Hey,

really cool project. I especially like your simple and clean design. I am in the process of switching from MSWord to something closer to raw text and was wondering what your reasons are for using markdown instead of json... especially because you know about jsonresume, which I just found out about because of your project.

So... why? :D

In the function:

def title(md: str) -> str:
    for line in md.splitlines():
        if line[0] == "#":  # IndexError if line is empty
            return line.strip("#").strip()
    raise ValueError("Cannot find any lines that look like markdown headings")

If there is an empty line in md, that line is just an empty string in the list returned by md.splitlines(). Accessing the empty string gives the IndexError.

Checking truthfulness on the line before accessing line[0] should be enough:

        if line and line[0] == "#":  # Check that line is truthy first

May I do a Pull Request if the issue and the resolution proposed is satisfactory?

See https://docs.github.com/en/free-pro-team@latest/github/creating-cloning-and-archiving-repositories/creating-a-template-repository for instructions and https://upptime.js.org/docs/get-started/#create-a-repository-from-the-template for an example.

I want to be able to align the date to the right when I am using a bulleted list.

I have this:

**Formación Complementaria**

- **Título de socorrismo y priemeros auxilios, Cruz Roja** [40h] <span>2022<span>

The problem is that the 2022 is not right aligned in the pdf nor the html.

I want the same output that generates this:

### <span>Título de socorrismo y priemeros auxilios, Cruz Roja</span> <span>2022</span>

But preceded with a bullet point.

I don't know if this only happens when it runs as a Github Action, or if it happens on regular Windows, but this is a pretty scary message for Windows users to see.

The access denied error is due to https://bugs.python.org/issue26660 (fixed in 3.8, but the Github windows environment uses 3.7, and I would rather not require 3.8 for users).

I am not sure why the exception we catch is bubbling up as "Exception ignored".

Could not delete C:\Users\RUNNER~1\AppData\Local\Temp\resume.md_3lhv1lk9
[WinError 5] Access is denied: 'C:\\Users\\RUNNER~1\\AppData\\Local\\Temp\\resume.md_3lhv1lk9\\CrashpadMetrics-active.pma'
Exception ignored in: <finalize object at 0x263c82064d0; dead>
Traceback (most recent call last):
  File "C:\hostedtoolcache\windows\Python\3.7.9\x64\lib\weakref.py", line 572, in __call__
    return info.func(*info.args, **(info.kwargs or {}))
  File "C:\hostedtoolcache\windows\Python\3.7.9\x64\lib\tempfile.py", line 797, in _cleanup
    _shutil.rmtree(name)
  File "C:\hostedtoolcache\windows\Python\3.7.9\x64\lib\shutil.py", line 516, in rmtree
    return _rmtree_unsafe(path, onerror)
  File "C:\hostedtoolcache\windows\Python\3.7.9\x64\lib\shutil.py", line 400, in _rmtree_unsafe
    onerror(os.unlink, fullname, sys.exc_info())
  File "C:\hostedtoolcache\windows\Python\3.7.9\x64\lib\shutil.py", line 398, in _rmtree_unsafe
    os.unlink(fullname)
PermissionError: [WinError 5] Access is denied: 'C:\\Users\\RUNNER~1\\AppData\\Local\\Temp\\resume.md_3lhv1lk9\\CrashpadMetrics-active.pma'

Hi! Your repo was very helpful!

Would you like a PR adding a requirements.txt file?

Hi thanks for the awesome project.

I got the following warnings after running make resume

WARNING: Ignored `box-shadow: 2px 2px 4px #AAAAAA` at 14:5, unknown property.
WARNING: Ignored `box-shadow: none` at 79:9, unknown property.
WARNING: Expected a media type, got screen/**/and/**/(max-width: 800px)
WARNING: Invalid media type " screen and (max-width: 800px) " the whole @media rule was ignored at 95:1.

Looks like weasyprint does not support box-shadow.

The python script is able to detect my chrome install, but fails at making the pdf. This fails regardless of whether or not the pdf already exists:

Wrote resume.html
Found Chrome or Chromium at C:\Program Files (x86)\Google\Chrome\Application\chrome.exe
[0419/104241.939:ERROR:headless_shell.cc(595)] Writing to file resume.pdf was unsuccessful, could not open file: FILE_ERROR_ACCESS_DENIED

The program hangs at this point, and I need to interrupt it to gain control again.

Description

Render markdown tables. Should be able to add something along these lines and have it render.

| asdf | fdas |
| --- | --- |
| 123 | 321 |

Could be nice to have some nice CSS for the tables too.

Minimal Fix

Change

python -m markdown -x smarty resume.md >> $@

to

python -m markdown -x smarty -x tables resume.md >> $@

Reference

weasyprint's dependencies are non-trivial to install on Windows and macOS (especially Apple Silicon).

