mermaid-cli 8.8.4 (new formula)

[BastianZim]

mermaid-cli 8.8.1 (new formula)

Added mermaid-cli

---

• Have you followed the guidelines for contributing?
• Have you checked that there aren't other open pull requests for the same formula update/change?
• Have you built your formula locally with brew install --build-from-source <formula>, where <formula> is the name of the formula you're submitting?
• Is your test running fine brew test <formula>, where <formula> is the name of the formula you're submitting?

---

I have not added a test yet as I do not know how to, unfortunately. mermaid-cli converts graphs written in markdown to images which is difficult to test for as a failed conversion will sometimes still result in an image.

What I would recommend is running mmdc -h which outputs

Usage: mmdc [options]

Options:
  -V, --version                                   output the version number
  -t, --theme [theme]                             Theme of the chart, could be default, forest, dark or neutral. Optional. Default: default (default: "default")
  -w, --width [width]                             Width of the page. Optional. Default: 800 (default: "800")
  -H, --height [height]                           Height of the page. Optional. Default: 600 (default: "600")
  -i, --input <input>                             Input mermaid file. Required.
  -o, --output [output]                           Output file. It should be either svg, png or pdf. Optional. Default: input + ".svg"
  -b, --backgroundColor [backgroundColor]         Background color. Example: transparent, red, '#F0F0F0'. Optional. Default: white
  -c, --configFile [configFile]                   JSON configuration file for mermaid. Optional
  -C, --cssFile [cssFile]                         CSS file for the page. Optional
  -s, --scale [scale]                             Puppeteer scale factor, default 1. Optional
  -f, --pdfFit                                    Scale PDF to fit chart
  -p --puppeteerConfigFile [puppeteerConfigFile]  JSON configuration file for puppeteer. Optional
  -h, --help                                      display help for command

but I'm not sure how to assert that a multi-line output is the same?

Another way could be to download a file as outlined in the cookbook but I would need to ask the maintainers to create and save such a file to the repo. Since I'm not the owner I cannot guarantee that they will allow this.

Edit: I'm actually not sure anymore if making sure that the image is correct, is needed. Since we're only testing if mermaid-cli works, we can establish that and any problem with the image should be taken up with mermaid-cli.
In that case, a test could be to run the following:

cat << EOF | mmdc
sequenceDiagram
    participant Alice
    participant Bob
    Alice->>John: Hello John, how are you?
    loop Healthcheck
        John->>John: Fight against hypochondria
    end
    Note right of John: Rational thoughts <br/>prevail!
    John-->>Alice: Great!
    John->>Bob: How about you?
    Bob-->>John: Jolly good!
EOF

(Not sure how to write that in ruby)

and then

assert_predicate testpath/"out.svg", :exist?

Any advice here would be appreciated.

---

• Does your build pass brew audit --strict <formula> (after doing brew install <formula>)?

Except for the test part it does.

---

Additional Comments:

1. Dependencies
   The project declares dependencies on their npm website which seem to be downloaded by brew when installing it (at least that's what happened to me).
   I don't know enough about dependencies though, unfortunately, to answer if they are needed, or not, so this is the list of dependencies I could find:
   1. https://www.npmjs.com/package/@mermaid-js/mermaid-cli
   2. https://github.com/mermaid-js/mermaid-cli/blob/master/yarn.lock
   3. https://github.com/mermaid-js/mermaid-cli/blob/master/package.json
      Any advice here would be appreciated.
2. The version on GitHub is 8.8.3 but the one on npm is 8.8.1. which I'm currently using. I have created an issue to check if npm is still being maintained. NPM version vs GitHub mermaid-js/mermaid-cli#78
   Edit: npm has been updated to 8.8.3-2 Should I update this in a new commit or create a new PR?
3. I'm not sure if a bottle is needed or not.

[chenrui333]

code style issue

==> brew style homebrew/core
==> FAILED
Formula/mermaid-cli.rb:17:3: C: test do should not be empty
  test do ...
  ^^^^^^^
Formula/mermaid-cli.rb:17:3: W: Empty block detected.
  test do ...
  ^^^^^^^

5358 files inspected, 2 offenses detected, 1 offense auto-correctable
Error: `test do` should not be empty
Error: Empty block detected.

