* Rename to jsonpull.cpp
* Clear for merge
* Convert jsonpull to C++ with shared_ptr and std::vector/std::string
Replace the manual malloc/realloc/free memory management in jsonpull
with std::shared_ptr ownership. Each json_object now owns its children
through std::vector<json_object_ptr>; raw back-pointers to parent and
parser remain valid by structural invariant and are cleared on
json_disconnect so detached subtrees can outlive their parser.
Strings become std::string, child arrays become std::vector, and the
old union becomes a struct so non-trivial members can coexist while
preserving the existing o->value.xxx access paths.
The old jsonpull.c is replaced by jsonpull.cpp, json_stringify now
returns std::string, and all callers across tippecanoe, tile-join,
tippecanoe-decode, tippecanoe-json-tool, tippecanoe-overzoom and the
unit tests are updated to use json_object_ptr / json_pull_ptr.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Subclass json_object so primitives shrink from 168 to 24 bytes
The previous "every member in a struct" layout cost 168 bytes per
json_object, even for JSON_NULL / JSON_TRUE / JSON_FALSE nodes that
have no payload. Splitting json_object into a small base class plus
json_number / json_string / json_array / json_hash subclasses brings
each instance down to just the size of its actual contents:
json_object (base, TRUE / FALSE / NULL) 24 bytes
json_number 48 bytes
json_string 48 bytes
json_array (empty) 48 bytes
json_hash (empty) 72 bytes
Other size wins along the way:
* Drop enable_shared_from_this<json_object> (its embedded weak_ptr
was 16 bytes per node). json_pull now keeps an explicit
container_stack and the parser no longer needs to resurrect a
shared_ptr from a raw `parent` walk.
* Remove the unused `refcon` slot from the string variant.
* No virtual destructor: shared_ptr keeps the deleter from the
original std::make_shared<json_xxx> call, so destroying a
shared_ptr<json_object> still runs the right subclass dtor.
The base class exposes type-tagged accessors (o->string(),
o->number(), o->array(), o->keys(), o->values(), o->large_signed(),
o->large_unsigned()) that assert the type matches and downcast to
the appropriate subclass storage. All call sites were swept from
the old `o->value.X.Y` field paths to these accessors. A raw-pointer
overload of json_hash_get() replaces the few external uses of
shared_from_this() that survived in geojson-loop.cpp.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Store hash key/value pairs in one ordered vector
Replace the parallel std::vector<json_object_ptr> keys / values on
json_hash with a single std::vector<json_entry>, where json_entry is
a small {key, value} aggregate. This still preserves insertion order
(the property the parallel vectors were providing) but removes the
"keep two vectors in lockstep" pattern, and call sites can now use
range-for with structured bindings:
for (auto &[k, v] : o->entries()) { ... }
Side effects:
* sizeof(json_hash) drops from 72 to 48 bytes (one fewer vector
header), matching json_array.
* The keys() and values() accessors on json_object are replaced by a
single entries() accessor returning std::vector<json_entry>&.
* All call sites were swept from the old paired-index pattern
(`o->keys()[i]` / `o->values()[i]`) to entry-based access. Where the
original pattern relied on `nprop = 0` to short-circuit iteration on
a null or non-hash `properties`, the rewrite now guards the loop
explicitly with `if (o->type == JSON_HASH)` so that calling
entries() doesn't trip the asserting downcast.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Move parser-only `expect` state out of json_object
`expect` was only meaningful while the parser was building a container,
and only ever read or written from jsonpull.cpp itself; once parsing
finished it was dead weight on every JSON_ARRAY and JSON_HASH (and
present-but-unused on every primitive too). Move it into the parser's
container stack, alongside the shared_ptr to the container it pertains
to:
struct json_pull::parse_frame {
json_object_ptr container;
json_type expect;
};
std::vector<parse_frame> container_stack;
The base class now only carries data-model state (parent, parser, type).
No external caller depended on `expect`, so no sweep was needed outside
jsonpull.cpp.
This change does not, in itself, shrink any json_object: the 4-byte
`expect` field used to live at offset 20 inside the base, where it was
already being eaten by alignment padding for the 8-byte-aligned first
member of every subclass (std::string, std::vector, double). The win is
in the data model, not the byte count -- the 4-byte hole is still
there, but it is now available for a future subclass whose first member
is small enough to slot into it.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Discriminate json_number's three numeric slots into one union
json_number used to carry three parallel 8-byte fields (a double plus
both a 64-bit unsigned and a 64-bit signed slot for the large-integer
cases) even though at most one of the integer slots is ever the
canonical value for any given number. Collapse them into a
discriminated union:
enum repr_t { REPR_DOUBLE, REPR_LARGE_UNSIGNED, REPR_LARGE_SIGNED };
repr_t repr;
union { double d; unsigned long long u; long long s; } value;
Callers keep the same read API: number() returns the appropriate
double, large_unsigned() returns the ull (or 0 if not currently stored
that way), large_signed() likewise. Writes go through new set_number /
set_large_unsigned / set_large_signed methods that keep the
discriminator and the union value in sync.
This was prompted by an observation that moving json_type to the end
of the object should shrink things via tail-padding reuse. Empirically
the type-at-end rearrangement saves nothing on its own (every
subclass payload is 8-byte aligned so it can't slot into the 4-byte
tail), but the discriminated-number redesign hits the same idea from
a different direction: adding the 4-byte `repr` to json_number makes
the class non-standard-layout, which lets the Itanium ABI pack `repr`
into the base's 4-byte tail padding at offset 20. The union value
then starts at the natural offset 24, and json_number ends at offset
32 -- a 33% reduction.
Per-node sizes:
json_object (TRUE/FALSE/NULL) 24 bytes
json_number 32 bytes (was 48)
json_string 48 bytes
json_array 48 bytes
json_hash 48 bytes
Numbers dominate real GeoJSON (every coordinate is one), so the net
memory win on a typical parse is substantial.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Fix bugs flagged in code review of jsonpull C++ port
- jsontool.cpp `out()`: route JSON_NUMBER (and anything else non-string)
through `json_stringify` instead of `o->string()`, which now asserts
on a non-string type and would crash `--extract` on numeric attributes.
- geojson.{hpp,cpp} `json_end_map`: take `json_pull_ptr` by reference so
the caller's shared_ptr is released, null-guard before touching
`jp->source`, and clear `jp->source` after delete to avoid a dangling
pointer.
- jsonpull/jsonpull.cpp: low-surrogate range check was comparing the
outer-loop byte `c` instead of the parsed code unit `ch`, breaking
surrogate-pair decoding for some \\uXXXX escapes. Pre-existing bug
preserved across the port.
- tile-join.cpp `handle_vector_layers`: require the field value to have
type JSON_STRING (and the key to be non-null) before calling
`string()`; the previous truthy `type` check would assert on a
non-string value.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Add jsonpull regression test for surrogate-pair decoding
Covers the `c` vs `ch` bug fixed in the previous commit: parsing
"\uD83D\uE000" (a valid high surrogate followed by a non-surrogate
BMP code point) used to mis-classify U+E000 as a low surrogate and
combine the two units into U+1F400 (F0 9F 90 80). The fixed code
flushes the stale high surrogate as standalone CESU-8 (ED A0 BD)
and then encodes U+E000 normally as EE 80 80. Verified the test
fails under the pre-fix logic.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Cheap perf wins in jsonpull C++ port
Profiling tl_2022_us_county.json (sample(1) on Apple Silicon) showed
~38% of parse time in allocator work and ~14% in std::string::push_back
during string-token construction. These changes target the low-hanging
fruit from that profile:
- Pre-reserve 2 slots in json_array and 4 slots in json_hash so
coordinate `[x, y]` pairs and typical GeoJSON property maps avoid
the 0 -> 1 -> 2 -> 4 vector-growth chain (and the shared_ptr copies
it incurs).
- Reuse a parser-wide std::string buffer for JSON_STRING tokens
instead of constructing a fresh local std::string per token. The
buffer is cleared (capacity preserved) at the start of each token
and copied into the final json_string, so once it has grown to the
longest string seen it stops reallocating entirely.
- std::move the freshly-created container shared_ptr into the parser
container stack in the `[` and `{` handlers, and move it out of the
frame on the matching `]` / `}`. Each move skips one atomic
inc/dec round-trip per container open and close.
On a tl_2022_us_county.json benchmark (4-iter user-time mean, Apple
Silicon, /usr/bin/time):
- main baseline: ~8.17s
- jsonpull-cpp before these changes: ~10.90s (+33%)
- jsonpull-cpp with these changes: ~9.33s (+14%)
So this commit recovers roughly half of the post-port regression.
The remaining gap is dominated by shared_ptr atomic refcount traffic
on the parse tree and per-node heap allocations, which would require
the larger unique_ptr/arena reworks to address.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Make json_free actually free the subtree
In the C++ port, json_free was just `o.reset()`, which dropped the
caller's reference but left the subtree alive: the parent's vector
slot kept it allocated, and for line-delimited streams the parser's
jp->root co-owned it until the next top-level value started parsing.
That defeated the geojson-loop pattern of calling json_free on each
feature after serializing it, which is supposed to release the
feature so it doesn't sit in memory while subsequent ones are parsed.
Restore the historical "remove this from the tree" semantics by
splicing the node out of its parent (sharing splice_from_parent with
json_disconnect) and clearing jp->root when the node is the parser's
current top-level value, then dropping the caller's reference.
Two unit tests pin this down: a pruning test parses
"[[1, 2], [3, 4], [5, 6]]" element-wise and confirms that calling
json_free on [3, 4] leaves the outer array with just [1, 2] and
[5, 6]; a top-level test uses a weak_ptr observer to confirm that
json_free on the parser's root really destroys the tree.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Migrate jsonpull to unique_ptr ownership
Replaces the shared_ptr-based json_object_ptr with a unique_ptr that
has a stateless custom deleter dispatching on json_object::type before
calling the right subclass destructor. Eliminates per-node atomic
reference-counting and the control-block allocation that shared_ptr
required for every node in the tree.
API now distinguishes owning and borrowing pointers explicitly:
- json_read / json_read_separators / json_hash_get return raw
json_object * (borrowed from the parser-owned tree).
- json_read_tree / json_disconnect return json_object_ptr (caller
takes ownership; back-pointers are cleared so the subtree can
outlive the parser).
- json_free / json_context / json_stringify take raw pointers.
- The parser's container_stack holds raw pointers; jp->root keeps
unique_ptr ownership of the most recent top-level value.
Internally, take_from_owner moves the unique_ptr out of whichever
parent vector / hash entry / parser root owned it, which both
json_free and json_disconnect rely on.
In the streaming parsers (parse_feature, parse_layers, the
geojson-loop callback), we are careful to free `j` only after we
have processed a complete Feature: json_read returns each token
as the tree is being built up, and freeing an intermediate node
would splice it out of the surrounding hash and corrupt the
in-progress feature.
Benchmark (tl_2022_us_county.json, -z0 --extend-zooms-if-still-dropping,
median of 5 runs on macOS arm64): 8.5s, vs 10.6s with shared_ptr
and 8.7s on the pre-refactor C baseline.
Co-authored-by: Cursor <cursoragent@cursor.com>
* Fix preprocessor mistakes identified by Copilot
* Make indent
* Skip non-string metadata.json entries instead of reading them as strings
dirmeta2tmp() warned about a metadata entry that was not a string/string
pair and then read it as a string anyway. Under the new type-tagged
accessors that trips the assert in json_object::string(); before them it
reinterpreted the node's storage as a char pointer, which segfaulted for
most values. Either way, tippecanoe-decode and tile-join could not read a
directory tileset whose metadata.json had a numeric minzoom or a nested
object, which is common in metadata.json files written by other tools.
Add the missing continue, and cover it in raw-tiles-test.
pmtilesmeta2tmp() handles the same case correctly but read the key with
string() before its own JSON_STRING check, so the assert would have fired
ahead of the check meant to catch a bad key. Hoist the check above the
read. The parser rejects non-string hash keys, so this is unreachable in
practice; the ordering is what makes the check meaningful.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Don't redefine _GNU_SOURCE in the C++ jsonpull port
The `#define _GNU_SOURCE` carried over from jsonpull.c, where it was
needed to get asprintf() declared. g++ already defines _GNU_SOURCE on the
command line for C++ translation units, so redefining it warns:
jsonpull/jsonpull.cpp:1: warning: "_GNU_SOURCE" redefined
Guard the define rather than drop it, so platforms whose C++ driver does
not predefine it still get asprintf() declared.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Add a unit test for json_disconnect
json_disconnect() is documented in jsonpull.h as the supported way to
splice a subtree out of the parser's tree and take ownership of it, but
nothing calls it: read_filter() and parse_filter() used to, and now get
the same guarantee from json_read_tree() clearing back-pointers on the way
out. Cover the behavior rather than leave the primitive dead and untested.
