On Mon, May 6, 2013 at 3:42 PM, C. Michael Pilato <[email protected]> wrote: > On 05/06/2013 03:11 PM, Branko Čibej wrote: >> It's always worked that way for me, but then, my APR doesn't change that >> often. Sounds like a different APR is loaded at runtime than was >> compiled against; but I can't understand why libsvn_subr would use the >> apr_crypto API? > > See libsvn_subr/crypto.c. The logic is currently unused by trunk code. Was > introduced for the body of work currently living on the master-passphrase > branch. (gstein felt that the crypto bits should live on trunk even if the > rest of the work was only on the branch.)
RIght. Best exposure there. And we could use it when we feel the time is right. If we expect some problems with it, then an #if 0 guard could be added. Maybe decide on that when you figure out what happened. Cheers, -g
