Daniel J. H.

Results 185 issues of Daniel J. H.

Hey folks, I ran into this issue when shipping an overview map displaying vector tiles with a single tile on zoom level zero. Meaning - I have a vector tile...

need more info

I just had a look at the Berlin GTFS feeds for this year. > https://daten.berlin.de/datensaetze/vbb-fahrplandaten-januar-2017-bis-dezember-2017 > stops.txt, CC-BY 3.0 licensed: http://www.vbb.de/de/datei/GTFS_VBB_Jan_Dez2017.zip I can see three concrete issues: 1/ In there...

Data
Feature Request

We are using `__builtin` calls in some of our bit manipulation functions. The idea is that the compiler will turn the builtin calls to native instructions, when compiled with the...

Please read https://github.com/tinygraph/tinygraph/issues/14 for context. This changeset adds a first take on succinct rank/select data structures based on http://www.cs.cmu.edu/~dga/papers/zhou-sea2013.pdf - named "poppy" > "Space-efficient, high-performance rank and select structures on...

We should provide an undirected graph abstraction. At the moment the csr graph is directed. For use what we have and turn it into an undirected graph, we'll need -...

We want to provide API and ABI stability - checking it can be done e.g. with the following tool - https://lvc.github.io/abi-compliance-checker/ - https://github.com/lvc/abi-compliance-checker Ideally we'd run this continuously e.g. in...

We are starting out with modern x86-64 hardware targets starting with Haswell+ for instructions like popcount, etc. - https://github.com/tinygraph/tinygraph/issues/11 In addition we should look into the neon instruction set, for...

We should look into implementing rank/select structures for bitset / elias-fano https://github.com/tinygraph/tinygraph/issues/13 For example have a look at > "Space-efficient, high-performance rank and select structures on uncompressed bit sequences" by...

One idea is run-length encoding offsets https://github.com/tinygraph/tinygraph/issues/12 - which will probably work quite well with very sparse graphs like road networks. Another idea is Elias-Fano coding, since offsets are monotonically...

At the moment we use calloc / malloc / free for dynamic memory allocation. We should allow for custom memory allocators. Look into how to best support custom allocators.