* The state of syntax stops is now properly included in the cache key for the
tokenizer-internal backtracking cache. This fixes some mis-parses when
re-parsing a bit of text with different flags.
* Clear the backtracking cache after each toplevelblock. This drops the peak
memory usage when expanding [[:en:Barack Obama]] from ~380M to ~110M.
Change-Id: Icdb879cae5907e4595903dd6acba2e686e8c2e4b
This patch fixes a tokenizer syntax error encountered on
[[:en:Template:JacksonvilleWikiProject-Member]] and [[:en:Template:Infobox
former country]] by allowing optional whitespace before start-of-line template
syntax.
Change-Id: Ic214a731de58bf766e51f23d5e24ea2ce6788f58
254 round-trip tests (up from 184) are now passing.
Also:
* tweaked runtests.sh slightly (use less -R instead of -r).
* made sure the EOFTk is preserved in phase 3 transforms
Change-Id: I1de22186bdb78e52019370e43f096877005b8f5a
* Tokens are now immutable. The progress of transformations is tracked on
chunks instead of tokens. Tokenizer output is cached and can be directly
returned without a need for cloning. Transforms are required to clone or
newly create tokens they are modifying.
* Expansions per chunk are now shared between equivalent frames via a cache
stored on the chunk itself. Equivalence of frames is not yet ideal though,
as right now a hash tree of *unexpanded* arguments is used. This should be
switched to a hash of the fully expanded local parameters instead.
* There is now a vastly improved maybeSyncReturn wrapper for async transforms
that either forwards processing to the iterative transformTokens if the
current transform is still ongoing, or manages a recursive transformation if
needed.
* Parameters for parser functions are now wrapped in abstract Params and
ParserValue objects, which support some handy on-demand *value* expansions.
Keys are always expanded. Parser functions are converted to use these
interfaces, and now properly expand their values in the correct frame.
Making this expansion lazier is certainly possible, but would complicate
transformTokens and other token-handling machinery. Need to investigate if
it would really be worth it. Dead branch elimination is certainly a bigger
win overall.
* Complex recursive asynchronous expansions should now be closer to correct
for both the iterative (transformTokens) and recursive (maybeSyncReturn
after transformTokens has returned) code paths.
* Performance degraded slightly. There are no micro-optimizations done yet
and the shared expansion cache still has a low hit rate. The progress
tracking on chunks is not yet perfect, so there are likely a lot of unneeded
re-expansions that can be easily eliminated. There is also more debug
tracing right now. Obama currently expands in 54 seconds on my laptop.
Change-Id: I4a603f3d3c70ca657ebda9fbb8570269f943d6b6
* [[:en:Barack Obama]] can now be expanded in 77 seconds using 330MB RAM,
while it would prevously run out of RAM after ~30 minutes. Wohoooo!
The token transform framework rework really paid off.
* 303 parser tests are passing in the new record time of 5.5 seconds. Two more
tests are passing since these tests expect the day of the week to be
Thursday. Won't be the case tomorrow.
Change-Id: I56e850838476b546df10c6a239c8c9e29a1a3136
This is a bit better than cloning tokens wholesale, but not by much. There is
a lot of potential for much better per-token caching with reduced token
cloning. Need to map out all dependencies besides token attributes expanded
from template parameters or other scoped state. Even if tokens themselves
don't need transformation, they might still need to be considered for other
token transformers, so simply keeping the final rank won't quite work even if
the token itself is fully transformed. As a minimum, a shallow clone would
need to be made and the rank reset (as in env.cloneTokens).
Change-Id: I4329113bb21750bae9a635229ed1b08da75dc614
* Added an LRU cache (using the lru-cache node module) for tokenizer output
* Mutation of nested attributes now replaces the containers. A shallow copy of
tokens is sufficient to isolate token transformations. Need to investigate
if we can actually get away without isolation and re-transformation for most
ordinary tokens.
Change-Id: I9136b1d7a1fbcc538183a319d4ecaa290d616fdf
Eat unbalanced external link parts within template parameters. This does not
produce the same output as the PHP parser
(try echo '{{YouTube}}' | node parse.js), but preserves a level of sanity.
Need to check how common this is for external links. If it is rare enough,
moving the ']' after the parser function manually would fix the rendering for
the YouTube case.
Change-Id: I597d808efff36baa22191e7946a0061cc31120e8
* DOM based on Wikia's thumb output: HTML5, clean caption without magnify
icon.
* basic RDFa annotations, but most options additionally in data-mw object-
might want to move more (or all?) of those into RDFa data using meta tags.
* no support yet for framed or other formats, image scaling etc
* also tweaked some config options in the environment
Change-Id: Ie461fcdce060cfc2dec65cc057709ae650ef3368
gets a bit closer to supporting table fragments passed through template
arguments. Next, we'll need a way to indicate start-of-line position to
enable sol block-levels in template parameters.
Example:
{|
{{#if: true|{{!}}Table cell|}}
|}
possible to support template / template argument expansion in image options,
and causes little trouble for wikilinks. Non-image wikilinks with multiple
text pipes are quite rare in the dumps, and concatenating description tokens
with a plain '|' is quite easy. 261 parser tests passing.
mediawiki.tokenizer.js module, and pass a reference to parse(). Faster
inline_breaks production using a JS function which seems to be generally
correct, but still breaks five tests when enabled. Seems to be some weird
interaction with peg.js, possibly something to do with caching.
other tokens. This is only the first half of the conversion. The next step is
to drop the type attribute on most tokens and match on the constructor in the
token transform machinery.
* TokenTransformDispatcher is now renamed to TokenTransformManager, and is
also turned into a base class
* SyncTokenTransformManager and AsyncTokenTransformManager subclass
TokenTransformManager and implement synchronous (phase 1,3) and asynchronous
(phase 2) transformation stages.
* Communication between stages uses the same chunk / end events as all the
other token stages.
* The AsyncTokenTransformManager now supports the creation of nested
AsyncTokenTransformManagers for template expansion.
The AsyncTokenTransformManager object takes on the responsibilities of a
preprocessor frame. Transforms are newly created (or potentially resurrected
from a cache), so that transforms do not have to worry about concurrency.
* The environment is pushed through to all transform managers and the
individual transforms.
The TokenTransformDispatcher now actually implements an asynchronous, phased
token transformation framework as described in
https://www.mediawiki.org/wiki/Future/Parser_development/Token_stream_transformations.
Additionally, the parser pipeline is now mostly held together using events.
The tokenizer still emits a lame single events with all tokens, as block-level
emission failed with scoping issues specific to the PEGJS parser generator.
All stages clean up when receiving the end tokens, so that the full pipeline
can be used for repeated parsing.
The QuoteTransformer is not yet 100% fixed to work with the new interface, and
the Cite extension is disabled for now pending adaptation. Bold-italic related
tests are failing currently.