The test pins that the subtree is removed from its parent, that the parser
keeps the rest of the tree, and that the detached subtree stays readable
after the json_pull is destroyed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Correct two stale comments in the jsonpull port
jsonpull.h said a json_number is 40 bytes; it is 32 (json_object is 24,
and the repr discriminator fits in the base class's tail padding, so the
8-byte union lands at offset 24).
plugin.cpp's parse_feature() said `j` is freed only just before returning
or as jp->root at end of stream, but there is a third json_free(j) at the
bottom of the loop, for a complete Feature whose geometry came out empty.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Add a changelog entry and bump the version for the jsonpull rewrite
The rewrite is meant to be behavior-preserving, but it carries four
user-visible bug fixes that warrant release notes: tippecanoe-json-tool
--extract on a numeric attribute, surrogate-pair decoding, tile-join
reading a non-string tilejson field type, and non-string values in a
directory tileset's metadata.json.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Encode U+FFFF as three bytes instead of an overlong four
The \uXXXX decoder tested `ch < 0xFFFF` before taking the three-byte UTF-8
path, so U+FFFF itself fell through to the four-byte branch and came out as
F0 8F BF BF -- an overlong, and therefore invalid, encoding of a code point
that fits in three bytes.
check_utf8() only checks that continuation bytes look like continuation
bytes, not that a sequence is the shortest form, so nothing downstream
noticed: a GeoJSON attribute containing U+FFFF put invalid UTF-8 into the
output tile, where a strict consumer would reject it.
Since `ch` is parsed from exactly four hex digits it cannot exceed 0xFFFF
on its own, so after this change the four-byte branch is reached only for a
code point assembled from a surrogate pair, which is the only way to name
one above the BMP.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Keep the parent links inside a detached jsonpull subtree
json_read_tree() and json_disconnect() cleared both back-pointers on every
node of the subtree they handed out. Clearing `parser` throughout is
necessary -- the json_pull can be destroyed while the subtree lives on, so
a surviving `parser` would dangle -- but clearing `parent` throughout cost
more than it bought.
`parent` is a non-owning raw pointer, so keeping it cannot form a reference
cycle or keep anything alive; there is nothing to leak. And within a
detached subtree it refers to nodes the caller now owns as a single unit,
so it stays valid for exactly as long as the subtree itself. Clearing it
only made the tree unwalkable upwards, and made json_free() and
json_disconnect() silently no-ops on interior nodes of a detached tree,
since both find a node's owner through o->parent.
So clear `parser` everywhere and clear `parent` on the detached root alone,
which is the one that pointed out of the subtree at a node the parser still
owns. Split the old clear_back_pointers() into clear_parser_pointers() plus
a detach_subtree() wrapper that adds the root's `parent`.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Cover the U+FFFF encoding and detached-tree parent links
Each of the new assertions fails against the previous behavior, so they
pin the two fixes rather than merely passing alongside them:
- the U+FFFF test, plus the U+FFFE boundary below it and a surrogate pair
above it, so the three-byte and four-byte paths are both held in place
- json_disconnect() leaving the parent links inside the subtree intact
while clearing the root's
- json_free() pruning an interior node of a tree whose parser is already
gone, which only works because those links survive
- json_free() of a hash value leaving the key paired with a JSON_NULL
placeholder, which is the documented behavior and not a removal
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Address review notes in jsonpull itself
- json_stringify walked c_str(), so it truncated at an embedded NUL even
though values are std::string now and carry one through faithfully.
Range over the string instead; the existing control-character branch
already escapes a NUL like any other, so the output stays valid JSON.
- Assert that the hash has an entry waiting before add_object() assigns to
entries().back(). It always does -- JSON_VALUE is only set by a colon,
which requires a pushed key -- but the derivation is not local.
- Drop fabricate_object(), a pass-through to make_object() with the
arguments reordered, kept only to preserve the old C name.
- Inline the string_append / string_append_c wrappers over push_back and
append, and note that json_print_one's JSON_HASH and JSON_ARRAY branches
are unreachable, since json_print handles both itself.
- json_hash_get's comment said nullptr meant "the matching value is null",
which reads as JSON null. A JSON null comes back as a JSON_NULL node;
nullptr means the value slot is not filled in yet.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Tidy jsonpull call sites flagged in review
- geojson.cpp and attribute.cpp passed key_pool::pool() and
set_attribute_accum() a c_str() from a std::string, forcing a needless
reconstruction (and truncating at an embedded NUL). Both overloads take
std::string, so pass it directly. The geojson.cpp one is the hottest
loop in the program.
