A bug in add-on validator

Anything to do with the Pale Moon add-ons website. (addons.palemoon.org)
Not for questions about add-ons themselves!
Forum rules
Important: This board is for specifics regarding the add-ons website (addons.palemoon.org) and not to report extension compatibility issues or discuss different extensions.
Please only post here when your topic is directly related to the add-ons website service so our moderators don't have to move your posts all the time...
yami_

A bug in add-on validator

Unread post by yami_ » 2018-11-11, 19:10

The add-on validator at http://addons.palemoon.org/special/validate/ seems to be only checking if some properties exists, rather than checking the value of each property (according to Parsed Install Manifest (JSON Encoded) section it is reading the values correctly). For example if add-on install manifest contains something like this:

Code: Select all

<em:bootstrap>false</em:bootstrap>
<em:hasEmbeddedWebExtension>false</em:hasEmbeddedWebExtension>
the All Checks (JSON Encoded) section will contain something like this:

Code: Select all

{
    "isBootstrap": true,
    [...]
    "hasEmbeddedWebExtension": true,
    [...]
}
In addition to that both a embedded WebExtension-related warning and a This add-on uses Bootstrap (Restartless) Technology information will be issued. This is of course not true.

New Tobin Paradigm

Re: A bug in add-on validator

Unread post by New Tobin Paradigm » 2018-11-11, 19:29

False is not actually valid. It should be omitted if not true which is WHY i only check for existance.. So whomever wrote the install.rdf is in error.

And seeing as Mozilla doesn't even have an rdf parser anymore let alone have all but completely removed XUL that means WE define the specification now and by that I deem it not valid thus your observations are RESOLVED WONTFIX.
Last edited by New Tobin Paradigm on 2018-11-11, 19:34, edited 2 times in total.

yami_

Re: A bug in add-on validator

Unread post by yami_ » 2018-11-11, 19:33

I guess this explains the validator behavior. Thanks for your explanation.

New Tobin Paradigm

Re: A bug in add-on validator

Unread post by New Tobin Paradigm » 2018-11-11, 19:35

I will however account for that in the new validation routine for submission.

Seems half the code in Phoebus 1.x is to deal with user error and I find that ridiculous. After nearly 15 years it is shocking that people still make these crazy mistakes in install.rdf
Last edited by New Tobin Paradigm on 2018-11-11, 19:38, edited 2 times in total.

yami_

Re: A bug in add-on validator

Unread post by yami_ » 2018-11-11, 19:52

New Tobin Paradigm wrote:I will however account for that in the new validation routine for submission.
I guess that the new validator version will suggest to remove such errors.
New Tobin Paradigm wrote:Seems half the code in Phoebus 1.x is to deal with user error and I find that ridiculous.
Yeah, this is ridiculous.

New Tobin Paradigm

Re: A bug in add-on validator

Unread post by New Tobin Paradigm » 2018-11-11, 21:02

The validator shall reject it.. Not suggest.

Locked