[BastianZim]

code style issue

Yes that's the missing test. I left that in so that the test can be added to the right section.

[SMillerDev]

assert_predicate testpath/"out.svg", :exist?

This test seems like a good idea. And please also update the version as you indicated.

[BastianZim]

@SMillerDev Thanks for the feedback, will incorporate.

Would you mind quickly helping me with the test?

This is what I currently have:

test do
  
  shell_output("\
      cat << EOF | mmdc -o testpath/out.svg\
      sequenceDiagram\
          participant Alice\
          participant Bob\
          Alice->>John: Hello John, how are you?\
          loop Healthcheck\
              John->>John: Fight against hypochondria\
          end\
          Note right of John: Rational thoughts <br/>prevail!\
          John-->>Alice: Great!\
          John->>Bob: How about you?\
          Bob-->>John: Jolly good!\
      EOF\
    ")

    assert_predicate testpath/"out.svg", :exist?
  
  end

How do I send the variable testpath into shell_output so that the file is saved at the correct location?

Also, just for posterity, my two other questions again, if you wouldn't mind helping me here as well:

1. Dependencies
   The project declares dependencies on their npm website which seem to be downloaded by brew when installing it (at least that's what happened to me).
   I don't know enough about dependencies though, unfortunately, to answer if they are needed, or not, so this is the list of dependencies I could find:
   1. https://www.npmjs.com/package/@mermaid-js/mermaid-cli
   2. https://github.com/mermaid-js/mermaid-cli/blob/master/yarn.lock
   3. https://github.com/mermaid-js/mermaid-cli/blob/master/package.json
      When installing it, brew automatically installs the dependencies so not sure if they should be declared separately, or not. Any advice here would be appreciated.
2. I'm not sure if a bottle is needed or not.

Thanks!

[SMillerDev]

Just updating your code so I can use the github review tools is much easier. Please push your (non-functional) code and I'll review it.

[BastianZim]

Sorry about that, one sec.

[BastianZim]

I cannot create a commit for the bottles and dependencies questions since this is more about setup ut everything else should be up-to-date now.

[SMillerDev]

The project declares dependencies on their npm website which seem to be downloaded by brew when installing it (at least that's what happened to me).

For certain languages this is automatically downloaded. JavaScript is one of these so no need to declare them manually.

I'm not sure if a bottle is needed or not.

Unless you're downloading a single binary that can run anywhere (like a .jar file) you need bottles. You don't need to do anything for those though.

[BastianZim]

Ok, thanks for the feedback and help! It's my first formula so really appreciated!

[BastianZim]

@SMillerDev I'm still not able to find the file with assert_predicate(testpath/"out.svg", :exist?) failing every time.
Could it be something with mermaid itself where it's not creating the file? Is there a way to check that?

[SMillerDev]

Could it be something with mermaid itself where it's not creating the file? Is there a way to check that?

It could be, if you run brew test -d <formula> it'll run in debug mode and give you an option to drop into a shell to check when it fails. I think it's probably due to the issue with brew style.

[BastianZim]

Just ran it with --keep-tmp. The test.mmd file is correctly created but the out.svg not. Could it be because it was created with system and resides in a different location?

The path is correctly substituted as the command is displayed as mmdc -i /private/tmp/mermaid-cli-test-20201127-75742-171gs6j/test.mmd -o /private/tmp/mermaid-cli-test-20201127-75742-171gs6j/out.svg but the file doesn't show up anywhere.

[SMillerDev]

Could it be because it was created with system and resides in a different location?

System just means "execute this". If you run brew test -d <formula> and then run the command manually in the shell, does it work then?

[gromgit]

@BastianZim , you have an indentation-related syntax issue in your formula (https://github.com/Homebrew/homebrew-core/pull/65521/checks?check_run_id=1464363067). Run brew audit --strict --online --fix mermaid-cli to fix it locally, then try again.

