feat: add jbang jdk install <version> <path> #1252

nandorholozsnyak · 2022-02-23T20:13:51Z

Ability to link pre-installed JDK directory to JBang
Documentation update

- Ability to link pre-installed JDK directory to JBang - Documentation update

codecov · 2022-02-23T20:18:50Z

Codecov Report

Merging #1252 (2b0d8fd) into main (8e77cef) will increase coverage by 0.73%.
The diff coverage is 83.87%.

@@             Coverage Diff              @@
##               main    #1252      +/-   ##
============================================
+ Coverage     58.28%   59.01%   +0.73%     
- Complexity     1077     1139      +62     
============================================
  Files            86       96      +10     
  Lines          5818     5995     +177     
  Branches        996      996              
============================================
+ Hits           3391     3538     +147     
- Misses         1925     1950      +25     
- Partials        502      507       +5

Flag	Coverage Δ
Linux	`57.78% <83.87%> (+0.71%)`	⬆️
Windows	`57.83% <83.87%> (+0.76%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/main/java/dev/jbang/util/JavaUtil.java	`51.19% <0.00%> (-0.62%)`	⬇️
src/main/java/dev/jbang/cli/Jdk.java	`25.67% <33.33%> (+2.06%)`	⬆️
src/main/java/dev/jbang/net/JdkManager.java	`64.44% <96.15%> (+6.84%)`	⬆️
src/main/java/dev/jbang/cli/Build.java	`75.00% <0.00%> (-14.48%)`	⬇️
src/main/java/dev/jbang/source/ResourceRef.java	`45.28% <0.00%> (-5.67%)`	⬇️
src/main/java/dev/jbang/cli/Run.java	`67.56% <0.00%> (-2.97%)`	⬇️
src/main/java/dev/jbang/cli/Edit.java	`46.95% <0.00%> (-1.05%)`	⬇️
...c/main/java/dev/jbang/cli/DependencyInfoMixin.java	`100.00% <0.00%> (ø)`
src/main/java/dev/jbang/source/ScriptSource.java
... and 27 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8e77cef...2b0d8fd. Read the comment docs.

quintesse

Looks good in general! Good job!

quintesse · 2022-02-23T22:31:54Z

docs/modules/cli/pages/jbang-jdk-install.adoc

@@ -56,6 +56,9 @@ Installs a JDK.
 _version_::
  The version to install

+_[existingJdkPath]_::
+  Already existing JDK's path on the filesystem. If path is given, its version will be checked against the incoming one, if not the same, the process will not finish successfully.


A minor thing but I think it's best to rewrite this using a positive view. Eg. ".. and if the same, the given JDK will be added successfully". (The use of "added" is somewhat vague and ambiguous, perhaps there is a better word?)

added could be linked maybe?

"linked" is good 👍

quintesse · 2022-02-23T22:35:04Z

src/main/java/dev/jbang/net/JdkManager.java

+	 * @param path    path to the pre-installed JDK.
+	 * @param version requested version to link.
+	 */
+	public static void linkToExistingDirectory(String path, int version) {


Minor: I'd prefer "linkToExistingJdk"

quintesse · 2022-02-23T22:39:03Z

src/main/java/dev/jbang/net/JdkManager.java

+		if (Files.exists(jdkPath)) {
+			Util.verboseMsg("JBang managed JDK already exists, must be deleted to make sure linking works");
+			try {
+				FileUtils.deleteDirectory(jdkPath.toFile());


Don't use this, use our Util.deletePath(), that already handles any errors (or you can tell it to be quiet).

quintesse · 2022-02-23T22:40:21Z

src/main/java/dev/jbang/net/JdkManager.java

@@ -203,6 +207,40 @@ public static void uninstallJdk(int version) {
 		}
 	}

+	/**
+	 * Links JBang JDK folder to an already existing JDK path with a soft link. It


Don't specify "soft", because the linking code will actually create a hard link if it can't create a soft link for some reason.

src/main/java/dev/jbang/cli/Jdk.java

quintesse · 2022-02-23T22:49:35Z

src/main/java/dev/jbang/net/JdkManager.java

+			Util.infoMsg("JDK " + version + " has been linked to: " + linkedJdkPath);
+		} else {
+			throw new ExitException(EXIT_INVALID_INPUT, "Java version in given path: " + path
+					+ " is not having the requested version " + version + ", please provide a proper path");


say " is X, which does not match the requested version Y". Remove the "please provide..." part.

