[PATCH] Marking weak alist vectors

Hi, The patch below fixes a bug in the garbage collection process of weak alist vectors. Without the patch, a value attached to a weak key (or the opposite) may be freed just _before_ the weak key in question is freed. The problem stems from the fact that `scm_i_mark_weak_vector_non_weaks ()' wo

Re: [PATCH] Marking weak alist vectors

[EMAIL PROTECTED] (Ludovic Courtès) writes: > The problem stems from the fact that > `scm_i_mark_weak_vector_non_weaks ()' would not mark a value > attached to a weak key if that weak key is unmarked. Therefore, > both the weak key and its associated value become unmarked during > the very same m