Chrome or chromium are much bigger dependencies, but much easier to install for most users (if not already installed).

This will generate resume.pdf from resume.html.

/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --headless --print-to-pdf="resume.pdf" --print-to-pdf-no-header resume.html

The difficulty is figuring out the path of Chrome on all platforms (and especially Windows).

python resume.py on macOS results in:

Wrote resume.html
Found Chrome or Chromium at /Applications/Google Chrome.app/Contents/MacOS/Google Chrome
[0316/071703.668323:WARNING:crash_report_exception_handler.cc(240)] UniversalExceptionRaise: (os/kern) failure (5)
Chrome died with <Signals.SIGABRT: 6> but you may find resume.pdf was created succesfully.

This appears to be harmless. resume.pdf is created just fine. But it's not a great UX.

[Feature Request]

Use case: when writing my resume, I want to make sure that my PDF output does not exceed one page. Currently, I have to update the .md or .css file and run the command every time and check the PDF output. A better UX would be to have a watcher that would run the .py script every time the .md or .css file changes.
It would be great if we did not have to run the command again and again when updating my resume.

I will get my hands dirty on this, next weekend. However, I can test this only on win32 and I will need support for testing on macOS and Linux.

The default CSS is pretty fragile because the HTML does not contain id or class attributes.

python-markdown has an attr_list extension that allows this. Using this would improve the CSS and make it easier for user's to customize.

Don't forget to update the README when this ticket is closed.

We're getting [0213/081040.860681:ERROR:sandbox_linux.cc(377)] InitializeSandbox() called with multiple threads in process gpu-process. on Linux (e.g.).

Ship the CSS file embedded in resume.py or in a data file documented in MANIFEST.in, and provide a command line option to point to custom CSS that should be embedded in the CSS file. (WIll need to verify CSS cascade behavior for this.)

(We don't want to go dependency crazy, but once this is done the additional burden of e.g. watchdog for #17 becomes small.)

Python-Markdown 3.4 has been released and makes the build fails.

It works well if I set the dependency back to version 3.3.7: pip install markdown==3.3.7.

While the requirements are not particularly onerous, it would be nice to be able to build with zero dependencies, just by pushing to Github. I've set up the machinery to do this on the workflow branch. It basically works fine, but the PDF output doesn't look great because Times New Roman is not installed in the container that does the build. I cannot figure out how to install it in a headless (i.e. non-interactive and license compliant way).

This issue can be resolved by either:

• Changing the default font to one that is installed or easy to install in the environment Github Workflows run
• Figuring out how to install TNR in that environment
• Deciding to live with the difference in output

Thoughts welcome!

Hey, apologies for the brief issue report. I'm using this to help with a school assignment and have to get back to writing it.

I tried with the latest Google Chrome on macOS (120.0.6099.62, installed with brew) , and with the Chromium that ships with the latest Debian stable (120.0.6099.71, tried via a Docker container), and it seems to print headers and footers on the output pdf.

I'm able to generate the expected pdf by pulling in a earlier debian container (buster) that ships with version 90 or so of Chromium, and I'm able to generate a correct pdf with that. I'm assuming it worked with 119, and broke with 120.

Thanks for this wonderful project! Let me know if I can help at all.

First of all, thanks for this repo! I really like this workflow. I did run into one issue when I inserted Font Awesome icons via a Font Awesome kit. The icons show up fine on the HTML version of the resume, but they do not show up in the PDF.

However, if I manually use my Chrome browser to print the HTML resume to PDF then the icons do show up in the saved PDF. This makes me think the base64 HTML encoding step might be causing an issue with the Font Awesome icons. Do you know if there is a different way I can encode the HTML content so that the Font Awesome icons will be preserved?

For reference below is the HTML content for a Font Awesome icon (you can set the kit to use either the web font or the SVG version).

Web font (default):

<i class="fa fa-envelope" aria-hidden="true"></i>

SVG:

<svg class="svg-inline--fa fa-envelope" aria-hidden="true" focusable="false" data-prefix="fas" data-icon="envelope" role="img" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" data-fa-i2svg=""><path fill="currentColor" d="M464 64C490.5 64 512 85.49 512 112C512 127.1 504.9 141.3 492.8 150.4L275.2 313.6C263.8 322.1 248.2 322.1 236.8 313.6L19.2 150.4C7.113 141.3 0 127.1 0 112C0 85.49 21.49 64 48 64H464zM217.6 339.2C240.4 356.3 271.6 356.3 294.4 339.2L512 176V384C512 419.3 483.3 448 448 448H64C28.65 448 0 419.3 0 384V176L217.6 339.2z"></path></svg>