Geoff Thorpe via RT schrieb: > Hi, Hello Geoff, [snip]
>>But in the code I checked the passed string is only referenced, >>not copied... > ahhh ... OK, right I'm with you now. So you see the problem is in the > ENGINE implementations, not in the library code that passes the strings > through to ENGINE_ctrl(). Yes, any ctrl() values that are passed in for > *later* use should be duplicated and stored inside the implentation. > This should be fixed. "There is no communication without errors..." >>Comparing the code of the different implementations it seems to me >>the first implementor made a mistake and all following >>implementors only copied the existing code (with small modifications) > > Heh, want to guess who that first implementor was? <gulp> Do I really want to know ? >>Is should be stored in a local copy of the string. >>Best would be to have the ENGINEs have a reverence to the >>internal data and store it there. > > Well an ENGINE implementation can attach its own data to an ENGINE > structure already, and some in fact do just this. Moreover, it's > intended for precisely this sort of "state", because you can also clean > up at the right time. >>But at the moment the passed data is stored in static variables. > Yup, that's not a Good Thing(tm). My point of view... >>??? >>I resubmited the patch on Tue, 18 Jun 2002 16:51:04 +0200... >> >>Should I send it again ? > > Nope, I've found it :-) Yes, your patch looks good - I'll double-check a > few things then commit. Please keep an eye on openssl-cvs and/or check > out the next snapshot. Thanks for your efforts. Simple: I need it myself... ;-) Bye Goetz -- Goetz Babin-Ebell, TC TrustCenter AG, http://www.trustcenter.de Sonninstr. 24-28, 20097 Hamburg, Germany Tel.: +49-(0)40 80 80 26 -0, Fax: +49-(0)40 80 80 26 -126
smime.p7s
Description: S/MIME Cryptographic Signature