It may not be able to infer the X value, cause this path could point to anywhere, it can have no Java inside it, I can do an evaluation and if it does not work then fallback to the original message if it is okay.

quintesse · 2022-02-23T23:12:56Z

src/main/java/dev/jbang/util/JavaUtil.java

@@ -81,6 +81,7 @@ private static int determineJavaVersion(Path javaCmd) {
 		String output = Util.runCommand(javaCmd.toString(), "-version");
 		int version = parseJavaOutput(output);
 		if (version == 0) {
+			Util.verboseMsg("Version is 0, reading it from java.version property");


say something like "Version could not be determined from: '$javaCmd -version', trying 'java.version' property"

quintesse · 2022-02-23T23:14:07Z

src/main/java/dev/jbang/util/JavaUtil.java

@@ -104,7 +105,7 @@ public static Path getJdkHome() {
 	public static int parseJavaOutput(String output) {
 		if (output != null) {
 			Matcher m = javaVersionPattern.matcher(output);
-			if (m.find() && m.groupCount() == 2) {
+			if (m.find() && m.groupCount() == 1) {


That does indeed seem to have been wrong all this time, good catch!

quintesse · 2022-02-23T23:18:55Z

Oh and add "Fixes #1248" to the end of the commit message, then GitHub will automatically close the issue when your PR gets merged (you probably know that and just forgot 😄)

- Changes for PR - Fixes jbangdev#1248

nandorholozsnyak · 2022-02-24T19:29:54Z

Mentioned things are fixed, I'm coming soon with the tests.

- Tests for new feature

nandorholozsnyak · 2022-02-24T20:23:24Z

Coverage should be a bit better now.

- Change in test

nandorholozsnyak · 2022-02-24T20:45:53Z

Any cool idea how to fix this error happening with the Windows build?

TestJdk > testJdkInstallWithLinkingToExistingJdkPathWhenJBangManagedVersionDoesNotExist(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdk > testJdkInstallWithLinkingToExistingJdkPathWhenJBangManagedVersionExistsAndInstallIsForced(File) FAILED
    java.io.IOException at TempDirectory.java:262
^"C:\Users\RUNNER~1\AppData\Local\Temp\junit5745621985287379323\jbang\cache\jdks\11\bin\java.exe^" -Dfoo=bar ^"-Dbar=aap^ noot^ mies^" -classpath ^"D:\a\jbang\jbang\build\classes\java\test\itests\hellojar.jar^" helloworld
java -Dfoo=bar ^"-Dbar=aap^ noot^ mies^" -classpath ^"C:\Users\RUNNER~1\AppData\Local\Temp\junit8100379350[68](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:68)2[69](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:69)4745\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc03[71](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:71)000e3c8d32b2b143a430c20900b.jar^" ^"-XX:ArchiveClassesAtExit=C:\Users\RUNNER~1\AppData\Local\Temp\junit8100379350682694[74](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:74)5\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc0371000e3c8d32b2b143a430c20900b.jar.jsa^" helloworld

TestJdkManager > testResolveJavaVersionFromPathWhenJavaVersionHasNewVersioning(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdkManager > testResolveJavaVersionFromPathWhenJavaVersionHasOldVersioning(File) FAILED
    java.io.IOException at TempDirectory.java:262

- Extra use cases

quintesse · 2022-02-24T21:14:00Z

Any cool idea ...

