Fork me on GitHub
#leiningen
<
2017-12-28
>
danielcompton20:12:22

It’s very very unlikely we’re going to remove unofficial JARs from Clojars as that would break people builds. I think we could do a better job of highlighting shadowing, and possibly deprecating artifacts that are clearly unofficial

danielcompton20:12:58

But the golden rule of Clojars is “don’t break the build” so any solutions will need to keep that in mind

danielcompton20:12:42

I like the idea of being able to communicate a new canonical repo through tooling too

seancorfield20:12:35

I suspect that a visible "caution" on the http://clojars.org/postgresql/postgresql page that the group ID shadows Maven Central (perhaps with a link to http://search.maven.org with that group ID) would have been enough to prevent @jgh from hours of pain there @danielcompton

danielcompton20:12:34

Seems sensible. The set of repos is small enough to do it manually

noisesmith20:12:09

maybe lein ancient and lein deps :tree could be extended to recognize warnable artifacts, and add eg. colorized output or explicit warnings

danielcompton20:12:57

As some authors deliberately publish to both like cemerick and tcrawley

jgh20:12:42

Yeah truth be told it doesn’t need to be a whole lot, just a clue that what you think you’re doing may not be what’s actually happening.

rduplain21:12:54

Hi! Is there a doc for what the s3p:// URL is? For context: I'm using [s3-wagon-private "1.3.1"]. I just hit an issue where a project.clj had s3:// and not s3p://. Using s3p:// works. Curiously, however, I cannot find any docs on it, in projects, on AWS, or via Google.

danielcompton21:12:44

It’s a custom wagon protocol

danielcompton21:12:18

Plugins register support for different protocols. S3-wagon-private adds s3p

danielcompton21:12:32

There’s probably a few that register S3

rduplain21:12:57

Ah, gotcha. That makes sense. I've seen it in a few S3 plugin projects, so was curious whether there's a common meaning for it. It appears to mean "private S3 bucket." That it's a custom protocol, not AWS, clarifies.

jgh21:12:44

another victim to the postgresql clojars repo in #beginners … the body count continues to rise 😉

danielcompton21:12:08

@jgh can you post any sites that reference the Clojars version on your issue and I’ll try and contact them

jgh21:12:05

yea i’ll see if i can figure out where i came across that repo

danielcompton21:12:08

Yep, but keep in mind s3 is also not a built in protocol either

danielcompton21:12:00

They’re all extensions, the s3 one happens to look like an AWS s3 resource link but it’s not “official” in Maven

rduplain21:12:44

Got it. Thank you!

noisesmith21:12:58

regarding artifact discovery and making people use the right version, maybe it would help to have a usable version of lein search that doesn’t download the entire list of every version of every artifact available

noisesmith21:12:23

right now it’s kind of unwieldy so even people that discover it don’t find it especially usable

danielcompton22:12:17

@noisesmith I think lein 2.8.0 has that feature now

danielcompton22:12:42

It uses the api for search rather than the index

noisesmith22:12:59

haha wow haven’t tried this in a while

$ lein search postgres
Searching central ...
[net.piotrturski.testegration/postgres "0.0.2-beta"]
[net.piotrturski.integrationjtest/postgres "0.0.1-beta"]
[com.cohesiveintegrations.ddf/postgres "1.3.1"]
[net.sf.squirrel-sql.plugins/postgres "3.5.0"]
[org.cloudhoist/postgres "0.7.0-beta.2"]
[org.umlg/sqlg-postgres-parent "1.4.1"]
[com.github.cafapi/java-postgres-container "1.13.0-35"]
[com.cohesiveintegrations.ddf.postgres/distribution "1.3.0"]
[com.cohesiveintegrations.ddf.postgres/source "1.3.0"]
[com.cohesiveintegrations.ddf.postgres/postgres-app "1.3.1"]
Searching clojars ...
org.xml.sax.SAXParseException: The value of attribute "description" associated w
ith an element type "result" must not contain the '<' character.
 at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.createSAXParseEx
ception (ErrorHandlerWrapper.java:203)
    com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.fatalError (Erro
rHandlerWrapper.java:177)
    com.sun.org.apache.xerces.internal.impl.XMLErrorReporter.reportError (XMLErr
orReporter.java:400)
    com.sun.org.apache.xerces.internal.impl.XMLErrorReporter.reportError (XMLErr
orReporter.java:327)
    com.sun.org.apache.xerces.internal.impl.XMLScanner.reportFatalError (XMLScan
ner.java:1437)

noisesmith22:12:45

wonder if that’s a leiningen bug or a clojars bug?

noisesmith22:12:34

also that seems like a garbage result set from central?

noisesmith22:12:24

I asked about this in #leiningen on freenode, if I don’t get a good response I’ll open a ticket I think

noisesmith22:12:42

follow up: it’s a bug in cljars code

noisesmith23:12:32

follow up follow up: it’s fixed

noisesmith23:12:49

the quality of the results is still low though

noisesmith23:12:05

@jgh btw we are discussing these issues right now on #leiningen on freenode

noisesmith23:12:19

(as mentioned in topic message for this channel)

jgh23:12:42

ok, gmas and ill join