search

a-hurst / klibs

A simple Python framework for writing cognitive psychology experiments
7 stars 1 forks source link

Inserting blocks with no factor mask results in crash #4

Closed brettfeltmate closed 4 years ago

brettfeltmate commented 4 years ago

When no factor mask is provided, insert_block() sets factors to None and passes that to the trial_generator instead of the dict it was expecting.

a-hurst commented 4 years ago

Figuratively just encountered this myself. Should be fixed in the latest commit!

jmwmulle commented 4 years ago

Austin—

I'm slowly getting back into klibs, here; rewriting NumpySurface entirely at the moment (though preserving backwards compatibility for now).

I pushed a branch called "test" that repairs a bug (tested) in the EyeLink class; when you get a chance you should merge that in to the main branch as the bug is fatal and pretty easy to trigger. Maybe next month we get together to talk about how to collaborate, going forward? I still have never had to manage versions and branches and multiple contributers before, perhaps you can guide me into that world. :)

j

-- "[A]t every stage of history our concern must be to dismantle those forms of authority and oppression that survive from an era when they might have been justified in terms of the need for security or survival or economic development, but that now contribute to---rather than alleviate---material and cultural deficit." - Noam Chomsky

On Wed, Feb 12, 2020 at 2:53 PM a-hurst notifications@github.com wrote:

Figuratively just encountered this myself. Should be fixed in the latest commit!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/a-hurst/klibs/issues/4?email_source=notifications&email_token=ABMXDHFD5OUEW6NIHUY65A3RCRATZA5CNFSM4KT7ECNKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOELR6RKI#issuecomment-585361577, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABMXDHBWVUF3LWYV5AV3AJTRCRATZANCNFSM4KT7ECNA .

a-hurst commented 4 years ago

Hi Jon, your fix from the test branch has been added to the main branch, thanks! Looks like you caught a mistake in the PyLink API docs: they say the function returns None if no new samples, but looking at the source it's just a blind wrapper of a C function that returns 0 if no new samples.

Next month works for me, though I also have next week mostly off (reading week) if you have any time to spare: my schedule's very open.