Is that error for real? Because that path looks very wrong! There are URLs and what not inside of it.

nandorholozsnyak · 2022-02-24T22:15:58Z

Hm, on a Windows based laptop the build is fine. Any guesses to the TempDir?
Which path is not looking good @quintesse ?

quintesse · 2022-02-24T23:08:36Z

This one :-)

"C:\Users\RUNNER~1\AppData\Local\Temp\junit8100379350[68](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:68)2[69](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:69)4745\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc03[71](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:71)000e3c8d32b2b143a430c20900b.jar^" ^"-XX:ArchiveClassesAtExit=C:\Users\RUNNER~1\AppData\Local\Temp\junit8100379350682694[74](https://github.com/jbangdev/jbang/runs/5324909962?check_suite_focus=true#step:6:74)5\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc0371000e3c8d32b2b143a430c20900b.jar.jsa^"

nandorholozsnyak · 2022-03-02T20:45:21Z

Sorry I still had no time to finalize the wrong tests, I ran the tests on Linux and on Windows too. I think there were some reasons why this part was not having tests :D @quintesse do you have any tips to resolve the problem? What could be causing the problem on the CI?

quintesse · 2022-03-03T01:20:20Z

Well actually I think the weird path must have been a copy&paste issue on your part, because if I look at the logs it looks normal:

TestJdk > testJdkInstallWithLinkingToExistingJdkPathWhenJBangManagedVersionDoesNotExist(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdk > testJdkInstallWithLinkingToExistingJdkPathWithDifferentVersion(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdk > testJdkInstallWithLinkingToExistingJdkPathWhenJBangManagedVersionExistsAndInstallIsForced(File) FAILED
    java.io.IOException at TempDirectory.java:262
^"C:\Users\RUNNER~1\AppData\Local\Temp\junit4891103843310719245\jbang\cache\jdks\11\bin\java.exe^" -Dfoo=bar ^"-Dbar=aap^ noot^ mies^" -classpath ^"D:\a\jbang\jbang\build\classes\java\test\itests\hellojar.jar^" helloworld
java -Dfoo=bar ^"-Dbar=aap^ noot^ mies^" -classpath ^"C:\Users\RUNNER~1\AppData\Local\Temp\junit8179139871602089251\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc0371000e3c8d32b2b143a430c20900b.jar^" ^"-XX:ArchiveClassesAtExit=C:\Users\RUNNER~1\AppData\Local\Temp\junit8179139871602089251\jbang\cache\jars\helloworld.java.6c29e59df0efda6a7a9783b1de6b823dc0371000e3c8d32b2b143a430c20900b.jar.jsa^" helloworld

TestJdkManager > testResolveJavaVersionFromPathWhenReleaseFileHasNoProperJavaVersion(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdkManager > testResolveJavaVersionFromPathWhenJavaVersionHasNewVersioning(File) FAILED
    java.io.IOException at TempDirectory.java:262

TestJdkManager > testResolveJavaVersionFromPathWhenJavaVersionHasOldVersioning(File) FAILED
    java.io.IOException at TempDirectory.java:262

Edit: actually it's an issue with copy&pasting from the logs. Somewhere GitHub decided to insert links for things it thinks are commit ids. ANyway, I cleaned it up and the above logs are displayed correctly now.

Edit2: Haven't had time to look at why it's failing though.

- Disabling tests on Windows that are using symlinks

nandorholozsnyak · 2022-03-03T20:03:30Z

On windows the symlinks are a bit clunky but I think you are aware of that, so I used @DisabledOnOs(OS.WINDOWS) on those tests that are creating symlinks.
This solves all other failings, that are really suprising.
@quintesse do I still need something to fix on it?

maxandersen · 2022-03-05T23:07:58Z

I thought we do symbolic linking for jdk install testing elsewhere so what is special about this usecase?

nandorholozsnyak · 2022-03-06T12:40:14Z

I thought we do symbolic linking for jdk install testing elsewhere so what is special about this usecase?

