Am Mittwoch, 12. November 2003 04:27 schrieb [EMAIL PROTECTED]: > >> do/protect protected-user.r > >> > >> Does this makes sense to any of you? > >> > >> is there already a way to do so? > > > >do/protect protected-user.r > >-> > >secure[file [allow read]] > >do %protect-user.r > > I'm not sure the above does exactly what I mean... > > I mean that once loaded, anything that is contained in the file is > protected. I don't care if this code writes on disk, I just don't want any > later code, to modify the content of everything which was loaded from that > file (passwords, for example) or paths to output files, or ip addresses to > remote servers. Any attempt at tampering them would cause an error, like > normal protected code.
No chance to do that AFAIK. RT uses launch instead. The reblets run in a clean vm, so whatever password the user enters in the desktop is protected. > > >for ram-cache, i wrote a ramdisk-scheme long ago. which loaded missing > > files from disk and cached. eventually i find it, if you ask. but since > > it was very simple and there are some scheme-examples now, you may have > > more fun to do it yourself ;) > >you have to change files to ram://file, but since the disk is readonly you > > wil trap wrong file-access. > > The goal is that you use it without knowing that you are using a protected > environment. so that any disk oriented code runs as-is, but without risk > of damaging your HD. also a simple log function would have to accompany > the tool so that you can diagnostic if its working ok. > If its for testing, i would copy the dirtree to a subdir and 'secure rebol to only write there. > basically it could be a replacement for the normal disk port file:// But > which has a listing of all tampered files/dirs along with memory allocated > for every file you tried to edit. > > I think Java has something similar, I just don't recall how its named. > > > -MAx -Volker -- To unsubscribe from this list, just send an email to [EMAIL PROTECTED] with unsubscribe as the subject.
