On Fri, 2009-07-31 at 00:57 +0200, Ludovic Courtès wrote: Hello! > > "Michael Gran" <spk...@yahoo.com> writes: > > > The branch, master has been updated > > via 77332b21a01fac906ae4707426e00f01e62c0415 (commit) > > from e5dc27b86d0eaa470f92cdaa9f4ed2a961338c49 (commit) > > Oops, I hadn't realized this was in `master'. Was it intended? (I > don't remember seeing a discussion, but I may have skipped it.) >
It was discussed after a fashion. This is a precursor to the tree that Andy reviewed, and he seemed to be okay with me committing some of it.[1] With all the exciting stuff going on, I was having a little trouble getting some mindshare. [2] [3] So I pushed this one since it was kind of your idea anyway. [4] > > Replace global charnames variables with accessors > > > > The global variables scm_charnames and scm_charnums are replaced with > > the accessor functions scm_i_charname and scm_i_charname_to_num. > > Also, the incomplete and broken EBCDIC support is removed. > > Does it have a user-visible effect? No change at all in the character names it will accept. A minor change on the output: writing U+0012 now gives #\ff over #\np. Certainly the removal of EBCDIC would be a user-visible effect if there were an EBCDIC user. > > (If so, please update `NEWS' for 1.9.1->1.9.2.) > > > * libguile/print.c (iprin1): use new func scm_i_charname > > > > * libguile/read.c (scm_read_character): use new func > > scm_i_charname_to_num > > > > * libguile/chars.c (scm_i_charname): new function > > (scm_i_charname_to_char): new function > > (scm_charnames, scm_charnums): removed > > These removals are incompatible in theory, but probably they don't > warrant a `NEWS' entry. Thoughts? If they were API, they weren't well documented as such. They look like internal information to me. > > > +const char *const scm_r5rs_charnames[] = > > Please follow the GCS when it comes to spacing, indentation, etc. If in > doubt, run GNU Indent. OK. > > > +int scm_n_C0_control_charnames = sizeof (scm_C0_control_charnames) / sizeof (char *); > > Scary name! ;-) > > Shouldn't it be private? And shouldn't it be a macro instead? OK. > > Thanks, > Ludo'. Footnotes: [1] http://lists.gnu.org/archive/html/guile-devel/2009-06/msg00084.html [2] http://lists.gnu.org/archive/html/guile-devel/2009-07/msg00163.html [3] http://lists.gnu.org/archive/html/guile-devel/2009-07/msg00077.html [4] http://lists.gnu.org/archive/html/guile-devel/2009-02/msg00073.html
