Abel Sinkovics
writes: Yes, this is what I meant. My bad. However, I still think it is not a really user-friendly way of making that code available. I'd prefer a more "standard" style of tutorial, where you can just copy/paste code from the webpage as you go, and into a source file not Metashell. However, this is a non technical point and I'd rather not lose time over it since I'm satisfied with the overall tutorial. What I'm thinking of doing is adding "what we have so far" links to the beginning of the sections (similar to the "copy-paste friendly version"
Well, I use ad-hoc scripts too, but they are pretty cool because everything can be controlled from CMake and they generate Javascripts charts that can be embedded in the HTML documentation :-). Also, my benchmarks are run on Travis and pushed to a special branch on GitHub whenever I push to master, which is handy for keeping them up to date. I'll try to send you a pull request setting up the basic system, but that might have to wait until the end of my own review. Also, what will happen with the CMake-based build system in the new Metaparse repository? So if I'm getting it right, your benchmarks are re-generated after every
Hi Louis, On 2015-06-04 00:27, Louis Dionne wrote: links) that point to pages containing only the code (so it is easy to copy them). Note that everything you can copy to Metashell, you can copy to a "regular" source file as well. The only difference is that it is your job to display the type then. push to GitHub? Very nice :) CMake build system: the Metaparse repository uses Boost's Jamfiles (no CMake). Regards, Ábel