On Monday, September 2, 2024 at 11:13:43 AM UTC-6 dima wrote: It appears that all these gunclone_deep() etc aren't documented in any proper way (and this is a root cause of this; same applies to memleaks we have in libsingular interface). Can you post a link to docs you are reading? I can only find an email exchange from 2018 where similar issues are discussed. It seems that gunclone_deep should be used sparingly.
Just a small addendum to this thread. The recommended way to deal with freeing elements of a "lazy vector" GEN -- recommended in both the email exchange in the the PARI documentation -- is to pass the GEN to obj_free. But if you look at the code you find that what obj_free does is to first verify that the last element in the GEN is a t_VEC and, if so, pass it to gunclone_deep. So, while the libpari documentation (section 5.6.2) says that gunclone_deep is only useful in the context of the GP interpreter, in fact they use it in obj_free within libpari itself.. "User's Guide to the PARI library void gunclone_deep(GEN x) is only useful in the context of the GP interpreter which may replace arbitrary components of container types (t_VEC, t_COL, t_MAT, t_LIST) by clones. If x is such a container, the function recursively deletes all clones among the components of x, then unclones x. Useless in library mode: simply use gunclone." -- 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/eb10b1b6-dfee-4ab6-a879-f74e87572af3n%40googlegroups.com.
