If you hash two files on different paths that have the same filename all references that get replaced will only use the same hash. For instance if you have MathJax in your path like:
vendor/js/MathJax/jax/input/AsciiMath/config.js
vendor/js/MathJax/jax/input/MathML/config.js
vendor/js/MathJax/jax/input/TeX/config.js
vendor/js/MathJax/jax/output/HTML-CSS/config.js
vendor/js/MathJax/jax/output/NativeMML/config.js
vendor/js/MathJax/jax/output/SVG/config.js
Then when you hash it you get something like:
vendor/js/MathJax/jax/output/SVG/config.02437fe5.js
vendor/js/MathJax/jax/output/HTML-CSS/config.5ada570d.js
vendor/js/MathJax/jax/output/NativeMML/config.53a6cc82.js
vendor/js/MathJax/jax/input/MathML/config.49dbdf64.js
vendor/js/MathJax/jax/input/TeX/config.5b950842.js
vendor/js/MathJax/jax/input/AsciiMath/config.20e7393e.js
So far so good, but when you look in a file that references one of those say vendor/js/MathJax/jax/input/MathML/config.49dbdf64.js for example, you'll find stuff like:
MathJax.InputJax.MathML.loadComplete("config.20e7393e.js");
Which should actually be:
MathJax.InputJax.MathML.loadComplete("config.49dbdf64.js");
Separating the processing into phases won't fix this, but separate tasks will. For a large project it can result in a lot of headaches.
If you hash two files on different paths that have the same filename all references that get replaced will only use the same hash. For instance if you have MathJax in your path like:
Then when you hash it you get something like:
So far so good, but when you look in a file that references one of those say
vendor/js/MathJax/jax/input/MathML/config.49dbdf64.jsfor example, you'll find stuff like:Which should actually be:
Separating the processing into phases won't fix this, but separate tasks will. For a large project it can result in a lot of headaches.