de/5d64f870.asc - KeyID: 5D64 F870 GPG Fingerprint: 5DB3 1301 375A A50F 07E7 302F 493E FB8E 5D64 F870 -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#501114: guile-1.6_1.6.8-6.1(sparc/unstable): FTBFS on sparc in test cases
Package: guile-1.6 Version: 1.6.8-6.1 Severity: serious There was an error while trying to autobuild your package: > Automatic build of guile-1.6_1.6.8-6.1 on lebrun by sbuild/sparc 98 > Build started at 20081004-1001 [...] > ** Using build dependencies supplied by package: > Build-Depends: libtool, debhelper (>= 4), libncurses5-dev, libreadline5-dev, > quilt [...] > Running posix.test > Running r4rs.test > ERROR: Unbound variable: eot > FAIL: check-guile > === > 1 of 1 tests failed > === > make[3]: *** [check-TESTS] Error 1 > make[3]: Leaving directory `/build/buildd/guile-1.6-1.6.8' > make[2]: *** [check-am] Error 2 > make[2]: Leaving directory `/build/buildd/guile-1.6-1.6.8' > make[1]: *** [check-recursive] Error 1 > make[1]: Leaving directory `/build/buildd/guile-1.6-1.6.8' > make: *** [debian/stamp/build] Error 2 > dpkg-buildpackage: failure: debian/rules build gave error exit status 2 A full build log can be found at: http://buildd.debian.org/build.php?arch=sparc&pkg=guile-1.6&ver=1.6.8-6.1 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]