Noah wrote in #280:
> Maybe we should just drop the test. The assumption when we wrote the
> test was that a race condition would not be so frequent in practice.
> We've seen that under the contrived conditions created by the test
> case, it is frequent. But that's ok
The arbitary value `25` was used but in prace it's quite possible
for more than 25 writing threads to have checked the cookie count
limit before adding their cookie. In practice we see Travis failing
on this test quite often, every few builds in fact.
I think using `threads.length` (i.e. 200) should cover the worst
case possibility where every thread reads a stale count and tries
to add their cookie.
Fixes#274
Jetty 9.4.12+ is required for TLS to work correctly under JDK11
(due to SSL handshake failures). In order to upgrade
Jetty we also need to upgrade Restlet.
There is one intentional change in behaviour to simplify
upgrading. We remove a workaround for an old [webkit bug]
where the browser claimed to prefer application/xml. The
bug was fixed in 2011.
[webkit bug]: https://bugs.webkit.org/show_bug.cgi?id=27267
Summary of Jetty API changes:
- package names changed such as org.mortbay -> org.eclipse
- SocketConnector and SslSocketConnector merged to ServerConnector
- HashUserRealm split into UserStore and HashLoginService
- SecurityHandler -> ConstraintSecurityHandler
Summary of Restlet API changes:
- some classes have moved package (Request, Response, Router etc)
- ServerResource replaces Resource
- represent(), acceptRepresentation() renamed to get(), post()
- constructors were replaced by an init() method
- setModifiable() was removed
- getRequest().getEntityAsForm() -> new Form(entity)
- Guard -> ChallengeAuthenticator
Fixes#275
Fixes#268 ssl handshake_failure. Support for jdk11 was added in jetty
9.4.14 but we may as well bump to the latest stable version.
In Jetty 9 it appears the request is logged after the response is sent.
Thus it was racing with the assertions that check the client IP. So to
fix this rather stashing the 'lastRequest' we just make the server
echo the client's IP in a response header.
Some other minor tweaks were needed due to changes in Jetty behaviour:
- We stop checking the length of the raw response. It doesn't tell us
anything and easily varies.
- Jetty now generates Set-Cookie with a space after the ;
- Jetty now lowercases the word "basic" in WWW-Authenticate header
- testLaxUrlEncoding(): Jetty now rejects bad paths with a 400 error so
we disable the response checks. The actual request line is
still checked which is the important thing.
Note: This patch does not affect the version of jetty used by the
Heritrix admin console. That will be tackled separately.
Without this change (or other measures), we sometimes get nulls in the
ExtractorYoutubeDL log for containing page information. We'll run this
on QA for a while and see if it causes any problems.
nlevitt [1:59 PM]
https://github.com/internetarchive/heritrix3/blob/master/modules/src/main/java/org/archive/modules/CrawlURI.java#L878
drops some stuff from `CrawlURI.data` after processing a uri, even if it needs to be processed again
there is a list of keys that shouldn’t be dropped (`persistentKeys`), but it is final and private
so if you’re writing your own heritrix module and you want to keep some information in CrawlURI.data, it usually works, except when the url is processed more than once (like when it needs a prereq like robots.txt the first time)
in practice it seems that most data is persisted, that is, most commonly used keys are in `persistentKeys`
in a crawl with pretty standard configuration i’m mostly seeing `prerequisite-uri` dropped and occasionally `fetch-completed-time` and `fetch-began-time` being dropped
i’m highly skeptical of the value of dropping keys at all and i’m tempted to get rid of this entirely, make all the keys persistent in other words
soliciting feedback (edited)
anjackson [2:39 PM]
My immediate reaction is HARD AGREE. It looks like Really Old Code though (https://github.com/internetarchive/heritrix3/blame/7d3eff5269142c77fa4b988396153f4c29d16caa/modules/src/main/java/org/archive/modules/CrawlURI.java#L878)
so the reasons for doing so may have been lost in time.
Hm, looking at usage: https://github.com/internetarchive/heritrix3/blob/a60b2ef3875ad47f57b0c6b3c0b19f86c40a12f7/engine/src/main/java/org/archive/crawler/frontier/WorkQueueFrontier.java#L954-L955
engine/src/main/java/org/archive/crawler/frontier/WorkQueueFrontier.java:954-955
curi.processingCleanup(); // lose state that shouldn't burden
// retry
I guess there's a concern that there may be state in there that is set during a fetch and may cause problems if the same CrawlURI is deferred?
But I'm not aware of anything in the fetch chain that behaves like that.
nlevitt [3:02 PM]
oh, i missed `CrawlURI.addDataPersistentMember(String)` et al. still...
There is already a clause for this in logging.properties, but it's using
log4j. It was dumping stack traces every time the client was dubious of
heritrix's self-signed certificate.
