Closed FagnerMartinsBrack closed 8 years ago
Hey, I don't know how I missed this. Thanks for taking interest in the project. I'm using Zlib for compression. Relevant code is here:
https://github.com/stefanrush/weightof.it/blob/development/lib/scale.rb
I think the difference in size is due to differences in libraries and compression levels. I'm using the default compression level for Zlib (6/9), which was admittedly an oversight. What's interesting is Zlib seems to compress some of the larger libraries more than the tools you linked to. I would chalk this up to differences in compression algorithms between libraries. I may change the compression level to 9 and reweigh the database at some point.
Hi @stefanrush. It would be very useful to use some sort of "standard" GZIP algorithm that most tools also use, that would prevent this kind of problem.
To my knowledge zlib is the standard library for gzip compression in Ruby. Would you recommend a different library?
I don't write Ruby so I am not aware that zlib is the standard library for that language. What I mean is try to use an optimistic tool that compress better, maybe zlib 9/9 will do that, w/e.
Unfortunately I don't know a better tool other than gzip-js
that I am using on js-cookie....
Hi, very interesting project.
I see that js-cookie is listed as having a gzip size of 916 bytes at http://weightof.it/category/utilities.
In the build I am using grunt-compare-size to calculate the GZipped size and it prints the following for https://github.com/js-cookie/js-cookie/commit/d9318bb5994db77c809a6b26266b1ea84a42a40f:
As you see it says it has 856 bytes gzipped.
Using closure compiler it shows the correct gzipped size in the "original size" section:
refresh-sf.com also doesn't get even close to 900 bytes:
Ok, but did you reduced the size recently?
As a matter of fact we did, but even then it never overflown to 900+ byte size.
Why bother?
We try to be as lightweight as possible, it is even documented that the lib is ~800 bytes gzipped. In this case I would like to make sure that the calculation is being done correctly to prevent misleading documentation in the README.md :)
How is the gzip calculation being done to show those numbers in the site?