Fork me on GitHub

Just wondering if anyone’s figured out a way to get tools.deps to use a http proxy when downloading Maven deps, other than applying the patch in ? I’m just being lazy I guess, putting off learning how to patch tools.deps and install and use the patched version…

Alex Miller (Clojure team)16:11:27

I don’t think there is any way to do it but to fix that (and would be happy to see a patch for it)

Alex Miller (Clojure team)16:11:41

oh, there is a patch for it, sorry

Alex Miller (Clojure team)16:11:51

well, then you just have to wait :)

Alex Miller (Clojure team)16:11:03

I’m probably not going to have much time to look at tools.deps stuff until post-conj


totally understandable, no worries! Thanks for the clear and helpful response!


Is there any documentation on how tools.deps resolves deps? Specifically wondering what happens when two dependencies depend on 'A but in different versions

Alex Miller (Clojure team)17:11:16

the general idea is to always take the newest


Ok, that's great. Does that "general idea" imply any exceptions to that rule where the older artifact would be chosen?

Alex Miller (Clojure team)17:11:59

the complexities come when you consider the full graph

Alex Miller (Clojure team)17:11:22

A -> B1 -> C2 -> B2 -> C1

Alex Miller (Clojure team)17:11:52

do you get {A, B1, C2} or {A, B2, C1} or {A, B2, C2}


I'm glad I'm not the one who needs to figure that out 😉

Alex Miller (Clojure team)17:11:01

it’s actually NP-complete I believe

Alex Miller (Clojure team)17:11:14

fortunately most actual cases are pretty “easy”


...unless they involve jackson, I'm guessing

Alex Miller (Clojure team)18:11:06

yes, that is the exception :)

Alex Miller (Clojure team)18:11:26

and unfortunately, every dependency set includes at least 9 versions of jackson

picard-facepalm 8

A new complexity class: jackson-complete

😆 4

Although that might be equivalent to: undecidable 🙂



😂 4

Oh, that is good. You even Deutsch-ified the spelling of Jackson. Love the attention to detail there 🙂

Alex Miller (Clojure team)19:11:28

now that’s a t-shirt with sufficiently small audience


A sufficiently small audience to .... not make it?

Alex Miller (Clojure team)19:11:06

a joke t-shirt with an audience of 1 is not fun, because you can’t share it. And a joke t-shirt with an audience of large N, is cliche. So you have to find that sweet spot in the middle. :)

Alex Miller (Clojure team)19:11:53

this is bigger than 1, but not by much :)


Got it. It is pretty easily explained to anyone at the Conj, if they don't get it immediately (I wouldn't without this conversation)


As a reminder, if you think Rich would see a "channeling Rich sayings" T-shirt as a tribute (as it is intended), not a weirdness or offensive, I would be happy to bring some to the Conj 🙂


But also happy to never bring it up again if he doesn't want to see such things.

Alex Miller (Clojure team)19:11:59

for someone with such great care for words, I hate to ascribe any words to him


Understood. He isn't the only person like that, and such people should get to be upset with people that misquote them, especially if the meaning is changed. FYI, I do take pretty significant care in transcribing his talks. I think I've listened to the ones I have transcribed at least 4 times each now.


Maybe we should make some

(rich (comment forms))
T shirts? off-topic

Alex Miller (Clojure team)19:11:22

I appreciate the care in the transcripts - he’s never complained about those


Stu talked about Rich Comment Forms in his "Running With Scissors" talk.

Alex Miller (Clojure team)19:11:48

and I know he appreciates having slides+text together for context

Alex Miller (Clojure team)19:11:57

oh, I haven’t watched it yet :)


And observing it was a pun on the (comment ...) forms being "rich" with exploratory code and it also being a habit of Rich to use comment that way.

Alex Miller (Clojure team)19:11:08

clearly the circle is too small for a t-shirt :)

Alex Miller (Clojure team)19:11:49

fwiw, I made a custom “O(log32 n) is good enough” t-shirt for Conj 1 :)

😄 4

if you donated that design to clojurist together i bet they could raise some money for projects by selling the t shirt. i would for sure get one

Alex Miller (Clojure team)20:11:17

well feel free to make one yourself, mine was certainly nothing special and I don’t have it

andy.fingerhut20:11:31 makes it really easy to create T-shirt designs, I found out recently


I have never bought anything from them, so don't have knowledge of the full experience front to back.


i made some tshirts for my team at my last company through them. I still get emails to make mouse pads and other things with that design 2.5 years later


but the shirts were/are great




I don't think I've even seen a mause pad for at least fifteen years


I guess we'll have to live with these continuing to be the only Rich-approved fan club T-shirts:

Alex Miller (Clojure team)19:11:12

well, all the conference t-shirts too :)


I still love my conj 2014 t-shirt.


my clojure west 2016 and strangeloop 2016 are my two favorite shirts


Not many other conference t-shirts stick around for long.


And here I thought the biggest benefit of the conferences was listening to the talks, and the hallway track 🙂


No, listening to the talks is totally overrated, it’s only the hallway track 🙂


At least for me, since I can watch talks online, but seldom get to chat with others who understand what I’m talking about.


But clearly I have missed out on some great T-shirt buying opportunities that are not available elsewhere. This should be in the conference advertisements!