Re: Bad control characters in GHCi 6.6 (Mac OS X PPC)

On Oct 23, 2006, at 8:59 AM, Björn Buckwalter wrote:

On 10/19/06, Simon Marlow wrote:

...

See this page that I just created:

http://hackage.haskell.org/trac/ghc/wiki/Building/MacOSX

I tried these instructions, first by using my existing (darwinports) readline in /opt/local, and when that failed by first installing readline in /usr/local per the instructions. Neither fixed the problem. I don't know why but perhaps I failed to remove the traces of my initial install completely. In the end I gave up and installed Greg Heartsfield's binary package which remedied my problems (thanks Greg!).

I think I have this partially figured out. When I install version 5.1 of readline, rl_readline_version is defined as follows in :

/* **************************************************************** */ /* */ /* Well Published Variables */ /* */ /* **************************************************************** */

/* The version of this incarnation of the readline library. */ extern const char *rl_library_version; /* e.g., "4.2" */ extern int rl_readline_version; /* e.g., 0x0402 */

However, in ghc-6.6, libraries/readline/configure has the following:

cat >>conftest.$ac_ext <<_ACEOF /* end confdefs.h. */

/* Override any gcc2 internal prototype to avoid an error. */ #ifdef __cplusplus extern "C" #endif /* We use char because int might match the return type of a gcc2 builtin and then its argument prototype would still apply. */ char rl_readline_version (); int main () { rl_readline_version (); ; return 0; } _ACEOF

which is looking for rl_readline_version as a *function*. This fails (at the link stage on Darwin/Intel), even though GNU readline is installed, because the symbol type is wrong. The configure script concludes that libedit is being used, even though it's using GNU Readline v5.1. Was rl_readline_version a function in earlier versions of GNU Readline? Deborah