Search results

My threat to commit is intended to provoke discussion, so thank you. I have not noticed a significant performance hit with this PR compared to running the main branch. Somewhat tongue in cheek I'd say the performance cost of a corrupted file is infinitely high and so, by definition, anything...

My understanding of the code is that uses effective is that it is really a special case to deal with weapons that use something other than the "expected" stat when determining damage. Effective does nothing different if the weapon is the July 4 saber, replica thereof or June Cleaver. If Moxie...

I forced the code to think it was headless and tests pass repeatedly. Specifically I hardwired SwinglessUIUtils.isSwingAvailable() to be false for several runs (and then restored it the exiting code which returns true in the test environment). I'm willing to commit.

I'm giving up. If someone can write a failing test where the maximizer does not emit commands to equip a weapon and show me which weapon it should have equipped because the score is higher then I'm back in. But every test I have tried has passed and there are definitely cases where no weapon...

Not sure where this best fits. I wanted to make the house of twigs and spit. I used the gong from the item manager. My session log says I then played one turn in the relay browser. I noticed I did not have form of a bird so investigated. I found no indication that I was in Bird or Roach...

The above has been in stdout/err from tests for some time. The file test_charsheet_zootomist.html contains the skill and was most recently change on Feb. 17, 2025. Just related trivia since there is no action required unless KoLmafia wants to work around KoL's "error"

Another one. Debug includes "You may only buy 3 of this item per day from this store. You have already purchased 1 in the last 24 hours." The probable portion of the session log is Not sure what was managing the recovery - a mood, a recovery script or what.

Well I stripped out items and non-passive skills until the bloated test and the stripped down one have the same results. Curiously the non-passive knife skill seemed to be needed. If the weapon and off hand slots are empty they are not filled and I have not convinced myself that is...

So I have a test case that attempts to recreate the beginning of run conditions from https://kolmafia.us/threads/maximizer-didnt-equip-a-weapon.29577/page-6#post-176370 My test case is suspect in that it doesn't generate the same scores. if I were more motivated I would also investigate what...

Been there. Done that. Maybe :) See itShouldMaximizeAndEquipSelectedWeapon in RuntimeLibraryTest and the tests nested under Effective in MaximizerTest My recollection is that during test development there were identical copies of some of the tests that differed only in the presence or...

I had a couple of instances where the log was generated but they seemed to be the "This store doesn't have that item at that price. Perhaps someone beat you to it, or the shop owner changed the price of the item while you were browsing." case. Are they of interest if they happen again?

There are a few tests already committed that could be copied and modified to reproduce the above. There are at least three distinct possibilities and the difference is important because the nature of testing and fixing varies. The maximizer selects a weapon and then fails to equip it. A test...

We should discuss what a fix might be. "nothing" + offhand is the right answer in some cases. If we can ever find a case where (something + offhand) > (nothing + offhand) but the maximizer picks nothing then we have a bug that needs to be fixed. But until we find that something what we...

@Yellowbeard First, thank you for reporting. This has been a long and difficult issue to debug. The folks at autoscend graciously agreed to assist in gathering data which is why you get the messages you do. The hope is that someone will come up with a set of conditions where no weapon is...

Local test failures seemed to be caused by caching. Blew away a couple of subdirectories and everything works locally again. Looking at the effects of actually writing to disk and Swing availability on the Deck test.

Well... I decided to run the coverage reports locally since that was a quick way to find out whether the lines referenced above were executed or not. The Deck test has failed locally twice now. Go figure and stand by.

I never ran headless but once upon a time saw the deadlock a lot running with the GUI. Since my confidence level is related to tests passing/failing it is worth looking at the Deck test to see whether headless/Swing could be a factor. April 1 was an arbitrary deadline anyway. Thanks.

For reference see https://github.com/kolmafia/kolmafia/pull/1889 #1889 was a solution to preference file corruption. It appeared to address the issue so it was committed. But real world experience indicated a lot of deadlocks started occurring including one that occurred fairly consistently...

Thank you. I forgot log was there when I committed. I was failing to launch Kolmafia, could not figure out why and log captured STDOUT and STDERR from the process I launched. My mistake was in how I set up the command to run in ProcessBuilder.

New version - it does a lot of what I hoped for here. First it downloads the latest version of itself and restarts using that. It deletes older versions of itself. If the local Java version is too low for mafia it reports that and stops. Otherwise it will download the latest version of...