Self hosted, wrong (fixed) player JS location

[alperyazar]

Hi,

sphinxcontrib.asciinema/sphinxcontrib/asciinema/asciinema.py

Line 40 in f1882d4

load = """<script src="../_static/asciinema-player_3.6.3.js"></script>"""

I think, this commit assumes that the position of the page with asciinema directive is fixed to _static directory. This assumption makes many pages broken.

It looks like the previous version is fine:

sphinxcontrib.asciinema/sphinxcontrib/asciinema/asciinema.py

Line 25 in 75a7168

def visit_html(self, node):

.css file is included correctly thanks to:

sphinxcontrib.asciinema/sphinxcontrib/asciinema/asciinema.py

Line 12 in f1882d4

def copy_asset_files(app, exc):

Thanks!

[mike-matera]

I can confirm that 0.4.0 does not display local cast files. Reverting to 0.3.8 fixed the problem.

[divi255]

will look into this. please stay on 0.3 until fixed

[divi255]

fixed both local and remote. sorry guys, was too busy to check the player upgrade commit. it broke completely everything.

0.4.1 works well

[alperyazar]

@divi255 Thanks for the fix, BUT

I tested the new 0.4.1 version today but for my website the self-hosted player looks broken, no player even shows up. I don't know the exact problem and don't know to debug but reverting back to 0.3.8 solves temporarily.

I have noticed a difference between 0.3.8 and 0.4.1. Both versions embed the cast file in HTML with Base64 which may not be the optimum solution in terms of size but it's fine for me. However, 0.4.1 embed the same Base64 string 3 times, one of just for div id. This potentially increases (unnecessarily ?) the HTML file if the recording is a long file.

0.3.8:

sphinxcontrib.asciinema/sphinxcontrib/asciinema/asciinema.py

Line 29 in d3fb128

'src="data:application/json;base64,{src}"></asciinema-player>')

0.4.1:

sphinxcontrib.asciinema/sphinxcontrib/asciinema/asciinema.py

Line 40 in 714264e

template = """<div id="asciicast-{src}"></div>

[divi255]

thanks for reporting. can you please upload somewhere an example where the player is broken? It was tested in both self-hosted and embedded mode and I found no problems with that. Maybe I'm missing some variation.

[divi255]

about div-id, agree, it should be compacted

[divi255]

div-id fixed in 0.4.2