Recent content by Irrat

I don't blame at all for not wanting to deal with it, reconciliation is a pain and no one wanted to play with prefs in the first place. It only became a thing due to issues with failed saves, then my complaints about io usage. I went in a slightly different direction regardless, for the atomic...

Alright, so what I'm looking at is adding _state which works the same way as _spec Which has stuffed into it Result a boolean modifier, simply if maximizer failed or not Failure Reason A ; seperated list of reasons for the maximizer failure. Combos Considered A numeric count of how many combos...

Duplicate of this, https://kolmafia.us/threads/29151-error-on-modifier-maximizer-npe.32729/ Apparently fixed in newer builds.

I'm not sure what the expected invocation is, but heartstone_middle_letter() does not understand when a monster is replaced. In this situation, I fought a Dusken Raider Ghost which gives the letter I. I then spoke to some fish which gives no letter. Indeed, I do not have the skill available...

Anyone have any thoughts about this before I mark the pr as ready?

Currently when you run maximize via some method that doesn't return a boolean, the script cannot tell if the maximize failed or not. I'd personally prefer not to call two maximizes, when I should only need to call one. Especially because you can, with the below method, limit the scope the...

What's the actual maximizer string you're calling? I thought maximize -shirt adv means it will not modify your shirt mode. Or maximize +equip Jurassic Parka (spikolodon mode) +adv will always have your parka in spike mode. It sounds like you're saying you didn't do any of these, and the bug is...

Yep, the largest remaining IO writes were because some files I had, grew larger than I anticipated. Which is a bit ironic given that they're only written to for about 5 minutes a day. So anyways, has anyone else crunched their own numbers for writes from mafia? I do find it a bit interesting...

So io says about 1.64gb written. 5.91gb written to disk as per smartctl. Although come to think of it, I had been doing some builds over that period, which would explain why it's still higher IO than normal. As for mafia itself still having somewhat high io, no idea. But it's clearly shrunk...

As a small update, been working on this PR; https://github.com/kolmafia/kolmafia/pull/3567 The approach I'm currently taking in it is writing preference changes as an encoded key=value\n. Encoded using the existing encoding setup. Each change is appended to a irrat_pref.journal, supporting...

Monitoring with csysdig, or sudo csysdig -d 10000 and changing view + sort by, there's one file that always topped the results with a refresh time of 10 seconds. irrat_prefs.txt About 3-4mb on average per 10s based over a ten minute sample. Sometimes it dropped below 1mb, and sometimes it went...

I believe I had tested against linux stats, not smartctl. I'll check those numbers later today after my day has run, can compare the smartctl then. I had forgotten to compare it originally, and the final numbers upset me enough that I ended up deleting the script I used to test. I hated the...

Yeah I'm not really sure about periodic saving either. Its really not great in comparison. I'm not sure if you meant kolmafia backups or private backups. I don't think there's much reason to be doing kolmafia backups as a log file is never being rewritten, only appended. We don't have to modify...

So while looking at this PR I was a bit surprised to find out that my 150kb prefs file is costing me about 5.4gb of writes a day from my 33k pref changes. We currently write the whole file out every time we change one value. To me that feels unacceptable, I want my drives to last longer than...

Universal Seasoning tracking in KoLMafia seems half done, as per the wiki. https://wiki.kingdomofloathing.com/Universal_Seasoning This is kind of an extension of this report, however it feels too much like it belongs in its own thread...