Clojure Q&A - Recent activity in Tools

Closed: does deps.edd support "exclusion" with wild cards ?

Mon, 23 Feb 2026 15:09:23 +0000

maven supports this:

<dependency>
      <groupId>org.apache.hadoop</groupId>
      <artifactId>hadoop-common</artifactId>
      <version>${hadoop.version}</version>
      <exclusions>
        <exclusion>
          <groupId>*</groupId>
          <artifactId>*</artifactId>
        </exclusion>
      </exclusions>

while Clojure does not , it seems

Closed: Improving curl options for CLI install scripts (brew-install)

Mon, 26 Jan 2026 23:23:15 +0000

Hi!

I'm the maintainer for Heroku's Clojure buildpack. It uses the install scripts from clojure/brew-install to install the CLI. I noticed intermittent install failures related to the download of the Clojure tools. I believe most of them could be avoided with a slightly changed curl command line options.

I opened a PR before realizing that PRs will not be accepted, so I'm starting this thread to properly get the discussion going. For reference, the closed PR can be found here: https://github.com/clojure/brew-install/pull/10

I believe these options would improve the download behaviour for all users, but especially when used in automations:

--connect-timeout 3 (https://curl.se/docs/manpage.html#--connect-timeout)
Prevents hanging when server is unreachable.

--fail (https://curl.se/docs/manpage.html#-f)
Exit with error code 22 for HTTP responses with status codes >= 400.

--max-time 60 (https://curl.se/docs/manpage.html#-m)
Maximum time in seconds for the entire transfer operation. Prevents hanging indefinitely with slow connections.

--no-progress-meter (https://curl.se/docs/manpage.html#--no-progress-meter)
Introduced in curl 7.67.0 (Nov 2019). Disables the progress bar for cleaner output. But less drastic than --silent and will still output diagnostic information (i.e. retries).

--retry-max-time 60 (https://curl.se/docs/manpage.html#--retry-max-time)
Retry only within this time period (60 seconds). After this time expires, no more retries will be attempted.

--retry 5 (https://curl.se/docs/manpage.html#--retry)
Introduced in curl 7.12.3 (Dec 2004). Retry up to 5 times on transient errors (timeouts, HTTP 408, 429, 500, 502, 503, 504, 522, 524).

--retry-connrefused (https://curl.se/docs/manpage.html#--retry-connrefused)
Introduced in curl 7.52.0 (Dec 2016). Consider connection refused as a transient error for --retry.

These are obviously opinionated and my ask is not for these options specifically. Happy to get anything that will improve the situation. :)

Manuel

Commented: proper XDG support

Tue, 30 Sep 2025 12:35:02 +0000

.clojure_history is not part of the Clojure CLI, I think that’s Leiningen?

Commented: Prepping - cache invalidation

Tue, 16 Sep 2025 13:35:03 +0000

Don't have anything planned currently.

Commented: how do you set -Djdk.attach.allowAttachSelf enabled so that I can interrupt a cider repl in emacs

Wed, 03 Sep 2025 06:45:25 +0000

You've accurately identified the timeout. CIDER (nrepl, in fact) first tries to interrupt the thread in a "legal" way, but if that doesn't succeed, it waits for 5 seconds and then kills the thread forcefully. It is not very "safe", might potentially corrupt memory or screw up locks, that's why it's a last resort, but it is totally fine to do that during development.

Welcome to Clojure! Hope you have fun.

Closed: Option in `clojure` CLI command to print the basis

Thu, 14 Aug 2025 21:37:20 +0000

It would be nice that clojure could print the basis for a given combination of aliases (or none).

A common use case is obtaining the computed :paths for a combination of aliases.

Answer selected: clojure -X:deps find-versions doesn't sort tags correctly?

