Timings may be expressed as numbers, not just integers about har HOT 15 CLOSED

Yes, if you change the schemas, that will fix validation (which is what I assume you're after).

from har.

Actually, the schema is also found in the bundled har viewer (lib/har/viewer/scripts/preview/harSchema.js). That's probably where I got the schemas used in the Ruby code originally.

Perhaps it's better to get the latest viewer from http://code.google.com/p/harviewer/ and re-extract the schema from there - assuming the viewer has the correct schema of course.

from har.

I had just come to the same conclusion. :) I'll take a look today.

On Tuesday, January 21, 2014, Jari Bakken [email protected] wrote:

Actually, the schema is also found in the bundled har viewer
(lib/har/viewer/scripts/preview/harSchema.js). That's probably where I got
the schemas used in the Ruby code originally.

Perhaps it's better to get the latest viewer from
http://code.google.com/p/harviewer/ and re-extract the schema from there

• assuming the viewer has the correct schema of course.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-32870515
.

from har.

Did you convert the harviewer PHP to HTML by hand? It seems that there's next to no actual PHP in them, merely including the Google Analytics script, so it may be safe enough to simply treat them as HTML or strip out the PHP "tags". What are your thoughts?

from har.

I don't recall there being PHP in the viewer code, but if it's for GA, let's strip them out.

from har.

Sounds good.

from har.

The edge version of harviewer has the same issue, and I've opened a ticket there. I've got an as-yet-uncommitted thor task that grabs the most recent harviewer master (via Bower), strips the PHP and renames files to HTML, then copies the whole shebang to vendor/viewer.

Because this has uncertain benefit, and it changes the behavior of the gem by no longer including the harviewer dependency code, I'm not sure if a pull request would be welcome. I'll probably commit and push to my fork tomorrow if you'd like to take a look.

from har.

Sure, ping me when you've pushed (or just submit the PR) and I'll take a look.

from har.

Here's the branch with the changes - I don't want to do a pull request, because I don't think Bower is actually appropriate, upon reflection. https://github.com/michaek/har/tree/feature/bower_dependency

from har.

Gotcha. I like the idea of making it easy to update the viewer though.

from har.

I made a note in that branch's Readme about what I think a better design would be for allowing easier updates. It seems like the ideal way to handle it would be to simply change a gem depencency: gem 'harviewer_ruby', '2.1.5' The thor task and Bundler dependency would just be part of the build process for harviewer_ruby, and the har gem's signature shouldn't need to change.

from har.

I suppose it's more conventional to name it ruby_harviewer. :)

from har.

Yeah, that would work. It seems a bit overkill to me though, having two gems for such a small utility.

from har.

harviewer has just fixed this in 2.0.16 janodvarko/harviewer#7

from har.

I've just pushed a 0.1.0 gem with harviewer updated to 2.0.16. I'm still seeing problems with Chrome-generated hars though:

I've opened a Chromium bug for this.

from har.