Re: [Zeek-Dev] Do we still need pysubnettree?
On Mon, Oct 14, 2019 at 15:24 +, Vlad Grigorescu wrote: > Does it still make sense to maintain pysubnettree? No strong opinion either way from my side. It looks like pytricia does indeed offer a very similar interface, and being able to stop maintaining a custom dependency would certainly be a plus. On the other hand, this might be a case of "if it's not broken, don't fix it". pysubnettree hasn't required a lot of work recently, and users would need to install a new dependency if we switched. I don't know what constraints LGPL imposes when applied to Python modules. Robin -- Robin Sommer * Corelight, Inc. * ro...@corelight.com * www.corelight.com ___ zeek-dev mailing list zeek-dev@zeek.org http://mailman.icsi.berkeley.edu/mailman/listinfo/zeek-dev
[Zeek-Dev] Do we still need pysubnettree?
>From what I can tell, trace-summary and zeekctl are the only things that use pysubnettree. pytricia seems to have become the de-facto module that's used for these structures in Python: https://github.com/jsommers/pytricia In fact, pytricia has a comparison section where it claim that it's faster (albeit only slightly) than pysubnettree. Does it still make sense to maintain pysubnettree? pytricia's interface looks very similar. A quick glance at how we're using pysubnettree makes me think that pytricia could just be a drop-in replacement. Are there build/packaging considerations? It looks like pytricia is LGPL licensed. On the flip side, I don't see many recent updates on pytricia. Although, it's straightforward enough, perhaps it doesn't need updates? Curious to hear thoughts. --Vlad ___ zeek-dev mailing list zeek-dev@zeek.org http://mailman.icsi.berkeley.edu/mailman/listinfo/zeek-dev
