diegoripley/tippecanoe
1
0
Fork 0
mirror of https://github.com/felt/tippecanoe.git synced 2026-10-05 18:05:42 +02:00
Code Issues Packages Projects Releases Wiki Activity

Shard the child tiles as widely as possible even if not skipping levels

Browse Source 
to help future parallelization
This commit is contained in:
Eric Fischer
2015-07-01 12:04:45 -07:00
parent b2fdcba6b0
commit af13a95dc1
1 changed files with 17 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tile.cc
+17 -1
View File
@@ -365,7 +365,23 @@ void rewrite(drawvec &geom, int z, int nextzoom, int file_maxzoom, long long *bb
for (xo = 0; xo < span; xo++) { for (xo = 0; xo < span; xo++) {
for (yo = 0; yo < span; yo++) { for (yo = 0; yo < span; yo++) {
int j = xo * span + yo; unsigned jx = tx * span + xo;
unsigned jy = ty * span + yo;
// j is the shard that the child tile's data is being written to.
//
// Be careful: We can't jump more zoom levels than MAX_ZOOM_INCREMENT
// because that could break the constraint that each of the children
// of the current tile must have its own shard, because the data for
// the child tile must be contiguous within the shard.
//
// But it's OK to spread children across all the shards, not just
// the four that would normally result from splitting one tile,
// because it will go through all the shards when it does the
// next zoom.
int j = ((jx & ((1 << MAX_ZOOM_INCREMENT) - 1)) << MAX_ZOOM_INCREMENT) |
((jy & ((1 << MAX_ZOOM_INCREMENT) - 1)));
long long bbox2[4]; long long bbox2[4];
int k; int k;