Branko Čibej <brane@xbc.nu> writes:
> The docstring for check_non_ascii in utf.c says:
>
> /* Return APR_EINVAL if the first LEN bytes of DATA contain anything
> other than seven-bit, non-control (except for whitespace) ASCII
> characters, finding the error pool from POOL. Otherwise, return
> SVN_NO_ERROR. */
>
> And the code does:
>
> if ((! apr_isascii (*data))
> || ((! apr_isspace (*data))
> && apr_iscntrl (*data)))
>
> (Those should be svn_ctype_is*, but no matter).
>
> So it seems that, if iconv isn't used, we forbid all ASCII control
> characters in any non-fuzzily converted string (which includes file
> names). But we allow horizontal tabs, vertical tabs and linefeeds.
Yes. These were deliberate decisions, and much-discussed at the time.
Unfortunately, I don't remember when that time was or what the names
of any of those threads were.
I think it had something to do with what characters to allow in paths,
though why we would embed that logic deep in the raw string conversion
code is a mystery to me.
> And we allow ASCII control chars in filenames if we _do_ have iconv,
> even though the conversion to/from UTF-8 is a no-op. How weird. I
> can't help feeling that this is a ... misfeature, to say the least.
I think it may just be a misimplemented feature, but yes.
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Sep 17 01:19:55 2005