- Replace the hand-maintained counters beside range-for loops in
attribute.cpp, main.cpp and tile-join.cpp with indexed loops, since the
index is only wanted for error messages.
- parse_json_args took json_pull_ptr by value and then copied it, costing
two refcount bumps per construction. Move it.
- Assert that the parser is still attached where geojson.cpp reads
geometry->parser->line. Only json_read results reach it today, but
json_read_tree and json_disconnect now clear every parser pointer, so a
detached tree would null-deref there instead of tripping an assert.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Pin the array-splicing fix, and close four test gaps
The existing pruning test does not discriminate: json_read hands back each
container as it completes, so the node it frees is always the most recently
added element of its parent -- the one case the old element-count-vs-byte-
count memmove got right, because it then moved zero bytes. Widening that
test to more elements does not change this; the shape is what matters, not
the size. Verified: the eight-element streaming variant still passes
against the pre-fix code.
Add a test that builds the array first and then prunes element 0 of eight,
asserting the identity of every survivor rather than just the resulting
count. That fails against the pre-fix code deterministically, with
arr[0] == arr[1] and the last element dropped. Note the limitation on the
streaming test so the next reader does not try to strengthen it in place.
Also cover, all previously untested:
- json_free of a hash key, and of both halves of a pair, where the entry
survives with a JSON_NULL stand-in until both are gone
- repeated json_read_tree over a line-delimited stream, which is what the
filter loaders and -L / -E do, asserting each detached tree survives the
next read and the parser's destruction
- json_stringify of a partially-parsed tree, the json_context error path
- json_stringify across an embedded NUL, which fails against the c_str()
walk this branch replaces
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Add changelog entries for three unadvertised fixes
The array-splicing fix goes first: it is a memory-corruption fix, and it is
the strongest illustration of why the ownership model is worth having,
since it is exactly the failure the model makes unrepresentable.
Also the uninitialized read when a filter emitted "properties": null, and
the evaluator.hpp include guard that defined EVALUATOR HPP and so never
guarded anything.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
* Consolidate the jsonpull comments
Comments were 25% of the added lines, and the ownership model was spelled
out in five places. Collect it into one block at the top of jsonpull.h and
point at it from the rest, cutting the ratio to 14% and the total by about
200 lines.
Removed the duplicate explanations of the deleter dispatch, of what detach
does to the back-pointers, and of "json_read returns intermediate
containers, do not free them". Trimmed the comments that argued for a
choice rather than described the code -- the reserve(2) / reserve(4)
rationales, the string-buffer copy, the pmtiles check ordering -- to a line
each, and shortened the test preambles, keeping the parts that say why a
test is shaped the way it is.
No code changes; the test suite is unchanged in both configurations.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017KNxyHKasyWrWcvre2yK4r
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Claude <noreply@anthropic.com>
* Track output position at the file level instead of within each tile
* Track file position where the child tile data begins
* Add option and document its intended behavior
* Changing the detail loop to account for stopping early
* I forgot I already added an option for this
* Stop early if we can make a complete tile
* Add a test of zoom truncation with limited feature count
* Forgot to commit the actual code change
* Make room for a vertex count in the header of each serialized tile
* Estimate tile complexity; don't try truncating when unlikely to work
* Be more conservative, because ever retrying a tile is a big speed hit
* If stopping early, don't simplify or clean; leave that to overzoom
* Add tiny polygon reduction / dust to overzoom
* Don't try to stop early in the children if we dropped anything by rate
* Fflush here too before pwriting
* Don't stop early if we ended up dropping any features.
Rework the can-the-next-zoom-stop-early logic to avoid going
one zoom further than needed.
* Fix warning
* Fix warnings
* Oops, checking for the wrong expected return value
* Cleanup from adding line simplification in overzoom
* Current (wrong) behavior when combining coalescing and truncating
* Keep a list of parent tiles to skip rather than truncating
* Now the coalesced tiles in z12 get children in z13
* Don't double-count feature dropping when the zoom level is retried
* Correct README description
* Remove todo about special case below basezoom, which is accounted for
* Be a little more aggressive in drop-densest determination
* Scale tile feature limit for megatiles in the same way as byte limit
* Fully deprecate -detect-shared-borders into an alias
* Track the distances found in the douglas-peucker recursion
* Serialize and deserialize the distance with the vertices
* Revert "Serialize and deserialize the distance with the vertices"
This reverts commit 753f1b7909.
