Sitebeam 5.0.13

This version was released on 27th September 2012.

Summary

Detailed list

Improvements

  • Added new Open Graph test.
  • Progress display greatly improved.
    • Detail of what the server is doing is now shown.
    • The estimated amount of time remaining is shown.
    • The time spent testing is now shown.
    • The rate at which pages are being downloaded from the site is now shown.
  • All reports are faster and more stable.
    • Sitebeam now respects caching headers, avoiding redundant downloads.
    • Tests now re-order and combine themselves for performance.
    • Transfers are now optimized into batches for performance.
  • Added Beta API for selected customers. Currently supports querying users, adding sites and running reports.
  • Adding a website can now be done with the keyboard – pressing Enter to submit the name field submits the whole form (previously, it was necessary to click the Done button manually).
  • The Freshness test now shows where it identified dates within the website.
  • The Broken links test now scales to handle much bigger sites than before.
  • Added ability to limit Sitebeam (WebScan not affected) to test all sites for specific languages only. This greatly accelerates testing where the language is always constant. See: Account > Languages.
  • Sitebeam is now much faster when the language list is already specified.
  • Added detection of additional date formats to Freshness test (partial dates without a year, e.g. “2nd July”).
  • Sites which have been created but have no reports are now automatically deleted after 60 days.
  • The Internal search test can now detect Google’s Javascript embedded GCSE (Google Custom Search Engine).
  • When exporting a PDF and using Chrome, an explanation appears which does not require the use of an external PDF exporter (as Chrome now supports native export to PDF).
  • The Cookie Law test now handles partial failures much more gracefully: the overall score is made informational, the issue is explained and diagnostic information is displayed.
  • Optimized some communications between servers for speed, particularly when a server fails.
  • Optimized low-level download requests for speed (removed superfluous hRequest caching in preference for hQuery caching).
  • Added support for detecting links to Privacy policies to the Cookie Law test.
  • (Silktide only) added substantial report profiling and diagnostics.
  • (Silktide only) added capability to test UI database load.
  • (Silktide only) added experimental support for proxy servers.
  • (Silktide only) added additional detail to test suite for Facebook detection.
  • (Silktide only) added ability to soft/hard restart a Test Server from progress screen.

Changes

  • The Search ranking test now tests Google via a physical server located in either the UK or US when testing those regions. This is because Google has started to return different results for English queries based entirely on location, even if the user manually overrides this. Our solution allows you to test UK results from the US and vice versa.
  • The Links test no longer includes a complete list of every link with an issue. This data was often so large as to occupy more space than the rest of the report combined; the test now simply lists pages with any issues.
  • The Visual interest test is now disabled by default. This test was relatively contentious and slow to run; you can re-enable it on demand under Account > Visual Interest.
  • The Broken links test now divides links into internal (within the site) and external. The test now scores links which are broken internally much more harshly than external links, so scores will change after this update.
  • Many more tests now hide themselves (consider themselves ‘boring’) if their results are non-actionable; this is particularly true of tests where scores are high. The scores are unaffected.
  • The penalty for sites using a Facebook personal profile is now lower. This is because some pages legitimately use Facebook profiles with the ‘Subscribe’ option enabled, instead of a personal page.
  • The Cookie Law test now awards 1 extra point (i.e. 1/10th of the total score) for sites that make any attempt to comply with the law whatsoever.
  • The Image usage test now penalises the absence of explicit image sizes less. This represents a gradual shift towards responsive layouts.
  • Refined the parsing of <sup> and <sub> for plaintext and spelling calculations (previously, these were considered run-on text, they are now separated).

Fixes

  • Fixed bug where an HTML page with a conditional comment wrapping the opening <html> tag would cause the HTML parser to incorrectly consider the closing conditional comment part of the page’s text.
  • Fixed bug where a badly formed HTML page could overload the body text detection, causing certain tests to crash.
  • Fixed bug where the transfer rate limiter could prevent reports from transferring back indefinitely, if a previous report crashed during a spawn process.
  • Fixed bug where transferring a new result type to a fresh install could fail if the UI had not yet invoked the required table.
  • Fixed the Incoming links test results not returning results (due to a change in the SEOmoz API).
  • Fixed a bug where results could appear duplicated for specific tests.
  • Fixed a bug where running the Writing report type could cause the language detection not to run as expected.
  • Fixed newly missing columns in the Search Engine Results table, as of v5.0.12.
  • Fixed bug where attempting to export a comparison report would crash.
  • Fixed bug where the Pages table inside the Speed test was missing the Size column in some reports.
  • Fixed bug where a Twitter URL that contained escaped entities after a hash symbol might be ignored.
  • Fixed bug where a <script> element inside a <script> element (e.g. inside a document.write statement) could fool Sitebeam into considering the nested <script> as an inline <script> at that point.
  • Fixed bug with Image usage test where the percentage of images in a non-web friendly format was not rounded to a sane number of decimal places.

Was this article helpful? Contact our support team if you have a question.