Why do we have so many identical log4j.xml's? 🤷♂️
Commas are allowed if they're in the middle of the URL. Consequently:
srcset="a,b,,c," => ["a,b,,c"]
srcset="a, b,, c," => ["a", "b", "c"]
They occur particularly commonly in data: URLs before the base64 value.
Commas are also allowed in descriptors if they are enclosed by parens:
srcset="a (b,c),d" => ["a", "d"]
Spec: https://html.spec.whatwg.org/multipage/images.html#parsing-a-srcset-attribute
DecideRuleSequence already has the optimization that I was looking for,
namely, don't bother evaluating a DecideRule if we know it won't change
the current result. For some reason PredicatedDecideRule, which is a
parent class to most of the decide rules in heritrix, didn't implement
onlyDecision(). Certain crawl configurations could see significant
performance improvement with this change.
I don't know why we've never seen this before, and now we suddenly have
a case of it, but this is the exception:
2018-07-23 17:47:01.123 SEVERE thread-2875088 org.archive.crawler.framework.CrawlJob.beansException() Failed to start bean 'scope'; nested exception is java.lang.IllegalStateException: java.nio.file.NoSuchFileException: /1/ait-h3-jobs/8144-20180723162745141/20180723174701/logs/scope.log.lck
org.springframework.context.ApplicationContextException: Failed to start bean 'scope'; nested exception is java.lang.IllegalStateException: java.nio.file.NoSuchFileException: /1/ait-h3-jobs/8144-20180723162745141/20180723174701/logs/scope.log.lck
at org.springframework.context.support.DefaultLifecycleProcessor.doStart(DefaultLifecycleProcessor.java:169)
at org.springframework.context.support.DefaultLifecycleProcessor.access$1(DefaultLifecycleProcessor.java:154)
at org.springframework.context.support.DefaultLifecycleProcessor$LifecycleGroup.start(DefaultLifecycleProcessor.java:335)
at org.springframework.context.support.DefaultLifecycleProcessor.startBeans(DefaultLifecycleProcessor.java:143)
at org.springframework.context.support.DefaultLifecycleProcessor.start(DefaultLifecycleProcessor.java:89)
at org.springframework.context.support.AbstractApplicationContext.start(AbstractApplicationContext.java:1236)
at org.archive.spring.PathSharingContext.start(PathSharingContext.java:115)
at org.archive.crawler.framework.CrawlJob.startContext(CrawlJob.java:455)
at org.archive.crawler.framework.CrawlJob$1.run(CrawlJob.java:428)
Caused by: java.lang.IllegalStateException: java.nio.file.NoSuchFileException: /1/ait-h3-jobs/8144-20180723162745141/20180723174701/logs/scope.log.lck
at org.archive.crawler.reporting.CrawlerLoggerModule.setupSimpleLog(CrawlerLoggerModule.java:298)
at org.archive.modules.deciderules.DecideRuleSequence.start(DecideRuleSequence.java:174)
at org.springframework.context.support.DefaultLifecycleProcessor.doStart(DefaultLifecycleProcessor.java:166)
... 8 more
Caused by: java.nio.file.NoSuchFileException: /1/ait-h3-jobs/8144-20180723162745141/20180723174701/logs/scope.log.lck
at sun.nio.fs.UnixException.translateToIOException(UnixException.java:86)
at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:102)
at sun.nio.fs.UnixException.rethrowAsIOException(UnixException.java:107)
at sun.nio.fs.UnixFileSystemProvider.newFileChannel(UnixFileSystemProvider.java:177)
at java.nio.channels.FileChannel.open(FileChannel.java:287)
at java.nio.channels.FileChannel.open(FileChannel.java:335)
at java.util.logging.FileHandler.openFiles(FileHandler.java:478)
at java.util.logging.FileHandler.<init>(FileHandler.java:344)
at org.archive.io.GenerationFileHandler.<init>(GenerationFileHandler.java:63)
at org.archive.io.GenerationFileHandler.makeNew(GenerationFileHandler.java:158)
at org.archive.crawler.reporting.CrawlerLoggerModule.setupLogFile(CrawlerLoggerModule.java:275)
at org.archive.crawler.reporting.CrawlerLoggerModule.setupSimpleLog(CrawlerLoggerModule.java:296)
... 10 more
ContentDigestHistoryTest does not close BdbModule. It results in failure to delete bdb directory in following tests.
Added tearDown() method that closes BdbModule.
Since we were only enforcing the size limit after reading each
line a single very long line could cause us to to run out of
memory.
This change enforces the character limit on each read character
not just on line boundaries.
We also correct the count of processed characters in the warning
message, which was not counting newline and linefeed characters.
Alex Osborne