Mon, 11 Aug 2025 01:59:31 +0000

git tags are sorted via:

git tag --sort=v:refname

which treats the tag name as a version. How suffix versions are sorted is dependent on your git config settings. It's possible to change this setting on your machine to consider suffixes before releases with:

git config --global versionsort.suffix -rc

It would be possible to resolve tags to commits and sort based on ancestor comparison but this would affect the performance (possibly a lot). Hard to say what the effect would be without some testing.

Answer selected: tools.deps.cli.api/mvn-install docstring typo

Mon, 11 Aug 2025 01:44:08 +0000

Yep, thx - fixed in source for next release.

POSIX-style command line arguments in tools (-T)

Thu, 03 Jul 2025 21:06:46 +0000

I have a couple command line tools written as libraries that I'd like to expose as proper Clojure Tools. They have classic POSIX-style command line interfaces, relying on tools.cli. I could add an -X kv-map interface, but there are a number of boolean flags that would not translate well to :foo true style (such as arbitrary file paths).

Is there a way to pass arbitrary arguments to a clojure Tool without using the kv-map style?

And if not, would there be interest in a flag or option that would allow for such a thing?

Commented: Improve clojure cli error msg: "The following libs must be prepared" could include "clojure -X:deps prep"

Mon, 09 Jun 2025 01:18:27 +0000

While I do have the basic prep command memorized, the variant that always takes me a bit of thinking is how to prep dependencies that come in via test dependencies (answer: add :aliases '[:test]' to the basic command).

Commented: How to uberjar without bundling source code?

Fri, 04 Apr 2025 07:26:05 +0000

I believe you need a bit more than just ommiting "src from the copy-dir call.
That would work if you don't use any dependencies of your own but most applications are assembled from a collection of other modules/libraries.
In that case, you want to specify the `:exclude` key in the parameter map passed to `b/uber`. It's a sequence of things to exclude where each item can also be a regex.

NOTE: If you are building your uberjar on macOS, I also suggest excluding `LICENSE` because it tends to cause issues with the case-insensitive file system.

Answered: a) how: programmatically set REPL ‘context’ ‘prj-1 aware’? b) how: programmatically switch REPL ‘context’ ‘prj-n aware’?

Sun, 23 Feb 2025 17:34:17 +0000

The implication seems to be to work on a different project one must (System/exit 0) VM and start another VM using clj

You don't have to exit the existing REPL process before starting a new one. You can have however many running clj instances, all inside different or the same projects.

If you want to switch to a different project from within the same window, be it a terminal or some IDE or something else, then it might also be possible but then it depends on what that "something" is. In a terminal, people usually use screen or tmux for something like that, or even Ctrl+Z with fg or bg if you're on Linux and maybe Mac or WSL.

This appears to be idiomatically (colloquially) referred to as ‘restarting the REPL’ I believe?

If it's all in the same process - yes. Although usually it means restarting a REPL within the same project, since nothing between projects conflicts with each other.
If there's a server process for the REPL and a client process, it might mean restarting both or restarting only the server if the client can reconnect without restarting.
But usually all the fine points are not important at all.

a required lisp feature, for me anyway: multiple simultaneous-concurrent (multi-threaded) REPLs, I refer to as ‘Clojure Listeners’ (‘in-process’ or ‘against the same runtime’)

I have never experienced or even heard of a need for that, so I'm curious what the actual use case is that isn't satisfied with multiple separate REPLs.

In any case, even that is possible if you start multiple socket REPLs.

non-socket non-nREPL non-LSP non-network non-middleware non-multi-VM based REPL approaches

But why all the limitations? Do they have any purpose?

SEE Multiple Clojure Listeners

I have no idea what I'm looking at, there's very little context on that image.

Could easily have multi-threaded socket-based listeners . . .

But you said "non-socket" above?..

Clojure Listeners: multi-threaded, full history, type ahead, line editing, completion, pop-up menus, colorized output, colorized-balanced-brackets-([{, etc. Also working on other listener features: multi-line editing, enhanced source lookup, doc lookup, repo-artifact search, embeddable .e.g. in morse, etc.

