search

lisagandy / synthesizer

Sythesizer Application available at http://cps-xena.cps.cmich.edu/uploader
1 stars 0 forks source link

Deploying with jake vs straight deployment #1

Closed jordangumm closed 10 years ago

jordangumm commented 10 years ago

Your instructions for deploying the app without building is confusing to me. Basically, I've never heard of deploying a straight Cappuccino app without building it to JS and subsequently pressing and flattening it to improve performance and overall application size. From my understanding Cappuccino is used mainly to ease development, not to deploy directly from. Anyway, when I host the deployed version the loader works (it doesn't with the straight Cappuccino version) but the table never does load. When I host the Cappuccino version the loader doesn't work but the table does display.

Do you have any suggestions?

jordangumm commented 10 years ago

Here are the different versions. The deployed is the compressed version created with Jake.

https://cps-xena.cps.cmich.edu/mergecolumns-deployed/ https://cps-xena.cps.cmich.edu/mergecolumns/

mikepj commented 10 years ago

I haven't gotten to the point of working on deploying and building the app with Jake. Right now, I've just been using the unoptimized version for the development cycle. After development is complete, I plan to work on the Jakefile to get everything compiled correctly so we can see that performance enhancement.

The straight Cappuccino version you linked to above isn't working because it doesn't have a column values CSV file to read from. You have to pass the filename in as an argument in the URI. In the AppController.j file, I set the default location of these values files to be in the web root under /input_data/. I'll add an input_data directory to the repository with a sample data file, but you will have to move it to the web root for it to work (or change AppController.j to look at /mergecolumns/input_data/). I'll post another message here once that's done.

mikepj commented 10 years ago

Alright, I added a new directory called input_data with a file in it named sample-values.csv. To get it to work, move or copy input_data to the web root, and then use the updated URL:

https://cps-xena.cps.cmich.edu/mergecolumns/#sample

mikepj commented 10 years ago

I spent some time learning the Cappuccino deployment process. There are 3 different states of a deployed build: a built state, a pressed state, and a flattened state. If you do a "jake deploy", I found that it will both build and press the code. Then you just have to flatten it to finish the process. Interestingly enough, the flattened state generates a larger archive than just the pressed state, but it launches in the browser a lot faster.

So here's what I did to deploy the app (after installing the Cappuccino tools available at https://github.com/cappuccino/cappuccino):

// cd to the root repository directory: mergecolumns jake deploy // Pressed build is placed in Build/Deployment/ColumnMerger flatten Build/Deployment/ColumnMerger Build/Deployment/ColumnMerger-flat // Flattened build now resides in Build/Deployment/ColumnMerger-flat.

Then copy the contents of ColumnMerger-flat to the web server. It should read from the sample data file when you use the same #sample URL format shown above.

jordangumm commented 10 years ago

Thanks for the quick response! I'll configure things as you said.