Fork me on GitHub
#sci
<
2022-11-08
>
mhuebert08:11:00

šŸ‘‹ good morning, looking into errors today. I noticed a couple of cases where stacktraces are missing or (maybe) missing information, not sure if they are bugs / expected / etc.

borkdude09:11:22

the throw case might be easily fixable - not sure about the anonymous function one, but could be worth investigating

borkdude09:11:51

can you make separate issues about each?

1
borkdude09:11:08

I can point you a bit where to look, if you're interested in looking into it

mhuebert10:11:37

I may have time - have to prioritize the final weeks of the Clojurists Together term, Iā€™d like to have some kind of deliverable although thereā€™s a huge surface area remaining šŸ˜…

Sam Ritchie17:11:31

Let me know if I can help you with any writing, demos, assignable stuff

1
borkdude10:11:08

no worries, if you can make the issue, then I will get to it eventually. this doesn't seem like a huge priority, but a nice to have, right?

mhuebert10:11:59

yeah, itā€™s not the biggest thing. One interesting thing Iā€™m thinking about is how I can resolve ā€œstack framesā€ back to individual blocks that have been evaluated. Right now everything will be <expr>:1:x or whatever, relative to the block it was evaluated in. Iā€™m not even sure how to phrase the question, but do you know if thereā€™s any way I can pass some context to eval that would be included in the call stack / error metadata somehow?

mhuebert10:11:03

or set the starting line/column

mhuebert10:11:22

eg ā€œeval this string, it is within a doc at position L3 C4ā€

borkdude10:11:29

There is :clojure.core/eval-file that you can set on the expression

borkdude10:11:58

I think clerk uses that too, SCI supports it as well

mhuebert10:11:12

is there a way to set that other than modifying the string?

borkdude10:11:55

you can also do this:

(sci/binding [sci/file ...] (eval-string* ...))

mhuebert10:11:05

iā€™m already using sci/eval-form maybe I can just do with-meta on the form