* Revert "Track the distances found in the douglas-peucker recursion"
This reverts commit e5361f8c22.
* Revert "Fully deprecate -detect-shared-borders into an alias"
This reverts commit 0698aeb766.
* Better tracking of whether we failed to make a full-detail tile
* Put a bloom filter in front of the binary search for shared nodes
* Forgot to take out this printf
* Improve dispatch of tiling tasks
* Still dispatch the biggest tasks first
* Track zoom truncation in the strategies list in the tileset metadata
* Prescan for small deltas before doing proper simplification
* Revert "Prescan for small deltas before doing proper simplification"
This reverts commit d1d8238b83.
* Update version and changelog
* Rename to --generate-variable-depth-tile-pyramid
* Read unidecode data, do some plumbing of it
* More unidecode plumbing
* Do the unidecode smashing, but it doesn't seem to be working
* Ah, that's better!
* Add missing header
* And reorder the includes too
* Shortcut when there is no unidecode data to work with
* Update version and changelog
* Avoid repeated unidecode smashing of the same constant string
* Scale the tile size limit up with the multiplier at low zooms
* Add a test to demonstrate that high zoom tiles can't be extra large
* Update changelog and version
* Fail more cleanly when a tile can't be made small enough
* Guard against a cluster where the start marker has been dropped
* Look harder for a working feature interval instead of giving up
* That change to the drop-smallest logic changed a test output
* Update changelog
* Add explanatory comment
* Speed up mvt_value comparison
* Converting repetitive ifs to cases
* More conversions from ifs to cases
* Optimize the always-true filter case
* Don't convert types of attributes without accumulators
* Unordered map seems to be faster than map
* Add missing header
* Fix some warnings
* Fix the warnings better
* Avoid an int->string->int conversion
* Lazily initialize layer key and values maps when actually needed
* More switches from maps to unordered_maps
* Sure, I'll take the microoptimization
* More emplacement
* Save some copies
* Emplaces and moves
* Lazy linear scan of attributes instead of building a map
* Extra printfs, missing header
* Avoid clipping if the input and output tiles are the same
* But do clip if the tile extent is being reduced
* Make sure I'm not constructing std::strings here at runtime
* More worrying about runtime string construction
* A couple more std::moves
* Const references!
* More const references
* Another std::move
* Make the string_value of mvt_value std::optional
* Reserve storage when decoding
* Provision for different mvt_values to share a string pool
* Use the string pool when decoding
* Avoid another string construction
* Try limiting the depth of the search for duplicate attributes
* Revert "Try limiting the depth of the search for duplicate attributes"
This reverts commit 9ec94a15ff.
* Update changelog
* Fix typo noticed during code review
* Starting to factor out attribute accumulation into its own file
* Continuing to factor out attribute accumulation
* Reduce duplicate code
* Plumbing the accumulate-attribute option around
* Call the attribute accumulator
* Test that accumulation works
* Add missing #includes
* Don't sort within individual multiplier clusters
Doing so throws off the spatial distribution of the low zooms
* Docs and changelog
* Add comments
* Start of externalizing polygon shard detection
* Completely untested external quicksort
* Add unit test for external quicksort
* Remember to clean up temporary files
* Sort and scan the vertices
* Bring over more vertex logic
* Make nodes from vertices
* Checkpoint on switching over to global shared nodes
* Do the thing
* Revert unintended change to coalesced linestring behavior
* Take shared nodes into account in early simplification
* Let it do more sorting in memory
* Fix overnoding of collinear linestrings
* Remove duplicate nodes, since only duplicate vertices now matter
* Remember to delete temporary files
* Fix out of bounds memory access below, apparently
* Fix the actual undefined behavior
* Still running out of memory in one case. Find out where.
* Forgot the conditional
* Try again to make it not run out of memory
* Update version and changelog
* Remove the concept of "separate metadata"
This was an extra level of attribute indirection (features point
to metadata records which point to key and value strings) which was
intended to reduce the size of temporary storage for features with
large numbers of attributes that were also spread across large numbers
of tiles at maxzoom.
For other kinds of features, the extra indirection slowed things down
instead, and, especially when maxzoom guessing was being used, many more
features were having their metadata externalized than could actually
benefit from it.