You're just describing an IDE, just with that small addition of "multi-threaded REPL" with a yet-unclear-to-me value add.

I want a fully dynamic long running VM under full clojure programmatic control.

That's a very, very vague statement, specifically the "fully dynamic" part.

I'm still not entirely sure what's going on here, but I feel that it might be useful to add that at least on JVM you'll have issues in your attempts to be "fully dynamic" in case it also means "changing dependencies within the same process". There are tools to add dependencies at run time, but you cannot remove them. You cannot have two of the same things as well. And you cannot update a dependency to a version that made a breaking change and expect things to continue working exactly as they did before.

With that being said an given the sheer length of your opening message and the length of the previous message, I'm absolutely certain that such a discussion is much better to be had via some more interactive medium, like Slack or Zulip. There's much less risk of getting stuck in the weeds only because you took a wrong turn, and it's much easier for both the writer and all the readers to unfold the idea piece by piece instead of having to deal with a truck load of vague info.

Comment edited: deps.edn support for :git/branch + :git/sha

Fri, 21 Feb 2025 00:39:57 +0000

From further discussions in the Clojurians Slack: Git deps as they exist today are more intended for publishing libs and tools, not for supporting organizations with large dependency graphs of private Git repos where you really want to track a main branch and not tags (e.g. service model spec repo → backend repo → infrastructure repo update propagation).

A workaround today would be to make all repositories have some shared structure under an imaginary "parent folder" and then use something like Git subtrees, Git submodules, or something else to assemble them all into a build directory. That way, all `deps.edn` files can use `:local/root "../other-dep"` to reference things.

Answered: a) how to start REPL with ‘context’ for ‘project-1 aware’? b) How to switch REPL ‘context’ to ‘project-n aware’?

Wed, 19 Feb 2025 21:15:06 +0000

If you are using the Clojure CLI, then running clj in a directory will start a repl that has "project context" to the project deps.edn in that directory, and that is the way that most Clojure developers (regardless of editor) create a connected repl. (Alternately, they may be using Leiningen with a project.clj, but same idea.)

Typically you do not need tools.build / build.clj at all to create and work in a project until the point that you want to build and release an artifact. (Note that because the Clojure CLI / deps.edn support projects in public git repos like github, you don't necessarily ever need an artifact at all.) So I would ignore this aspect at the beginning, unless it affects your end goals.

It's unclear to me what you need at this point that you don't have. Do you need help with getting a project REPL connected in Emacs?

Commented: Confusing error message on deps.edn wrong git/url dependency

Thu, 05 Dec 2024 20:21:49 +0000

This kind of thing is usually a security measure to prevent fishing around for private repos. Private repo https authentication has gotten a lot harder in the last few years with github (for probably good reasons). You can create and use a fine-grained personal access token here as your password now (https://github.com/settings/tokens?type=beta). But really, I'd recommend using ssh auth instead.

Answer selected: why does clj -X pass in options as a clojure.lang.Symbol?

Tue, 26 Nov 2024 01:04:24 +0000

-X will edn/read-string each argument

Answer selected: Is there a plan to have completion for "clojure" command?

Tue, 19 Nov 2024 08:31:46 +0000

Would love to have completion support but it may require some planning wrt performance as it will likely require edn parsing, which implies Clojure, which implies a JVM.

Commented: Add retries when downloading dependencies?

Tue, 08 Oct 2024 20:04:18 +0000

Yes, tools.deps.git just shells out to git so if this is a credentials or timeout issue, that's what you need to fix.

Answer edited: How can I make Calva use a relative path when loading a file in cljs?

Thu, 26 Sep 2024 05:44:47 +0000

Hello. I don't think there is a way to configure that currently. You are welcome to file an issue about it. We can probably make this work.

Meanwhile.

Is shadow-cljs or Figwheel managing the loading of the project? If so, saving the file should load it .