I do not really see any tests that are installing the JDKs (of course there are but not in the TestJdk class). What I could basically debug is that the creation of hard links on Windows works only if you link one file, not a whole directory to another (like in the TestEdit test cases).
On my local Windows environment AccessDeniedExceptions are being thrown, but on the CI IOExceptions and not only for the now disabled 2 test cases but another 4. I was assuming that the TempDir annotation and its lifecycle is doing some magic behind it but as I read it's javadoc and it is clear that each method is having its own temporary directory, cause we have no BeforeAll only BeforeEach method in the BaseTest.
I assume that in the CI when the build is running and the linking is happening after the temporary directory can not "delete" itself, but as I do not see the full stack trace I can not tell if that is the case.

As I read, hard linking folder on Windows required admin privileges or different policies, so this current folder/directory linking might not be the best solution for this particular feature, but on Linux it works fine.

Should I come up with some other alternative ?

- Changes in test and reason for disabling test on Windows

maxandersen · 2022-03-06T12:56:32Z

i'll need to test on my windows setup but I thought we used linking for jdks - but if we dont we might need a "jdklocations.json" instead of 100% file path based solution.
@quintesse wdyt?

nandorholozsnyak · 2022-03-06T13:14:51Z

I was also thinking on a config file to locate JDKs, it should work everywhere. Let me know if I should rework it

maxandersen · 2022-03-06T13:44:29Z

I think that could be useful.

- CI config change to check what is the output of the tests, will get reverted

nandorholozsnyak · 2022-03-06T14:13:16Z

BTW, so if you are using a privileged powershell or command line then the whole feature works on Windows too. The symbolic linking on Windows has been already announced as it could only work with admin privileges only. Knowing this as an already announced limitation on Windows, should it be still reworked to the config file based JDK location?

I just changed the CI build config to see the more information in the logs, I will revert it as soon as I just understand the situation within the CI build.

- Retesting old version with custom JBANG_CACHE_IDR

maxandersen · 2022-03-06T14:48:12Z

symbolic link was supposed to only affect jbang edit; for jdk location that is used for running i think we need to find a way to support it across platforms.

nandorholozsnyak · 2022-03-06T14:50:31Z

Allright, I was not able to reproduce the error that was happening in multiple builds... I'm going to rework it with a config file.

maxandersen · 2022-03-06T15:02:21Z

a reason we didnt do config file before was that the jdk mechanism needs to work in the various bash/powershell/cmd script too - so we need to have that in mind how that will work.

nandorholozsnyak · 2022-03-06T15:17:48Z

@maxandersen
By the way, I just tested the jbang jdk default on Windows, and it is also not working, but it is stated in the documentations as well: https://www.jbang.dev/documentation/guide/latest/javaversions.html#managing-jdks

This config based change could be very "big", I would be totally fine to use admin privileges to set and install JDKs because of that limition, so what is going to be the final decision? Keep it as is, cause it has been stated earlier or should I do the refact on it? I have no Mac to test, but I can imagine it does not have this limitation only Windows.

maxandersen · 2022-03-06T15:22:26Z

hmm okey - if windows already has this issue then maybe better we add this now (with its known limitations) and look at config file based approach separately.

@quintesse wdyt ? you know the nastiness around these

- Solution using BaseTest prepared JBANG_CACHE_DIR - Removing --info from gradle build

quintesse · 2022-03-06T19:12:13Z

I'll take a look tomorrow. But a quick remark would be that we do indeed use symlinks on all platforms and that has always worked. Using any kind of "index" will make the whole "default" system not work. (You could never use a path like that for PATH or JAVA_HOME, you'd have to change them each time you switch the default JDK).

maxandersen · 2022-03-06T19:17:17Z

Yes, but there are two separate but related features here.

Jbang jdk install to point to a specific alternate java vm to use when resolving java version.

Jbang jdk default install location.

