Awkward dependency of parts.js.rb and incomplete.json.rb

[Craig Russell <ap...@gmail.com>]

 When parts.js.rb is updated to add more reasons to reject a document, a corresponding change must be made to incomplete.json.rb. 

Is there an easy way for incomplete.json.rb to somehow pick up the changes, perhaps by redefining the structure of the reasons to include all of the parts: the name of the variable, the text to be displayed on the tool, and the text to be inserted into the email?

Just asking because I had this same issue last time I updated the checkboxes and didn't update the email text.

Thinking out loud, each reason would be e.g.
reason[1] = [@unsigned, ' unsigned', ' the document appears to be unsigned']
reason[2] = [@script_font, ' script font', ' a name typed in a script font is not a signature']

But maybe it would help if the parts.js.rb would calculate the reasons before invoking the incomplete message processing? In other words, just move the message calculation from incomplete.json.rb to parts.js.rb.

Craig L Russell
clr@apache.org

Re: Awkward dependency of parts.js.rb and incomplete.json.rb

[Craig Russell <ap...@gmail.com>]

Hi Sebb,

Thanks for taking the time to respond.

I guess it would be possible to compose the list of reasons on the front end and pass them to the back end to compose the email. There would be a bit more network traffic because the entire text of the reasons would have to be sent.

I admit I missed the comment at the beginning of the reject section which specifically refers to corresponding changes needed in incomplete.json.rb.

It is a bit awkward to stitch together changes from multiple files into one branch/PR but that's just me. 

All told it's probably not worth spending more energy on this.

Regards,
Craig

> On Jun 11, 2022, at 03:57, sebb <se...@gmail.com> wrote:
> 
> On Sat, 11 Jun 2022 at 00:15, Craig Russell <ap...@gmail.com> wrote:
>> 
>> When parts.js.rb is updated to add more reasons to reject a document, a corresponding change must be made to incomplete.json.rb.
>> 
>> Is there an easy way for incomplete.json.rb to somehow pick up the changes, perhaps by redefining the structure of the reasons to include all of the parts: the name of the variable, the text to be displayed on the tool, and the text to be inserted into the email?
>> 
> 
> Note that parts.js.rb is converted to Javascript, and runs in the
> browser, whereas incomplete.json.rb runs as a script on the server.
> 
>> Just asking because I had this same issue last time I updated the checkboxes and didn't update the email text.
>> 
>> Thinking out loud, each reason would be e.g.
>> reason[1] = [@unsigned, ' unsigned', ' the document appears to be unsigned']
>> reason[2] = [@script_font, ' script font', ' a name typed in a script font is not a signature']
>> 
>> But maybe it would help if the parts.js.rb would calculate the reasons before invoking the incomplete message processing? In other words, just move the message calculation from incomplete.json.rb to parts.js.rb.
> 
> Maybe; I'm not convinced that the effort is worthwhile compared with
> (say) ensuring that the code has comments reminding maintainers what
> to do.
> 
>> Craig L Russell
>> clr@apache.org
>> 

Craig L Russell
clr@apache.org

Re: Awkward dependency of parts.js.rb and incomplete.json.rb

[sebb <se...@gmail.com>]

On Sat, 11 Jun 2022 at 00:15, Craig Russell <ap...@gmail.com> wrote:
>
>  When parts.js.rb is updated to add more reasons to reject a document, a corresponding change must be made to incomplete.json.rb.
>
> Is there an easy way for incomplete.json.rb to somehow pick up the changes, perhaps by redefining the structure of the reasons to include all of the parts: the name of the variable, the text to be displayed on the tool, and the text to be inserted into the email?
>

Note that parts.js.rb is converted to Javascript, and runs in the
browser, whereas incomplete.json.rb runs as a script on the server.

> Just asking because I had this same issue last time I updated the checkboxes and didn't update the email text.
>
> Thinking out loud, each reason would be e.g.
> reason[1] = [@unsigned, ' unsigned', ' the document appears to be unsigned']
> reason[2] = [@script_font, ' script font', ' a name typed in a script font is not a signature']
>
> But maybe it would help if the parts.js.rb would calculate the reasons before invoking the incomplete message processing? In other words, just move the message calculation from incomplete.json.rb to parts.js.rb.

Maybe; I'm not convinced that the effort is worthwhile compared with
(say) ensuring that the code has comments reminding maintainers what
to do.

> Craig L Russell
> clr@apache.org
>