* also fix new FormLoginProcessorTest.testFormLoginExtraInputs() in context of form-login-multipart changes
* master:
fix NullPointerException when using old HtmlFormCredential login
to make this branch only about multipart/form-data, undo the changes to handle more complicated forms with extra input fields; add unit test; store CrawlURI submit data in http-library agnostic way and remove httpcomponents-specific code from HTMLForm, refactor creation of post data accordingly
Be sure to return at least one of the candidates
Remove unneeded constant
Make sure to check form encoding type when addign form to curi data
More changes for https://webarchive.jira.com/browse/ARI-4656
More changes
Save enctype of form in submission curi
Add ability to submit forms with enctype="multipart/form-data"
If multiple username inputs in login form, favor the one with the word 'login' in it.
* master:
license header
check that sourceTag of CrawlURI actually matches configured sourceTag
remove already-outdated stuff from javadoc
handle multiple clauses for same user agent in robots.txt
Hook in submitted seeds properly.
avoid spurious logging
try very hard to start url consumer, and therefore bind the queue to the routing key, so that no messages are dropped, before crawling starts (should always work unless rabbitmq is down); some other tweaks for clarity and stability
* seed-limits:
change class originally known as SeedLimitsEnforcer to SourceQuotaEnforcer; make it a Processor instead of a DecideRule (because checking quota at link scoping time doesn't work, since many urls which would go over quota can be added to the frontier); support quotas on any of the fields tracked by CrawledBytesHistotable
fix checkpointing problems with new statsBySource
SeedLimitsEnforcer (contrib) DecideRule that rejects CrawlURI if source seed byte or document limit has been reached
SourceSeedDecideRule applies the configured decision for any URI with discovered from one of a set of seeds
add support to StatisticsTracker to keep a CrawledBytesHistotable per source tag when trackSources is enabled; integration test for this functionality
The previous capture may also have been a 304 but there must be an
original 200 capture at the start. This assumption may fail when using
an index generated by a version of Heritrix without this fix. But even
then you should be no worse off than before.
Hunter Stern