On Wed, 2021-10-27 at 13:02 +0200, Mark Rogers wrote:
> Is option 3) worth investigating?
Hi,
I think it makes sense, at least from my point of view. I do not know
what others would think.
An alternative to your proposal is to detect the cycles before the
limit is reached. That would help, not only on the performance, but
also because the limit can break valid files. I recall a generated PDF
file with a structure, which triggered the limit, only because it had
not been structured efficiently. It had been shared on the list, if I
recall correctly, but I do not have pointers handy, I'm sorry. In any
case, it can be combined with your proposal (thread-local storage),
only not counting the deep level, but rather remember visited objects.
Bye,
zyx
_______________________________________________
Podofo-users mailing list
Podofo-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/podofo-users
