Segmentation Fault
Analyzing the core
  • Home
  • About

Working on the tinderbox’n

Web Add comments

I’ve been writing an extension that uses part of Tinderbox’s (56K warning) json.js file. It’s an interesting experience, since I haven’t done much work with JSON before.

At over a meg, this json file takes quite a while to load. While parsing it and playing around with it for my own purposes, I noticed a few things that I would like to see :

  • A JSON formatter refuses to touch json.js because it is too big. So I had to do one of my own (need to upload it once I pretty it up).
  • JavaScript reportedly can load compressed javascript files. It would be mighty dandy for it to load compressed json (shrinking it down to a much smaller 84KB). Maybe it can! I have not been very successful
  • Tinderboxe’s JSON output isn’t real JSON, but that has been noted and filed in bugzilla. Hmm, I wondered why an error message was being written to my console ;)
  • I haven’t yet found a (simple) way to associate a check-in with a time/person, so I can’t “blame” a burning build on anyone. It’s got to the point where I was just about to comment asking them to reopen the bug, but loading in a new json.js file I noted some things that were not in the previous file. Mainly, the last json.js file I downloaded all had ‘undef’ in one section, and this one has a few names and id so I can sorta match when they checked in.
  • There are files littered in tinderbox to a bunch of this data that json.js is supposed to replace (See Tinderbox’s README file, Other Files section). When I just started using JSON, the almost CSV file was both direct to the point and pretty much what I wanted out of the JSON file anyways. But it was still missing some things, like who checked in, the log file, the stats. And another file sorta had that information. So it was spread out. I am really hoping that json.js consolidates and really fixes this problem. But at the same time, it is also fairly complex.

Anyways, it will be all fun and worth it when this is done. At least, I’ll be using it :)

Tags: json, mozilla, personal, seneca, tinderbox


May 6th, 2008 |

Tags: json, mozilla, personal, seneca, tinderbox


2 Responses to “Working on the tinderbox’n”

  1. Brian Crowder
    May 8th, 2008 at 9:36 pm

    The beautifier I link in the website URL might beautify your JSON well enough, give it a try.


  2. Robert Helmer
    May 8th, 2008 at 11:35 pm

    Yeah the JSON output didn’t come out exactly as I would’ve liked; what I ended up doing was dumping the internal object that Tinderbox server uses using a custom JSON converter (all the Perl modules I tried failed for different reasons!). The original plan was to release a saner subset in a cleaner format.

    gzipping the content does work well. My people.mozilla.org account is no longer active, but I had a bunch of examples of how to do parse this data, as well as a Javascript library to use it.

    The way I did gzipping was to have a Python app polling and downloading the json.js when it changed, and saving it as a compressed file.

    I’ll try to get access and put this stuff back up.


Leave a Reply

  • Pages

    • About
  • Recent Posts

    • addParser.py
    • A (Use) Case for self-signed certs
    • Now for something completely different
    • New Editor tool finally landed on AMO
    • Calvin and Hobbes
  • Tags

    editor extension fennec google chrome html5 intern json lazy mozilla personal prism python ria security seneca sleep svg tinderbox tip toronto Web wildon
  • Archives

    • November 2008
    • October 2008
    • September 2008
    • August 2008
    • July 2008
    • June 2008
    • May 2008
    • April 2008
  • Meta

    • Log in
    • Entries RSS
    • Comments RSS
    • WordPress.org
  • Blogroll

    • Andrew Smith
    • Anthony Hughes
    • Michael Mullin
  • Links

    • My Gallery
    • My Repository
RSS XHTML CSS Log in
Copyright © 2008 Segmentation Fault All Rights Reserved
Wp Theme by i Software Reviews
Proudly Powered by Wordpress