If the Mermaid syntax depends on proper indentation, this may also be the source of your current headache.

[BastianZim]

If you run brew test -d <formula> and then run the command manually in the shell, does it work then?

Just to confirm, you mean:

1. Run brew test -d mermaid-cli
2. Get the following prompt:

==> mmdc -i /private/tmp/mermaid-cli-test-20201127-97029-g7kwsl/test.mmd -o /private/tmp/mermaid-cli-test-20201127-97029-g7kwsl/out.svg
/usr/local/Homebrew/Library/Homebrew/debrew.rb:18:in `raise'
Test::Unit::AssertionFailedError: <#<Pathname:/private/tmp/mermaid-cli-test-20201127-97029-g7kwsl/out.svg>>.exist? is true value expected but was
<false>
1. raise
2. ignore
3. backtrace
4. irb
5. shell

3. Run 5
4. Execute mmdc -i /private/tmp/mermaid-cli-test-20201127-97029-g7kwsl/test.mmd -o /private/tmp/mermaid-cli-test-20201127-97029-g7kwsl/out.svg again

Tried that and it doesn't work. Strange this is, when I then just run mmdc -i /private/tmp/mermaid-cli-test-20201127-97524-i2mfnv/test.mmd -o /private/tmp/mermaid-cli-test-20201127-97524-i2mfnv/out.svg on its own, it works completely fine.

[BastianZim]

you have an indentation-related syntax issue

Thanks for the heads-up @gromgit, should be fixed with c866b45

If the Mermaid syntax depends on proper indentation, this may also be the source of your current headache.

Unfortunately, mermaid is fine with the indentation. I ran the output file through it separately and it worked perfectly fine, just not in this context.

[BastianZim]

Just to sum it up:

If I run brew test --keep-tmp mermaid-cliI get:

==> Testing mermaid-cli
==> mmdc -i /private/tmp/mermaid-cli-test-20201127-98550-1clnvz3/test.mmd -o /private/tmp/mermaid-cli-test-20201127-98550-1clnvz3/out.svg
==> Kept temporary files
Temporary files retained at /private/tmp/mermaid-cli-test-20201127-98550-1clnvz3
Error: mermaid-cli: failed
An exception occurred within a child process:
  Test::Unit::AssertionFailedError: <#<Pathname:/private/tmp/mermaid-cli-test-20201127-98550-1clnvz3/out.svg>>.exist? is true value expected but was
<false>
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:55:in `block in assert_block'
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:1631:in `_wrap_assertion'
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:53:in `assert_block'
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:1392:in `block in assert_predicate'
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:1636:in `_wrap_assertion'
/Library/Ruby/Gems/2.6.0/gems/test-unit-3.2.9/lib/test/unit/assertions.rb:1383:in `assert_predicate'
/usr/local/Homebrew/Library/Taps/homebrew/homebrew-core/Formula/mermaid-cli.rb:33:in `block in <class:MermaidCli>'
/usr/local/Homebrew/Library/Homebrew/formula.rb:1884:in `block (3 levels) in run_test'
/usr/local/Homebrew/Library/Homebrew/utils.rb:500:in `with_env'
/usr/local/Homebrew/Library/Homebrew/formula.rb:1883:in `block (2 levels) in run_test'
/usr/local/Homebrew/Library/Homebrew/formula.rb:901:in `with_logging'
/usr/local/Homebrew/Library/Homebrew/formula.rb:1882:in `block in run_test'
/usr/local/Homebrew/Library/Homebrew/mktemp.rb:63:in `block in run'
/usr/local/Homebrew/Library/Homebrew/mktemp.rb:63:in `chdir'
/usr/local/Homebrew/Library/Homebrew/mktemp.rb:63:in `run'
/usr/local/Homebrew/Library/Homebrew/formula.rb:2129:in `mktemp'
/usr/local/Homebrew/Library/Homebrew/formula.rb:1876:in `run_test'
/usr/local/Homebrew/Library/Homebrew/test.rb:43:in `block in <main>'
/System/Library/Frameworks/Ruby.framework/Versions/2.6/usr/lib/ruby/2.6.0/timeout.rb:93:in `block in timeout'
/System/Library/Frameworks/Ruby.framework/Versions/2.6/usr/lib/ruby/2.6.0/timeout.rb:33:in `block in catch'
/System/Library/Frameworks/Ruby.framework/Versions/2.6/usr/lib/ruby/2.6.0/timeout.rb:33:in `catch'
/System/Library/Frameworks/Ruby.framework/Versions/2.6/usr/lib/ruby/2.6.0/timeout.rb:33:in `catch'
/System/Library/Frameworks/Ruby.framework/Versions/2.6/usr/lib/ruby/2.6.0/timeout.rb:108:in `timeout'
/usr/local/Homebrew/Library/Homebrew/test.rb:42:in `<main>'