We can do the first via config file and second can keep using symlink and be limited

quintesse · 2022-03-07T10:42:51Z

Sure, we could, but we're already using the linking on Windows and we already accept that in certain cases the linking on Windows isn't ideal. So why introduce a new system just to work around one of those issues and not both? It seems to me that a better investment of time for the future would be to figure out how others are able to reliably make linking work on Windows (eg. Scoop).

quintesse · 2022-03-07T10:51:31Z

Btw, for me this PR works perfectly on Windows 👍 🙂

quintesse · 2022-03-07T10:55:16Z

src/main/java/dev/jbang/cli/Jdk.java

 			throws IOException {
 		if (force || !JdkManager.isInstalledJdk(version)) {
-			JdkManager.downloadAndInstallJdk(version);
+			if (StringUtils.isNotBlank(path)) {


Prefer the usage of https://github.com/jbangdev/jbang/blob/238ef90b0f65aba896b241b06bc8cde796364382/src/main/java/dev/jbang/util/Util.java#L1339_L1349 over the use of external libraries. (This might change in the future but at least that way it would be easy to find all occurrences)

quintesse

LGTM. The comments I left are mostly clean-up.

quintesse · 2022-03-07T10:57:40Z

src/test/java/dev/jbang/BaseTest.java

@@ -31,10 +31,11 @@
 	@BeforeEach
 	void initEnv(@TempDir Path tempPath) throws IOException {
 		jbangTempDir = Files.createDirectory(tempPath.resolve("jbang"));
+		jbangTempCacheDir = jbangTempDir.resolve("cache");


The changes in this file are not necessary (and unwanted really). See my comments on TestJdk.

quintesse · 2022-03-07T11:01:01Z

src/test/java/dev/jbang/cli/TestJdk.java

-
-		final File testCache = initJBangCacheDir();
-		final File jdkPath = new File(testCache, "jdks");
+		final File jdkPath = new File(jbangTempCacheDir.toFile(), "jdks");


Use JdkManager.getJdksPath()

Hehe, that is a private method. I can make it public to make it work.

Yeah, probably okay to do that 👍

quintesse · 2022-03-07T11:01:23Z

src/test/java/dev/jbang/cli/TestJdk.java

+	void testJdkInstallWithLinkingToExistingJdkPathWhenJBangManagedVersionDoesNotExist(@TempDir File javaDir)
+			throws IOException {
+		initMockJdkDir(javaDir);
+		final File jdkPath = new File(jbangTempCacheDir.toFile(), "jdks");


Same here, use JdkManager.getJdksPath()

quintesse · 2022-03-07T11:06:35Z

Personally I'd merge this PR as-is because it works great, even on Windows as long as the user has made the appropriate policy changes (which were already necessary on Windows). We can then think about how to improve the linking issue later.

nandorholozsnyak · 2022-03-07T11:43:52Z

I will fix the findings after work. Thank you.

- Removing StringUtils usage in Jdk - Rework test to use JdkManager's method

- Fixing failing test case

nandorholozsnyak · 2022-03-07T19:15:36Z

Everything should be fixed an the builds should be working.

quintesse · 2022-03-08T01:44:46Z

Good job, thanks @nandorholozsnyak !