Also, I think you most often get away with selecting the whole file and evaluate selection. You can configure a keyboard shortcut like this to give yourself something quite similar to the command that doesn't work:

  {
    "command": "runCommands",
    "key": "ctrl+alt+c ctrl+alt+enter",
    "args": {
      "commands": [
        "editor.action.selectAll",
        "calva.evaluateSelection",
        "cursorUndo"
      ]
    }
  },

You can add a when clause to it to only operate on cljs files. I think it would be resourceExtName == .cljs. See https://code.visualstudio.com/api/references/when-clause-contexts

That said, it's probably fine to let it load Clojure files this way as well. You can copy the existing when clause from the command in the in Keyboard Shortcuts UI.

Commented: clojure is not starting on databricks file system (DBFS)

Sun, 22 Sep 2024 17:59:42 +0000

The one thing the CLI could do is to offer a way to set the place for project specific .cpcache which is **always** used if given, and not only as fallback when local dir is not writable. This would remove it's reliance on a working "-w" , which on some filesystems (at least on DBFS) is not working.

Closed: No way to specify Clojure version across all deps.edn files in a monorepo

Tue, 17 Sep 2024 02:47:03 +0000

Suppose you have deps.edn at the root and then a bunch of sub-directories with their own deps.edn files.

If you want to have a set of common dependencies for all of the projects, you can simply have e.g. a common directory where common/deps.edn specifies all of the common dependencies and all the rest of deps.edn files list common as one of their deps.

However, that doesn't work with org.clojure/clojure, making it implicitly special.

As I mentioned on Slack, it seems that clojure.tools.deps/root-deps brings the default top-level version of org.clojure/clojure and nothing else can override it but another top-level version.

Closed: Feature Request for Clojure CLI: Treat empty deps.edn the same way as a non-existing one

Tue, 17 Sep 2024 02:45:47 +0000

When clojure is called from a directory with an empty deps.edn it exits with a non-zero value. I wonder if it could be considered to treat that situation the same as when there is no deps.edn file at all?

This would be improve the Calva Ux a great deal. Calva defaults to start clojure-lsp at activation. One activation trigger is that a Clojure file is opened. Creating a deps.edn file in VS Code will create it as an empty file which the user then can edit. When clojure-lsp starts because of the file creation, clojure is called by clojure-lsp to figure out the classpath. If the file is still empty by then, the command fails, and no clojure-lsp for the user in that session.

It gets extra important for beginners with Clojure, who we want to instruct how to create a minimal deps.edn project and use a Clojure editor to connect to the REPL. As it is now the instructions get complicated.

Commented: Refresh/Reload deps.edn changes in Calva + VS Code in WSL Ubuntu

Tue, 20 Aug 2024 05:46:58 +0000

There is no command in Calva for reloading the project. With Clojure 1.12, add-libs can be used for loading a library into the project dynamically. Generally it is what you tried to do, restarting the REPL. But of course, the project files need to be fine . I see that you found the issue there.

Answered: Using log4j-layout-template-json with Leiningen

Thu, 13 Jun 2024 17:10:43 +0000

If you depend on multiple libraries that each have their own plugins cache (the .dat file), you need to merge those as part of your build process -- which is what the log4j2-plugins-cache plugin is for, with Leiningen.

In log4j 3.x, they plan to do away with this cache file and use something more amenable to build tooling that doesn't require custom merging like this.

I had to write something similar for tools.build / build.clj for use with the Clojure CLI: https://github.com/seancorfield/build-uber-log4j2-handler

Answer selected: install-latest does not use coords to find versions in self-hosted git

Thu, 06 Jun 2024 08:10:40 +0000

Logged as https://clojure.atlassian.net/browse/TDEPS-253

Answer selected: tools.tools install-latest throws sha error with :coord and never upgrades with :tool

Thu, 06 Jun 2024 08:10:20 +0000

Logged as https://clojure.atlassian.net/browse/TDEPS-265

Thanks!

Answer selected: Does tools.deps treat a version (e.g. 1.2.3) as a range (i.e. [1.2.3,))?

