CPAN Testers is only made possible with the support of our sponsors.
For more information on sponsoring, please visit the I ♥ CPAN Testers website.

Upgrade Notice

The CPAN Testers Blog site has been upgraded since you last accessed the site. Please press the F5 key or CTRL-R to refresh your browser cache to use the latest javascript and CSS files.

News & Views

During August the CPAN Testers Reports server suffered a problem with the RAID array. Thankfully, it just needed the disk to be re-mirrored. However, this did mean that for a couple of weeks the server, and the websites did seem unresponsive. To lessen the load I reduced the builder processes, and only ran them for a few hours a day. As such, you may not have had your pages updated as frequently. The RAID array is now fully operational, and the pages are now back up to date. Currently reports look to be about an hour behind and the less frequently used pages are less than a day behind.

The new CPAN Testers Admin site is coming along. The bulk of the site is now completed, and work has already started on the integration for CPAN Testers Reports Builder and the CPAN Testers Statistics, to ensure removed reports are accounted for. The first part of the site to be launched will only be looking at historical reports. At some point in the future, I would also like to make it easier to report runaway smokers, so we can (un)block them more pro-actively.

On the mailing list last month, a thread which started the previous month by Nat Goodman, was carried over when Nat mentioning he was using a testing pragma. Steffen Schwigon picked up on this, and asked Nat to elborate, which he did. It seems due to a problem with building pre-requisites, Nat has employed a pragma to help catch when a situation isn't satisfied, in his case when the tools to connect to a MySQL database aren't loaded. It's an interesting use of a pragma and the TAP parser.

Dave Cantrell picked up on the fact that we no longer update the SQLite version of the cpanstats database. As mentioned previously, SQLite has not being storing results correctly for several years. Both Andreas and myself have tried to diagnose the problem without any luck. As the file itself is several gigabytes in size, having this downloaded several times a day was taking up valuable disk IO. As such, the processes to update the file have now been disabled. The recommended way to retrieve these reports is now through the use of CPAN::Testers::WWW::Reports::Query::Reports, such that you can create your own database, and update with only the latest reports, rather than trying to download them all every time.

Neil Bowers asked about getting summary data for a particular distribution. For those interested in this sort of data, there are two distributions you can use, CPAN::Testers::WWW::Reports::Query::AJAX, providing basic data, or Leo Lapworth's more detailed CPAN::Testers::Reports::Query::JSON. Both get the same data, but present it in a slightly differnt format. Rocky Bernstein, also asked how to get the raw reports. However, there is a slight misunderstanding of what a raw report is. In the olden days, we used email, and as such the reports were all in plain text. With CPAN Testers 2.0 we now create structure data. We still display a plain text version, but more of the data is being built as components, to enable better analysis in the future. As such, the reports that appear on the Reports website, are not necessarily the original plain text versions, but rather constructed to look like plain text versions. Thus clicking the 'raw' button doesn't really present the raw email format that it used to be, but a close approximation inside PRE tags. The main reason for this is encoding. HTML handles encoding slightly better than plain text sent to a browser. We have had problems with this in the past, and it has proved more reliable to handle the encoding HTML than plain text. At some point, alongside the 'raw' button, there will be a 'data' button, which will return the report as structured data in JSON format.

Ron Savage highlighted a problem with IO::Pipe and Proc::Fork Tests hanging on Windows. David Golden pointed out that forking on Windows has always been problematic, and can cause unexpected results. The conclusion for Ron was to drop Capture::Tiny, however, if you're coding to be cross platform, it is always worth bearing in mind that Windows and fork() are not the bext of friends.

That's it for this month. Hopefully, we'll have more news about the Admin site next month, and may be even some server news.

Comments

No Comments


Add A Comment

Comments now have limited rich text support thanks to TinyMCE.

*
*