Is there a way to ignore non breaking spaces (   ) changes ?

[justme6666]

Hi guys,

I've been trying to ignore the non breaking spaces from my css selector results, but I am having trouble finding the right regex to ignore it. Or changedetection.io does not support this feature yet ? I did select "Ignore Whitespaces" from the Global configuration, but that doesn't seem to help, as I keep getting false positives.

I am using the CSS selector to check the changes for the "amount" field as shown in the snippet bellow:

Changedetection.io successfully checks the changes in the field "amount" , however the problem arises when changedection.io also checks for the &nbsp changes along with the "amount" field . How to ignore the &nbsp from the searches and only keep the changes in "amount" field ?

I attached a screenshot of the changes as shown in my notifications. As you can see, there are lots of notifications for changes in &nbsp even though the amount of "1" is the same, so this is causing me a lot of false positives, because I only wish to follow the changes in the "amount" field and ignore all non breaking spaces

Your help is greatly appreciated.

[dgtlmoon]

as shown in my notifications.

we are not mind readers, please be more specific, we are not sitting next to you

• what version exactly
• what notification backend exactly
• provide a snippet of the HTML that causes it (attach it as a text file)
• provide your CSS rule

[josxha]

This discussion seems to be stable so I post my data instad since I'm having the same issue.

• Tested on version v0.39.14
• Url to watch: http://asdfast.beobit.net/api/?type=word (Returns new text on every page visit)
• Notification backend: Telegram (using the global settings)
• Notification title:

{watch_title}

• Notification body:

{watch_url}
{diff_url}
{diff}

• Fetch method: Basic fast Plaintext/HTTP Client
• Global filters:
  • Ignore whitespace
• No additional settings set for the request

The content changes are visualized on the website as intended:

But this is the content of the message:

http://asdfast.beobit.net/api/?type=word
https://[my-domain]/diff/a9c5f242-f4ea-474e-9f6b-0596d654d35f
(changed)     "text": "Tur libero sdonec rproin hendre.",
(into   )     "text": "Tur laoree urient ndisse enimphas.",

I think the problem is that Telegram doesn't convert   to a space character.

[justme6666]

Yes, I was using Telegram as well in my report....

[VS-X]

This is an issue with Apprise: caronc/apprise#585. Apparently already fixed in master, so we're just waiting for the release.

Edit: Built the image myself, replacing the current Apprise version with git+https://github.com/caronc/apprise.git - can confirm, it works like a charm, no more   in Telegram notifications

[josxha]

Thanks a lot for the details and reference to the apprise issue.

[caronc]

All of these changes have been pushed to latest version (0.9.9) as of last night

[Pauld-1]

Apologies for replying to an old discussion but I'm having a similar problem with Discord notifications, here's an example

This page https://www.ntlite.com/changelog/

appears as

This is the notification body

{{watch_url}}

{{triggered_text}}
****
{{current_snapshot}}

with the notification format: HTML

I'll be grateful for any suggestions on how I could tidy up the notification please.

[raphaelp-github]

Same here!

[XBadFellaX]

@dgtlmoon

[meadim]

Facing the same using discord for webhook notifications

[caronc]

I can't personally provide much solace, as no changes to the Apprise formatting were made. But just from a glance, it appears the "Change Detection templates" are accepting markdown but then being converted to HTML before being passed into Apprise?

Apprise fully supports this, however, if you also tell Apprise that you're feeding it 'TEXT' (and provide it HTML) it will make sure that your &nbps; (and all other HTML characters) get passed along (literally)... you can also choose to tell Apprise you're feeding it Markdown, or feeding it HTML (maybe this is what you want?) and it will act acordingly as well (supporting both formats without problem). A 4th option is to just not tell Apprise what you're feeding it at all..... If you do this Apprise will not manipulate the data in any way; it just passes it 'as-is' (it assumes the sending systme knows what the format of the receving end is expecting). The 4th option may not be ideal though; especially if this is a program that should know in advance what it is feeding into Apprise.

Long story short; It appears to be in the initialization of the Apprise(body_format='text') . If this worked before, then (possibly) the logic in ChangeDetection should possiby shift to:

body_format = 'html' if using_custom_markdown_to_html_templating else 'text'
Apprise(body_format=body_format).

This is purely guessing though; i'm not quite sure how it was implimented. Just replying as i keep getting pinged on this ticket from 2022 🙂 ; Just offering my two-cents

[XBadFellaX]

Hey Guys, I created an issue for this as this is getting annoying. Can you guys chip in there so we can provide dgtlmoon a bit more information to track down the root cause? #3721

[dgtlmoon]

https://discord.com/developers/docs/reference#message-formatting discord only supports plaintext / basic markdown

Discord utilizes a subset of markdown for rendering message content on its clients, while also adding some custom functionality to enable things like mentioning users and channels. This functionality uses the following formats:

So therefor, the &nbsp is because you have HTML format set and there are white-spaces in the content layout that its trying to preserve.

  = "Non breaking white space" in HTML markup.