I actually like this idea a lot. My main concern is that it may not be immediately clear where a hash is being called for a user (for example, using it as an argument to a @cached_method will call __hash__). So when they try to mutate the vector/matrix and it fails, it may be difficult for them to find out where in the code they need to remove a hash.
An awkward "solution" would be to save the stack when you set it immutable so that you can print it when the user tries to modify the object. But I certainly don't want to add that amount of memory to each vector. Maybe a debugging mode for the parent could implement this, but discovery would be a problem.... David On Mon, Aug 9, 2021 at 10:43 PM Lorenz Panny <l.s.pa...@tue.nl> wrote: > > Have we considered the idea to simply call self.set_immutable() whenever > __hash__ is invoked? I think that would solve Nils' original problem and > it would mean things run smoothly for users who don't know about mutable > and immutable objects. In that case, the exception for trying to modify > immutable vectors should probably explain why the vector was frozen. > > > On Fri, 6 Aug 2021 01:25:08 +0200, Nils Bruin <nbr...@sfu.ca> wrote: > > Ideas? > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/sage-devel/20210810104118.6c1c6116%40l. > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/CAChs6_kX1nqV0fq6CJRLtcBrg4%2B95pQnG2h-y278OQP51O_OTg%40mail.gmail.com.
