Oh, and .html/.xml files with DOCTYPE headers and the like confuse the
update as well. Ends up putting the copyright header in the middle of
the DOCTYPE in many cases.
On 8/20/10 5:48 PM, Justin Lee wrote:
> When running this against the grizzly tree, I've noticed a couple of
> hiccups:
>
> 1. js (and json) files aren't handled.
> 2. it complains about know binary files such as .jar, .jpg, .gif,
> .png, etc.
> 3. really doesn't like eclipse project files (probably should just rm
> them from the repo anyway but ...)
>
> Here's a filter I wrote while applying this so that I could filter out
> some noise:
> grep -v .bundle | \
> grep -v .class | \
> grep -v .gif | \
> grep -v .ico | \
> grep -v .jar | \
> grep -v .jks | \
> grep -v .jpg | \
> grep -v .ods | \
> grep -v .png | \
> grep -v .project | \
> grep -v LICENSE.txt | \
> grep -v dojo/ | \
> grep -v behaviour.js | \
> grep -v moo.fx | \
> grep -v prototype.js
>
> The dojo and .js files at the end are because we include some files
> from the dojo project and they don't necessarily need our copyright
> info. They're .js files on top of that so they won't get corrected
> anyway.
>
> These aren't show stoppers by any means but without filter these out,
> it makes the hudson copyright checking jobs very difficult to give
> meaningful errors when someone munges the copyright info.
>
> On 8/20/10 4:39 PM, Bill Shannon wrote:
>> n't have any copyright information. In cases where they do,
>> one of the challenges is that if a copyright comment is follo
>