Matt Johnston writes: > On Wed, Aug 31, 2005 at 01:01:50AM -0700, Eric Anderson wrote: > > > It would appear that the version that did &string[0] to get a > > writeable pointer to the string was the version that was committed. > > Was this intentional? Clemens ([EMAIL PROTECTED]) was > > concerned that doing this wasn't safe under the STL spec (although it > > works in practice because of how strings are implemented). > > > > I send a second version (Subject: [Monotone-devel] Incremental binary > > test, try 2) which instead used a char array[size], and changed the > > signature of guess_binary in file.hh. > > Taking another look now, I see the same attachment in both > messages. Could you resend the second diff?
I probably attached the wrong file. The attached patch should have the version with a char array[]; it should set the type of guess_binary to: bool guess_binary(const char *buf, int buflen); -Eric
incremental-binary-test.patch2
Description: Binary data
_______________________________________________ Monotone-devel mailing list Monotone-devel@nongnu.org http://lists.nongnu.org/mailman/listinfo/monotone-devel