Clan Tools has a number of tasks that run on an "out of band" scheduler; the most outwardly noticeable of these tasks is battle loading and payout calculations.
This scheduler stops working from time to time, seemingly at random. I had thought it was a memory issue and thus thought it was most likely to fail during the daily maintenance tasks which tend to eat up a lot of memory, so I based my (manual) monitoring around that assumption. Yesterday has shown my existing approach is insufficient.
I have created a semi-automated way to monitor the scheduler, and to detect when it failed. This should, at the very least, decrease the downtime when it does fail in addition to giving me a better picture of what is failing.
About two weeks ago I started a MSCS program which--in contrast to the past few months before where I effectively had nothing I needed to do, or even the a BSCS program I was in--has left me with a lot less time to spend working on Clan Tools. This is especially true of long chunks of time which are ideal when working on the more complex features.
To be clear, I have no intention of stopping my work on maintaining and improving Clan Tools. I'm writing this simply to let everyone know what my situation is and why updates have been, and will continue to be, slower to come out relative to the changes they bring.
The filter is primarily a (case insensitive) exact match "contains" search, though missing spaces and dashes (-) will still result in a match (e.g. e100 will match E 100 despite the lack of a space in the search).
Detailed Tank Stats for individual tanks (example) can now be viewed via new "Details" link (in the tank list). Note that the "maximum" data isn't available (via the API) for some game modes.
Additionally, if any of your battles from last night (2015-05-11) loaded encounter on province battles at attacks, I can manually fix them if you let me know which ones need to be corrected.
WG released an update to clan wars, at least on the NA server. This change included a change to the formatting of some of the data Clan Tools uses to load clan wars battles to the current level of detail.
I believe I've made the necessary changes to deal with the new format. However, all of my testing is currently limited to assumptions as no battles have had their times fixed (which battle loading relies heavily on). Additionally, I suspect that past four participants (3 attackers + one defender) the approach being used might break down, but again without actual data to test with I can only make assumptions.
Additionally, even if the additional data cannot be loaded, the system will now still load the battle, though the level of detail in the information presented may be lower. Specifically, the opposing clan won't be known, and there will be no distinction between an attack and an encounter on a province (when two or more clans chip the same province) (any non-defense battle for province will be considered an attack).
Regardless, I will monitor the situation and attempt fix any issues that arise as quickly as possible.