* Shave a few bytes off temporary files by using more unsigned integers
* Flush stderr after logging progress
* Revert "Shave a few bytes off temporary files by using more unsigned integers"
This reverts commit eef29084ec.
* Limit the size of the string pools and trees to fit in memory
* Add missing #include
* Move the string pool and search tree from mmap to allocated memory
* Sort in allocated rather than mapped memory too
* Also use pread instead of mapping to read in the data to sort
* When the pool gets too big, switch to just the file, not memory
* Switch string pool from memory to disk when memory is 10% full
* Add to-memory versions of the serialization functions
* Crashy work in progress toward compression
* Fix the pointer bug that was causing the crash
* Serialize features into memory rather than straight to disk
* Compress individual features in the temporary files
* Don't need to store the length of the geometry
* Remove per-feature compression; move minzoom back into the object
* Start adding a stream compressor object
* Track file position within fwrite_check()
* Add compressed stream writer functions
* Pull the writing of the serialized feature out to the callers
* Starting toward compression again from a different point
* Hook up more compression functions
* Remove unused code from the other day
* Make enough deflate calls to flush out all the buffered data
* Start on decompression
* Tile number is uncompressed, tile content is compressed
* Work on alternating compressed and uncompressed in decompression
* Closer, but still doesn't work
* Sort of works
* Works until we get to concatenated tiles
* More attempts that don't work
* One bug down
* It made a tileset!
* Handle nonzero initial zooms
* Fix seeking within compressed feature streams
* Tests pass!
* Remove debug spew
* Oops: remember to delete the temporary files so they don't hang around
* Test that fails with the current compression code
* Properly account for bytes read while closing the compressed stream
* Limit the number of warnings about bad label points
* A little more armor when closing decompression
* This time for sure
* A different, less fragile, test that failed previously with compression
* Move feature stream compression to its own file
* Remove now-unused code to deserialize from a file
* Forgot to add the new files
* Remove a little debugging logging
* Add a couple of comments on what it means to be within decompression
* Fix indentation
* Update changelog. Remove stray debugging comment.
* Add an option to limit geometry vertex count
* Update docs and changelog
* Track desired feature count and geometry size in strategies
* Remove dead code for a long-forgotten inaccessible option
* Fix the option name in changelog
* Refine coalesce-smallest to only coalesce onto other small features
* Clean polygons before coalescing-as-needed
* Don't accumulate tiny polygon holes as negative dust
* Add the option not to limit the feature count at maxzoom
* Add options to limit feature count more abruptly in each tile
* Revert "Add the option not to limit the feature count at maxzoom"
This reverts commit ace173ab7a.
* Revert "Fix the option name in changelog"
This reverts commit 38dbce8405.
* Revert "Add an option to limit geometry vertex count"
This reverts commit d379fdf06b.
* Remove test for reverted option
* Remove more leftovers from geometry size limiting
* Update changelog
* Add an option to retain extra coordinate precision at maxzoom
* Make sure not to shift away the extra detail from coordinates
* Add an option to convert double-precision attributes to single
* Sort attribute values in tiles to make them compress a little better
* Slightly improve polygon simplification
By choosing a point that would be retained after simplification
to be the start/end point that always gets retained
* I regret making all of these tests involve polygons
* Add an option to specify the size of tiny polygons
* Fix accidental requiring of argument for --single-precision
* Guard against duplicate points when generating "sizes" for them
* Restore the intended behavior that tiny polygons don't get simplified
* Make the extra detail settable rather than always maximizing it
* Revert "Improve maxzoom guessing for tightly-clustered point data sources (#4)"
This reverts commit fec5e8354c.
* Add an option to prevent choosing a base zoom higher than the maxzoom
* Keep the drop rate high enough when the basezoom gets constrained
* Revert "Revert "Improve maxzoom guessing for tightly-clustered point data sources (#4)""
This reverts commit db6bc27d9e.
* Add --order-by and --order-descending options
* Accept multiple --order-by and --order-descending-by sort keys
* Handle points too when dropping or coalescing the "smallest" features.
* Add statistics of tile size reduction strategies to tileset metadata
* Update changelog and version
* Update documentation
This moves filtering from the serialization stage to the
tiling stage so that the zoom level can be known to the filter.
The side effect is to carry null attributes much further through
the pipeline than previously.
