[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

what to do with clisp/libsigsegv?



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hello,

After a bit of reworking I managed to redo the clisp package. This
package should fix quite a number of bugs, however it does require a new
libsigsegv and splits the package into multiple sub packages (as
requested by upstream).

I'll give the changelog entries, the work-in-progress can be seen on the
git repositories.

for clisp we have:

> clisp (1:2.49-1) UNRELEASED; urgency=low
> 
>   * Use libsigsegv 2.8 should fix a few issues. (Closes: 566686)
>   * conflict against libreadline6 as it would be illegal to link 
>     against that. (Closes: 553741)
>   * Use bash for clisp-link (Closes: #530054)
>   * New upstream version (Closes: #462742)3
>   * Redid debianization from scratch. 
>     (Closes: #504514,#177057,#462085, #488042, #462088, #433592, #433596)
>   * Dropped conflicts with ancient versions of clisp. 
>   * Changed build system on powerpc, the resulting image
>     passes all tests, so should Closes: #592768 
>   * We got a fix the 'file with wrong *pathname-encoding* in ~/' bug.
>     (Closes: #443520)
> 
>  -- Peter Van Eynde <pvaneynd@debian.org>  Wed, 01 Sep 2010 05:46:26 +0200

I'm currently working on fixing the sparc build and I have high hopes
that I'll be able to also make that one work.

for libsigsegv:

> libsigsegv (2.8-1) UNRELEASED; urgency=low
> 
>   * New upstream version 2.8 which we track with git. Closes: #541027
>     Major changes:
>    + Support for platforms that follow POSIX:2008, not POSIX:2001.
>    + On Linux platforms, libsigsegv now prefers the POSIX way of defining the
>      signal handler over than the traditional one, when both are supported.
>      As a consequence, on Linux/i386 and other Linux platforms, the type
>      'stackoverflow_context_t' is now typedefed to 'ucontext_t *' rather than
>      'struct sigcontext *'.
>    + sigsegv_leave_handler is changed. Previously it was a normal function with
>      no arguments. Now it is a function that take a non-returning continuation
>      function and three arguments for it as arguments.
>      Where you had code like
>         int my_handler(void* fault_address, int serious)
>         {
>           ...code_before()...;
>           sigsegv_leave_handler();
>           ...code_after()...;
>           longjmp(...);
>         }
>      you now have to write
>         void my_handler_tail(void* arg1, void* arg2, void* arg3)
>         {
>           ...code_after()...;
>           longjmp(...);
>         }
>         int my_handler(void* fault_address, int serious)
>         {
>           ...code_before()...;
>           #if LIBSIGSEGV_VERSION >= 0x0206
>           return sigsegv_leave_handler(my_handler_tail, arg, NULL, NULL);
>           #else
>           sigsegv_leave_handler();
>           my_handler_tail(arg, NULL, NULL);
>           /* NOTREACHED */
>           abort();
>           #endif
>         }
>    + Improved distinction between stack overflow and other fault on NetBSD,
>      OpenBSD, FreeBSD, Linux, AIX, Solaris. Contributed by Eric Blake.
>    + GNU gnulib now has an autoconf macro for locating libsigsegv:
>      http://www.gnu.org/software/gnulib/MODULES.html#module=libsigsegv
>   * Added debian/README.building file
>   * Now use debhelper v7
>   * Has updated autoconf files. (Closes: #533680)
>   * The SONAME became 2.0.1, so we need to rename the binary package
> 
>  -- Peter Van Eynde <pvaneynd@debian.org>  Sun, 08 Aug 2010 09:40:32 +0200

Both uploads should fix a lot of bugs, so I'm inclined to get this into
squeeze...

Comments?

Best regards, Peter
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=WXLV
-----END PGP SIGNATURE-----


Reply to: