From:  John Lenz <>
Date:  13 Jun 2014 23:05:47 Hong Kong Time

Re: Adding checksums to source maps


URN for this sounds pretty good to me but  if we wanted a "normalized line
ending md5" how do we defined it?

On Thu, Jun 12, 2014 at 3:46 PM, Brian Slesinsky 

> I think that sounds pretty reasonable. Agreed that we're not going to put
> UUID's in source files, so a checksum makes more sense.
> For matching the JavaScript to the sourcemap, I don't think we'd want to
> require an official UUID, but rather any unique identifier generated in a
> way that mismatches are unlikely. (For a build system that requires its
> output to be a deterministic function of its input, we would want to use a
> checksum of some sort.)
> Perhaps put the scheme in front:
> //# sourceMapId=sha1:3da541559918a808c2402bba5012f6c60b27661c
> //# sourceMapId=uuid:db99a770-f281-11e3-ac10-0800200c9a66
> This looks suspiciously like a URN so maybe we could just use that:
> On Thu, Jun 12, 2014 at 3:09 PM, Andy Sterland <
>> wrote:
>> I think the main reason I was leaning away from the embedded identifier
>> approach is that one of the scenarios where the feature needs to work is
>> when the generated JS file has no source map comments. That enables
>> developers debugging production sites that strip comments to use source
>> maps still. Take for example jQuery which now removes the sourceMappingURL
>> comment from the minified jQuery file.
> The jQuery file is here:
> They have a comment at the top, so it seems they're not opposed to all
> comments?
> In the release notes [1], they say: "If you want to use the map file for
> debugging the minified code, copy the minified file and add a //#
> sourceMappingURL comment to the end of the file."
> So they're aware of the issue and they could have kept the
> sourceMappingURL comment if they wanted to, but decided it was better to
> make the developer do it.
> We would have to ask them why they took it out, but I would guess that one
> issue is that it's unclear where the URL should point to avoid a broken
> link. (Absolute versus relative URL, and so on.) This wouldn't be an issue
> with a URN, so maybe they'd be fine with leaving it in?
> - Brian
> [1]