Wed, 05 Jun 2024 17:55:37 +0000

No, tools.deps versions specify an exact version only, no ranges, no soft versions:

$ clj -Sdeps '{:deps {org.clojure/data.json {:mvn/version "2"}}}' -Stree
Error building classpath. Could not find artifact org.clojure:data.json:jar:2 in central (https://repo1.maven.org/maven2/)
$ clj -Sdeps '{:deps {org.clojure/data.json {:mvn/version "2.5"}}}' -Stree
Error building classpath. Could not find artifact org.clojure:data.json:jar:2.5 in central (https://repo1.maven.org/maven2/)
$ clj -Sdeps '{:deps {org.clojure/data.json {:mvn/version "2.5.0"}}}' -Stree
org.clojure/clojure 1.11.3
  . org.clojure/spec.alpha 0.3.218
  . org.clojure/core.specs.alpha 0.2.62
org.clojure/data.json 2.5.0

But note of course that you only control the versions if you are the top project - transitive versions can always be overridden from above.

Retagged: Combining :exec-args from multiple aliases with user-defined semantics

Wed, 22 May 2024 19:40:15 +0000

I was looking for a way to invoke an nREPL server using clojure -X and specify extra dependencies and middleware in separate aliases which would be composed together on the command line.

For context: nREPL middleware are specified as a list of functions/runtime-resolved symbols, so ideally there would be some mechanism for each alias in the deps.edn to specify their own vector-of-middleware and have them concatenated in the order which the aliases were listed on the command line. (See below for concrete demonstration of this)

Unfortunately I couldn't find a satisfactory way of achieving this via the clojure CLI - the issue is that :exec-args maps from different aliases are simply merged, with latter values overriding earlier ones. The clojure.org reference mentions deep merging but this appears to be a doc error.

(Note that the widespread use of -M for this use case does not afford any way of composing command-line string args, forcing each alias to couple together all middleware and dependencies.)

Of course there shouldn't exist some magic set of semantics for merging exec-arg values (vectors => concat, maps => merge etc.)... Nevertheless there might be a usability gap to fill here, letting users specify their own merging semantics declaratively rather than having to resort to custom-coded task runners .

An example of how this would look like from the command line, given concat semantics for the :middleware arg:

$ clojure -X:nrepl
Starting nREPL server with: nil

$ clojure -X:nrepl :port 1234
Starting nREPL server with: {:port 1234}

$ clojure -X:cider:nrepl :port 5678
Starting nREPL server with: {:port 5678 :middleware [cider.nrepl/cider-middleware]}

$ clojure -X:portal:cider:nrepl
Starting nREPL server with: {:middleware [portal.nrepl/wrap-portal, cider.nrepl/cider-middleware]}

$ clojure -X:cider:cljr:portal:nrepl
Starting nREPL server with: {:middleware [cider.nrepl/cider-middleware, refactor-nrepl.middleware/wrap-refactor, portal.nrepl/wrap-portal]}

see this conversation in Slack thread for more context:
https://clojurians.slack.com/archives/C6QH853H8/p1711461045176759

Closed: clojure -X:deps aliases fails for alias-as-data

Wed, 01 May 2024 18:32:56 +0000

(~/clojure/t)-(!2011)-> cat > deps.edn
{:aliases {:foo ["some" "data"]}}

Tue Apr 30 22:33:28
(~/clojure/t)-(!2012)-> clj
Clojure 1.12.0-alpha11
user=>

Tue Apr 30 22:33:37
(~/clojure/t)-(!2013)-> clojure -X:deps aliases
Execution error (ClassCastException) at clojure.tools.deps.cli.api/aliases$fn$fn (api.clj:170).
class java.lang.String cannot be cast to class java.util.Map$Entry (java.lang.String and java.util.Map$Entry are in module java.base of loader 'bootstrap')

Full report at:
/tmp/clojure-4544942395618193305.edn

Answer selected: The clojure/clj man pages are missing the -Stree and -Spom options.