because no file was created.

If I then only run mmdc -i /private/tmp/mermaid-cli-test-20201127-98550-1clnvz3/test.mmd -o /private/tmp/mermaid-cli-test-20201127-98550-1clnvz3/out.svg again, the file is correctly created.

[BastianZim]

To add: Running system "mmdc", "-i", "#{testpath}/test.mmd", "-o", "#{testpath}/out.svg" in a IRB works and successfully creates out.svg, the only time it doesn't is when it's in the script. There, the command is shown but no out.svg is created.

I also created a file containing only the command and that worked as well. Could it be that something in brew disables something in mermaid? Because, to be honest, I have absolutely no idea why it's just not doing anything, but only when it's called in the brew file, with everything else working completely fine.
The reason, why I'm thinking that it's mermaid, is that I also ran system "touch", "#{testpath}/test.txt" and successfully created a file.

Also, running

test = `mmdc -h`
puts test

correctly outputs

Usage: mmdc [options]

Options:
  -V, --version                                   output the version number
  -t, --theme [theme]                             Theme of the chart, could be default, forest, dark or neutral. Optional. Default: default (default: "default")
  -w, --width [width]                             Width of the page. Optional. Default: 800 (default: "800")
  -H, --height [height]                           Height of the page. Optional. Default: 600 (default: "600")
  -i, --input <input>                             Input mermaid file. Required.
  -o, --output [output]                           Output file. It should be either svg, png or pdf. Optional. Default: input + ".svg"
  -b, --backgroundColor [backgroundColor]         Background color. Example: transparent, red, '#F0F0F0'. Optional. Default: white
  -c, --configFile [configFile]                   JSON configuration file for mermaid. Optional
  -C, --cssFile [cssFile]                         CSS file for the page. Optional
  -s, --scale [scale]                             Puppeteer scale factor, default 1. Optional
  -f, --pdfFit                                    Scale PDF to fit chart
  -p --puppeteerConfigFile [puppeteerConfigFile]  JSON configuration file for puppeteer. Optional
  -h, --help                                      display help for command

so the tool itself works.

[gromgit]

Checked out and tried your PR on my own box, and I can confirm that something in the test environment is stopping mmdc from Doing The Right Thing.

Then I looked at the mmdc code and the installed artifacts, which suggests that it asynchronously launches a headless Chromium instance to do the rendering and output:

$ ls -lR /usr/local/opt/mermaid-cli/
...
/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer:
total 60
drwxr-xr-x 3 aho admin    96 Nov 28 10:12 .local-chromium/
-rw-r--r-- 1 aho admin  1132 Oct 26  1985 CHANGELOG.md
-rw-r--r-- 1 aho admin 11344 Oct 26  1985 LICENSE
-rw-r--r-- 1 aho admin 22441 Oct 26  1985 README.md
-rw-r--r-- 1 aho admin  1289 Oct 26  1985 cjs-entry-core.js
-rw-r--r-- 1 aho admin  1274 Oct 26  1985 cjs-entry.js
-rw-r--r-- 1 aho admin  2792 Oct 26  1985 install.js
drwxr-xr-x 4 aho admin   128 Nov 28 10:11 lib/
-rw-r--r-- 1 aho admin  4005 Oct 26  1985 package.json
-rw-r--r-- 1 aho admin  2053 Oct 26  1985 typescript-if-required.js