maxandersen · 2022-03-10T00:11:48Z

@all-contributors please add @nandorholozsnyak for code

allcontributors · 2022-03-10T00:11:50Z

@maxandersen

@nandorholozsnyak already contributed before to code

feat: add jbang jdk install <version> <path>

5433f06

- Ability to link pre-installed JDK directory to JBang - Documentation update

nandorholozsnyak mentioned this pull request Feb 23, 2022

add jbang jdk install 17 <path> #1248

Closed

quintesse requested changes Feb 23, 2022

View reviewed changes

feat: add jbang jdk install <version> <path>

0ddb94a

- Changes for PR - Fixes jbangdev#1248

feat: add jbang jdk install <version> <path>

5f05ac8

- Tests for new feature

feat: add jbang jdk install <version> <path>

2dea823

- Change in test

feat: add jbang jdk install <version> <path>

68c9d7d

- Extra use cases

feat: add jbang jdk install <version> <path>

5b575e4

- Disabling tests on Windows that are using symlinks

nandorholozsnyak requested a review from quintesse March 5, 2022 20:07

feat: add jbang jdk install <version> <path>

3f55924

- Changes in test and reason for disabling test on Windows

feat: add jbang jdk install <version> <path>

7e51630

- CI config change to check what is the output of the tests, will get reverted

nandorholozsnyak added 2 commits March 6, 2022 15:25

feat: add jbang jdk install <version> <path>

0308b46

- Retesting old version with custom JBANG_CACHE_IDR

feat: add jbang jdk install <version> <path>

54ed3ce

- Retesting old version with custom JBANG_CACHE_IDR

feat: add jbang jdk install <version> <path>

15dcdd7

- Retesting old version with custom JBANG_CACHE_IDR

feat: add jbang jdk install <version> <path>

b0b6fe6

- Solution using BaseTest prepared JBANG_CACHE_DIR - Removing --info from gradle build

quintesse reviewed Mar 7, 2022

View reviewed changes

quintesse previously approved these changes Mar 7, 2022

View reviewed changes

feat: add jbang jdk install <version> <path>

adca10b

- Removing StringUtils usage in Jdk - Rework test to use JdkManager's method

nandorholozsnyak dismissed quintesse’s stale review via adca10b March 7, 2022 18:51

feat: add jbang jdk install <version> <path>

2b0d8fd

- Fixing failing test case

quintesse approved these changes Mar 8, 2022

View reviewed changes

quintesse merged commit c8bcd40 into jbangdev:main Mar 8, 2022

nandorholozsnyak deleted the feature/1248-install-jdk-with-path branch March 8, 2022 07:24

feat: add jbang jdk install <version> <path> #1252

feat: add jbang jdk install <version> <path> #1252

Conversation

nandorholozsnyak commented Feb 23, 2022

codecov bot commented Feb 23, 2022 • edited Loading

Codecov Report

quintesse left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nandorholozsnyak Feb 24, 2022 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

quintesse commented Feb 23, 2022

nandorholozsnyak commented Feb 24, 2022

nandorholozsnyak commented Feb 24, 2022

nandorholozsnyak commented Feb 24, 2022

quintesse commented Feb 24, 2022

nandorholozsnyak commented Feb 24, 2022

quintesse commented Feb 24, 2022

nandorholozsnyak commented Mar 2, 2022

quintesse commented Mar 3, 2022 • edited Loading

nandorholozsnyak commented Mar 3, 2022

maxandersen commented Mar 5, 2022

nandorholozsnyak commented Mar 6, 2022

maxandersen commented Mar 6, 2022

nandorholozsnyak commented Mar 6, 2022

maxandersen commented Mar 6, 2022

nandorholozsnyak commented Mar 6, 2022

maxandersen commented Mar 6, 2022

nandorholozsnyak commented Mar 6, 2022

maxandersen commented Mar 6, 2022

nandorholozsnyak commented Mar 6, 2022

maxandersen commented Mar 6, 2022

quintesse commented Mar 6, 2022

maxandersen commented Mar 6, 2022

quintesse commented Mar 7, 2022

quintesse commented Mar 7, 2022

Choose a reason for hiding this comment

quintesse left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

quintesse commented Mar 7, 2022 • edited Loading

nandorholozsnyak commented Mar 7, 2022

nandorholozsnyak commented Mar 7, 2022

quintesse commented Mar 8, 2022

maxandersen commented Mar 10, 2022

allcontributors bot commented Mar 10, 2022

codecov bot commented Feb 23, 2022 •

edited

Loading

nandorholozsnyak Feb 24, 2022 •

edited

Loading

quintesse commented Mar 3, 2022 •

edited

Loading

quintesse commented Mar 7, 2022 •

edited

Loading