Mon, 29 Apr 2024 15:45:06 +0000

Thanks, I've updated that for next release. FYI, these options are being phased out in preference of -X:deps pom and -X:deps tree which is probably why I removed them (but I don't actually remember for sure).

Commented: Race condition in .cpcache directory creation

Sun, 25 Feb 2024 22:39:08 +0000

I investigated this a bit, and it seems to me the real problem here is that Java's mkdirs does not explain why the directory could not be made. Typically the way to overcome a race for a mkdir is to check for EEXIST and if that's the error, then continue with no error, because we know that the other process won the race.

java.nio.path.createDirectories will throw FileAlreadyExistsException when it loses the race.

I see that other code in that repo already uses java.nio, so I guess converting from mkdirs to createDirectories would be possible.

Hope this is helpful!

Answer selected: Add workspace level config

Thu, 15 Feb 2024 00:51:37 +0000

I'm going to redirect this to a dup - please vote/comment there

Closed: Typo in :clojure.tools.deps.specs/claspath-args

Tue, 16 Jan 2024 00:16:22 +0000

Hi Alex,

I have a project in which I run tests with instrumentation turned on. The project has a dependency on tools.deps and I have the clojure.tools.deps.specs namespace loaded.

When I run the tests in my project they fail with the following error:

Unable to resolve spec: :clojure.tools.deps.specs/claspath-args

It looks like there is a typo in a the usage of a spec here:

https://github.com/clojure/tools.deps/blob/master/src/main/clojure/clojure/tools/deps/specs.clj#L151

I think it should use the ::classpath-args spec, instead of ::claspath-args.

Could you please take a look at this?

Thanks, Roman.

Answered: Does the clojure cli have an option to include new extensions for dependency resolution?

Mon, 18 Dec 2023 23:09:14 +0000

Added clj-2820 to track this issue.

Commented: In tools.deps, get the source of dependencies without downloading

Tue, 12 Dec 2023 14:10:24 +0000

That would be useful to Nix + https://github.com/jlesquembre/clj-nix users too.

Commented: passing command line arguments to clojure via `cider-jack-in` (GNUEmacs 27.1, CIDER 1.2.0)

Wed, 06 Dec 2023 09:59:19 +0000

Cesar, you suggest using `(setq cider-clojure-clj-aliases ":dev")`
When i look in the docs i only see the option `cider-clojure-cli-aliases` (so cli instead of clj)

I think that's what we should use! Greetz

Commented: clojure cli bug on windows

Fri, 01 Dec 2023 20:29:36 +0000

Not really anything Clojure side can fix, quoting is environmental interpretation that affects what the CLI sees. I have filed an issue to extend the existing quoting docs with this info (https://github.com/clojure/clojure-site/issues/678).

Answer selected: Is tools.build considering adding a :license argument?

Tue, 10 Oct 2023 13:32:46 +0000

tools.build v0.9.6 now includes an additional :pom-data attribute that can be used to provide license (and any other pom data) in hiccup style to put in the generated pom.

Docs: https://clojure.github.io/tools.build/clojure.tools.build.api.html#var-write-pom

Example:

(b/write-pom
  ...
  :pom-data 
    [[:licenses
      [:license
        [:name "Eclipse Public License 1.0"]
        [:url "https://opensource.org/license/epl-1-0/";]
        [:distribution "repo"]]]])

Comment edited: Test suit doesn't find function implemented in my solution

Sun, 01 Oct 2023 14:25:42 +0000

This question relies on some info that's not stated. "Submitted" - where exactly? Also note that you have asked this in the `Portal` section, but I don't see how Portal is relevant to the question.

Answer selected: clj/clojure scripts throws a stacktrace with unknown command-line argument

Sat, 16 Sep 2023 00:05:24 +0000

Logged as https://clojure.atlassian.net/browse/TDEPS-254

Tagged literals in Clojure CLI arguments and deps.edn

Tue, 12 Sep 2023 21:02:35 +0000

As it stands there is no way for the Clojure CLI to extend the reader with tagged literals for use within arguments or deps.edn.

