On Sat, Jun 04, 2011 at 06:17:22PM +0200, Stefan Sperling wrote: > On Sat, Jun 04, 2011 at 07:11:05PM +0300, Daniel Shahaf wrote: > > s...@apache.org wrote on Sat, Jun 04, 2011 at 12:32:28 -0000: > > > +/* Detect the mime-type of the file at LOCAL_ABSPATH using MAGIC_COOKIE. > > > + * If the mime-type is binary return the result in *MIMETYPE. > > > + * If the file is not a binary file or if its mime-type cannot be > > > determined > > > + * set *MIMETYPE to NULL. Allocate *MIMETYPE in RESULT_POOL. > > > + * Use SCRATCH_POOL for temporary allocations. */ > > > +svn_error_t * > > > +svn_magic__detect_binary_mimetype(const char **mimetype, > > > + const char *local_abspath, > > > + svn_magic__cookie_t *magic_cookie, > > > + apr_pool_t *result_pool, > > > + apr_pool_t *scratch_pool); > > > + > > > > As I said on IRC: I think this API is tailored for the caller and it > > would seem more natural to me to move the "Ignore the result if it's > > a binary type" to the caller. > > Feel free to change that if it really bugs you. > But do we ever care about auto-detecting non-binary mimetypes in svn?
If you still plan on doing this then please take into account the comment added in r1133367. We should keep flagging UTF-16 files as "application/octet-stream" (or some other binary type) for now. But libmagic usually reports them as "text/plain", and the current code happens to deal fine with that. (Maybe we should add a regression test for this but I don't have enough time right now...)
