From: Sam S. <sd...@gn...> - 2010-01-27 19:20:22
|
Yaroslav Kavenchuk wrote: > Sam Steingold wrote: >> in that case you can try comparing the contents of the conf* dir >> generated by "./config.status --debug" of CVS head and release. > > http://sites.google.com/site/kavenchuk/diffs.tgz 1. one of the config.h files has CR-LF line terminators, the other - LF. the diff is useless. 2. looks like config.status were called with different arguments. more precisely, they were generated by configure scripts called with different arguments. >>> work of config.status depends on what else? >> nothing, I think... >> > > no - config.h.in so, what's the diff? > when I copy old config.h.in in CVS environment and run config.status: > > $ grep -i readline config.h > /* Define to 1 if you have the declaration of `rl_gnu_readline_p', and > to 0 if > #define HAVE_DECL_RL_GNU_READLINE_P 1 > /* Define to 1 if you have the declaration of `rl_readline_name', and to > 0 if > #define HAVE_DECL_RL_READLINE_NAME 1 > /* have a working modern GNU readline */ > #define HAVE_READLINE /**/ > /* Define to 1 if you have the <readline/readline.h> header file. */ > #define HAVE_READLINE_READLINE_H 1 > #define READLINE_CONST const > /* The readline built-in filename completion function, either > #define READLINE_FILE_COMPLETE rl_filename_completion_function > > What next? > try to figure out why the old config.h.in works and the new one does not. |