I propose that some mechanism for this should be added, which can enable additional tooling to use the intended use case of putting data within aliases in deps.edn where they otherwise would be restricted by data readers.

An example of a current project which would benefit from this is Codox, which has support for regular expression based filters on namespaces to be documented, however as the default EDN reader does not include support for regular expressions, this feature is unusable.

If tagged literals support was added to the Clojure CLI, this problem could be resolved by a #jvm/regex tag or similar being added for that usecase.

If this feature is added it should also be considered what happens when a deps.edn file or argument to the CLI includes a tagged literal which is not bound in the current invocation.

There are multiple ways that support could be added to the CLI, and below I present a few of my own thoughts on the matter, but this is not a request for this to be the way that the feature is implemented, just some ideas.

Tagged literals could be added via aliases in the user deps.edn which are then used by default in all projects
Tagged literals could be specified by an environment variable, which users can then set with direnv or other tooling

Answered: Accepting "-" as argument to read non edn data fails with `-X` cli invocation

Tue, 29 Aug 2023 13:44:21 +0000

Sorry, this usage is not going to work now as the clj tool itself uses - to accept args from stdin now as part of -X handling.

You could still use cat data.json | clojure -X:viscous '{:json-file -}' with the trailing map support.

Answer selected: clojure.tools.deps/calc-basis can fail with "Unable to resolve" due to a version range

Sat, 26 Aug 2023 16:13:49 +0000

This was caused because of the lack of :mvn/repos in my call.

Support for binary resources and large resources via deps cli.

Fri, 25 Aug 2023 00:58:32 +0000

I've written a few libraries that depend on binary resources (eg. native shared libraries) and other libraries that have very large resources (eg. the llm weights for llama).

In principle, both of these types of dependencies can be satisfied by the deps tools without any changes since the deps cli can fetch resources and add them to classpath.

However, there are few additional challenges which makes using the deps cli for large and/or binary resources awkward.

Large resources are Large

I've avoided using the deps machinery for large resources because there are several affordances missing:

a) no yes/no prompt before downloading dependencies.

It seems reasonable to download maybe 100's of megabytes of dependencies when invoking the cli, but it doesn't feel right for the cli to download gigabytes of data dependencies without asking first

b) no progress indicator when downloading dependencies

Downloading dependencies is usually pretty quick, but it's nice to have a progress indicator if several gigabytes of dependencies are being downloaded.

Native dependencies are often large

Native dependencies can range from a few megabytes all the way to a gigabyte (eg. Chromium embedded framework).

Native dependencies are platform/OS dependent

I think there's some maven magic that helps with this which I haven't had a chance to look into. Anyway, the artifacts required for running depend on the platform that the program is running on. Usually, it's ok to include the artifacts for all possible platforms, but that may be unnecessarily using resources and might not work in particular cases.

Shared libraries must be extracted from jars before loading

Many ffi libraries (like JNA) will do this for you. It's implemented in an adhoc way depending on the ffi library. It may or may not work depending on whether the shared library is standalone or depends on other shared libraries.

Potentially, the deps tools could help extract shared libraries using an acceptable, uniform method.

Precompiled linux shared libraries barely work

It's possible to precompile shared libraries for mac osx and windows with a wide range of compatibility that "just work". With linux, there are lots of caveats. Generally, it seems like you can get something that mostly works if you have a standalone shared library and compile it with zig (eg. llmdb), but compatibility goes way down if there are multiple dependent shared libraries that don't use zig (eg. graphviz).

Native dependencies should be compiled from source?

Some package managers ship precompiled binaries (eg. conda), but it seems like many package managers tend to compile from source (eg. pip, macports, homebrew, etc).