/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium:
total 0
drwxr-xr-x 3 aho admin 96 Nov 28 10:12 mac-818858/

/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858:
total 0
drwxr-xr-x 3 aho admin 96 Nov 28 10:12 chrome-mac/

/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858/chrome-mac:
total 0
drwxr-xr-x 3 aho admin 96 Nov 28 10:12 Chromium.app/

/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858/chrome-mac/Chromium.app:
total 0
drwxr-xr-x 7 aho admin 224 Nov 28 10:12 Contents/

/usr/local/opt/mermaid-cli/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858/chrome-mac/Chromium.app/Contents:
total 16
drwxr-xr-x  3 aho admin    96 Nov 28 10:12 Frameworks/
-rw-r--r--  1 aho admin 10387 Nov 28 10:12 Info.plist
drwxr-xr-x  3 aho admin    96 Nov 28 10:12 MacOS/
-rw-r--r--  1 aho admin     8 Nov 28 10:12 PkgInfo
drwxr-xr-x 59 aho admin  1888 Nov 28 10:12 Resources/
...

Sure enough, /var/log/system.log contains this error (reformatted for readability):

Nov 28 10:52:44 rover-air61 com.apple.xpc.launchd[1] 
  (com.apple.xpc.launchd.domain.pid.Chromium Helper.95300): 
    Path not allowed in target domain: 
      type = pid, 
      path = /usr/local/Cellar/mermaid-cli/8.8.3-2/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858/chrome-mac/Chromium.app/Contents/Frameworks/Chromium Framework.framework/Versions/88.0.4298.0/XPCServices/AlertNotificationService.xpc 
      error = 147: The specified service did not ship in the requestor's bundle, 
      origin = /usr/local/Cellar/mermaid-cli/8.8.3-2/libexec/lib/node_modules/@mermaid-js/mermaid-cli/node_modules/puppeteer/.local-chromium/mac-818858/chrome-mac/Chromium.app/Contents/Frameworks/Chromium Framework.framework/Versions/88.0.4298.0/Helpers/Chromium Helper (GPU).app

[BastianZim]

Awesome, that's it, reproduced it on my side. Thanks for investigating that!
Yes, Puppeteer launches Chromium to render the svg file and it seems like it doesn't like the path.

Is there any way to mitigate that? Otherwise, I think, I'm forced to use mmdc -h as a test as creating a file is the only thing mermaid does.

[SMillerDev]

Is there any way to mitigate that? Otherwise, I think, I'm forced to use mmdc -h as a test as creating a file is the only thing mermaid does.

I'd consider this a bug that upstream needs to fix. I wouldn't accept a help test just because the browser they ship with the CLI tool doesn't like sandboxes.

[BastianZim]

Found the problem: Puppeteer (and then Chrome) launches in its own sandbox which doesn't play nice with brew apparently. (Also see https://github.com/puppeteer/puppeteer/blob/main/docs/troubleshooting.md#setting-up-chrome-linux-sandbox)

When using --no-sandbox everything works.

[carlocab]

This LGTM. @BastianZim, can you rebase this against master (and squash your commits) so that CI can re-run it?

[BastianZim]

@carlocab Thanks for the review. I tried to follow the GitHub guide here but it pulled in all commits. Would you mind showing me how to do it correctly, or should I just open a new PR?

Sorry about the mess, it's my first time.

[carlocab]

Hmmm. The problem seems to be that you made your changes in the master branch, and didn't check out a different one to do it.

I think this might work (back up your formula first in case something goes awry):

git checkout origin/master
brew update
git checkout BastianZim/master
git rebase origin/master

Once you've done that, you'll want to do git rebase -i HEAD~N where N was the number of commits in your original PR. That'll allow you to squash them together.

[BastianZim]

@carlocab Thanks for the feedback! Yeah, master seems to be the problem as it's always pulling in all the commits when rebasing. I followed your advice but still have all the commits.

If it's alright with you, I would open a new PR, use a separate branch this time and ping you from there.
Sorry again!

[BastianZim]

Closing in favour of new PR: #67804