Summary: | dev-lisp/clisp-2.49-r8 FFI::FIND-FOREIGN-VARIABLE: foreign variable #<FOREIGN-VARIABLE "rl_readline_state" #x0000000000847B40> does not have the required size or alignment | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | tman <cornicx> |
Component: | Current packages | Assignee: | Common Lisp Bugs <common-lisp> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | gentoo, jstein, toralf, ua0zeh |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | /mnt/portage/logs/dev-lisp:clisp-2.49-r8:20160930-082530.log |
Description
tman
2016-09-30 08:33:19 UTC
Created attachment 448490 [details]
/mnt/portage/logs/dev-lisp:clisp-2.49-r8:20160930-082530.log
Getting the same error here. Getting the same error here. *** Bug 595700 has been marked as a duplicate of this bug. *** Getting the same error here. commit beb52002a3f12695af63071a74b52736eea5fe1c Author: Andrey Grozin <grozin@gentoo.org> Date: Sun Oct 2 10:38:44 2016 +0700 dev-lisp/clisp: 2.49-r9 for readline-7 Bug: 595624 Package-Manager: portage-2.3.0 commit 6a82015e0302846aaae7def7c4c1527bc2fd588c Author: Andrey Grozin <grozin@gentoo.org> Date: Sun Oct 2 09:53:11 2016 +0700 dev-lisp/clisp: in stable 2.49-r8, readline must be <7.0 Package-Manager: portage-2.3.0 (In reply to Andrey Grozin from comment #6) > dev-lisp/clisp: 2.49-r9 for readline-7 [..] > dev-lisp/clisp: in stable 2.49-r8, readline must be <7.0 The (upstream) patch should only be applied when readline-7 is installed, with readline < 7.0, it'll break again. Just as I've did in my patch in https://bugs.gentoo.org/show_bug.cgi?id=595624: # bug #594552 if use readline && has_version ">=sys-libs/readline-7.0"; then epatch "${FILESDIR}"/${P}-readline-7.patch fi See my comments there. *argh* I meant https://bugs.gentoo.org/show_bug.cgi?id=594552 of course. Oops. Commented first in <https://bugs.gentoo.org/show_bug.cgi?id=594552>. But the same applies. Why the f... have you, Andrey, not used the conditional application of the readline-7 patch??? And I'd like to reopen this bug. (In reply to David Haller from comment #9) > Oops. Commented first in <https://bugs.gentoo.org/show_bug.cgi?id=594552>. > But the same applies. Why the f... have you, Andrey, not used the conditional > application of the readline-7 patch??? > > And I'd like to reopen this bug. I have already explained this in my comment 19 in the bug #594552. Your "solution" with the patch applied conditionally is unacceptable. If you don't understand, I'll go in small steps. 1. clisp compiled with readline-7 (with the patch, of course) is binary incompatible with clisp compiled with <readline-7. Switching from one of them to the other one requires recompiling all dependencies: maxima[clisp], fricas[clisp], maybe others. 2. Hence clisp/readline-7 and clisp/<readline-7 *must* have different subslots. maxima and fricas ebuilds depend on clisp:= and will be recompiled. 3. Hence clisp/readline-7 and clisp/<readline-7 *must* be distinct revisions. 4. Hence there is no need to write this conditional logic in either of these 2 ebuilds: in one of them the condition is always true, in the other one it is always false. Anything else you don't understand? Please don't commit any changes to clisp which will break maxima and fricas. |