Compiling from source may help with some of the following:

compiling from source sidesteps many of the challenges of providing native dependencies on linux
native deps could be git deps!
Some devs really dislike precompiled binaries and strongly prefer compiling from source.

Obviously, supporting compiling from source has its own challenges.

The above covers many of the problems with large and/or binary resources. Here are some concrete use cases:

clj-cef: wraps the chromium embedded framework. The framework itself is about 0.5-1.5gb depending on the platform and includes both shared libraries and large resource files. There's also an additional small, shared library required.
clj-graphviz: wraps the graphviz c libraries. Extra tricky because there are multiple shared libraries and each shared libraries have additional dependencies (eg. libpng).
llama.clj: wraps the llama.cpp library for running llms locally. It would also be nice to add specific llm weights as dependencies, but they can be very large (eg. a few gigabytes up to hundreds of gigabytes). Further, llama.cpp can be compiled with gpu support, but I haven't figured out how to do in a portable way without a compile from source option.

I know there's a lot of overlap with existing tools, so maybe there's some way for the deps cli to integrate with those existing tools:

fetching native deps: pip, conda, homebrew, macports, apt, scoop, etc.
fetching data dependencies: hugging face, https://dvc.org/

Anyway, I've made multiple attempts at finding a good approach for including native and/or large dependencies and all of the options still feel pretty awkward.

Comment edited: Problem with leiningen and a local library

Tue, 22 Aug 2023 20:17:42 +0000

Thank you very much for your help.

The reason of the problems where indeed the underscores in the filenames. I've known, that in Clojure / Clojurescript the filenames must have underscores instead of dashes. But as an old Lisp programmer I'm always using dashes instead of underscores in filenames without thinking about it.
But however, I tested my code in html-lib-ui with defcard and tests without any problems. It seems to me, that Clojurescript has not as much problems with dashes in filenames than Clojure.
Are there any linters or packages for the Emacs to get a warning about dashes in filenames?

Unable to locate a Java Runtime

Sat, 29 Jul 2023 18:13:11 +0000

I try to learn Clojure and follow Clojure for the Brave and True book.

In core.clj file, I typed M-x cider-jack-in command. I see following error:

error in process sentinel: Could not start nREPL server: The operation couldn’t be completed. Unable to locate a Java Runtime.

I tried both 8 and 17 versions of Java, I verified them via which java and echo $JAVA_HOME

Lein commands works by themselves.

Can you help me?

Answered: Path to installed man pages by Clojure POSIX installer script

Tue, 25 Jul 2023 13:23:57 +0000

Added issue on the site repo here: https://github.com/clojure/clojure-site/issues/658

Answered: make-classpath2 in tools.deps cannot be run from a directory other than the project one

Sun, 16 Jul 2023 17:59:41 +0000

This should not be considered a public api - it changes regularly in breaking ways, and using it for any directory other than the current directory is not supported (this is only called from the CLI, which assumes current directory).

To programmatically make a classpath from deps.edn, you should use clojure.tools.deps/create-basis (which can specify a project deps.edn in any directory).

Answered: Feature request for Clojure CLI: Some way to specify OS-conditional values for :local/root in deps.edn

Wed, 12 Jul 2023 07:49:22 +0000

FWIW I would recommend just using relative paths for :local/root. Absolute paths are not necessary and just makes things complicated.

In my own setup I have all my projects in ~/code. So ~/code/project and ~/code/some-lib, in deps.edn I just use :local/root "../some-lib" from the project and everything works. Regardless of whether thats executing in WSL/Windows or whatever else.

I also have setups where everything is in the same ~/code/project and ~/code/project/packages/some-lib where I again just use :local/root "packages/some-lib" instead of the absolute path.

(Most of the projects actually still use project.clj, but the same method applies)

If you avoid absolute paths your entire problem goes away. ;)

Answered: Could the CLI check for the latest stable version?

Wed, 21 Jun 2023 16:14:02 +0000

Created ticket at https://clojure.atlassian.net/browse/TCLI-103