The content of the site is formatting the result on MG

https://metager.de/meta/meta.ger3?eingabe=repair+cafe+chat+irc&submit-query=&focus=web&s=&f=&ff=&ft=&m=

Chat Apropo Romania - 𝐕𝐞𝐜𝐡𝐢𝐮𝐥 𝐀𝐩𝐫𝐨𝐩𝐨 𝐂𝐇𝐀𝐓 𝐑𝐨𝐦𝐚𝐧𝐞𝐬𝐜

very interesting! This specific result however is not formatting the title to bold with html tags but with UTF-8 characters.
for example:
Chat Apropo Romania - 𝐕𝐞𝐜𝐡𝐢𝐮𝐥 𝐀𝐩𝐫𝐨𝐩𝐨 𝐂𝐇𝐀𝐓 𝐑𝐨𝐦𝐚𝐧𝐞𝐬𝐜

You can copy paste that text into any editor and it appears bold. Maybe we should limit valid utf-8 characters?

@phil what do you think?

Maybe we should limit valid utf-8 characters?

Do you mean transform or limit? The content could be viewed in normal characters. Maybe there is a library for doing that?!

Transforming those characters would be the best. But we would need a list of characters that we would want to transform back to "normal" letters. I found an answer on Stackoverflow that seems to be a good start.

Hm, afaik Unicode defines normalization functions for more or less exactly this purpose. I will look into whether we will have to use a special library for this.

Or we could just use the function from Stackoverflow.

After delving a bit into Unicode I think we should just use string replacement.

The list hasn't got:

• Unicode Characters in the Enclosed Alphanumerics Block https://www.fileformat.info/info/unicode/block/enclosed_alphanumerics/list.htm
• some chars of DingBats can be transformed too. https://www.fileformat.info/info/unicode/block/dingbats/list.htm

That might be unimportant, on the other hand could it be helpful to access more informations from sites with graphics and diagrams. If you tell me the data format, i can help to create the list.

Edit: Like the list from StackOverflow?

I'm not quite sure yet which characters we do want to transform. As those characters are not harmful despite the fact that specially bold characters might distract the user or catch the users attention more than those that have a normal font-weight. That's why I don't think we need to transform every circled number etc. but rather only those that look specially bold. Is that correct?

but rather only those that look specially bold. Is that correct?

It's a design decision. Maybe some users even prefer the original version. There could be a difference between the clean result page and a preview version for one snippet, that is shown by using a mouseover at a picture symbol.

Imho:

stay the same

• sans
• monospace
• fullwidth

should be converted

• *bold
• *bolditalic
• *italic at least in the header, maybe even in the snippet
• *script
• *fraktur

unsure

• doublestruck
• strikethroughSingle
• doublestruck
• capitalized
• serif
• circled
• parenthesized
• underlinedSingle
• underlinedDouble
• strikethroughSingle
• crosshatch

https://metager.de/meta/meta.ger3?eingabe=++++++Chat+Apropo+Romania+-+%F0%9D%90%95%F0%9D%90%9E%F0%9D%90%9C%F0%9D%90%A1%F0%9D%90%A2%F0%9D%90%AE%F0%9D%90%A5+%F0%9D%90%80%F0%9D%90%A9%F0%9D%90%AB%F0%9D%90%A8%F0%9D%90%A9%F0%9D%90%A8+%F0%9D%90%82%F0%9D%90%87%F0%9D%90%80%F0%9D%90%93+%F0%9D%90%91%F0%9D%90%A8%F0%9D%90%A6%F0%9D%90%9A%F0%9D%90%A7%F0%9D%90%9E%F0%9D%90%AC%F0%9D%90%9C&submit-query=&focus=web

added Fehler label

Is unicode allowed in the URL? Is it possible to fake URLs by "changing the font"? Maybe the URLs should stay the same?!

< ;) > If a snippet is more eye-catching by using bold, it could be shown smaller. Just count the weight between black and white points per char and add it for the whole snippet. e.g. used in ascii art: ### and ... </ ;) >

created merge request !1814 (merged) to address this issue

mentioned in merge request !1814 (merged)

I think we should just start by replacing characters in title and description as we have examples of abusage for those and keep an eye open if we find examples where it might be used in other places.

Reported results seem to look good with the fix applied:
http://2-review-1031-the-c-3ttjhv.review.metager.de/meta/meta.ger3?eingabe=Chat+Apropo+Romania+-+%F0%9D%90%95%F0%9D%90%9E%F0%9D%90%9C%F0%9D%90%A1%F0%9D%90%A2%F0%9D%90%AE%F0%9D%90%A5+%F0%9D%90%80%F0%9D%90%A9%F0%9D%90%AB%F0%9D%90%A8%F0%9D%90%A9%F0%9D%90%A8+%F0%9D%90%82%F0%9D%90%87%F0%9D%90%80%F0%9D%90%93+%F0%9D%90%91%F0%9D%90%A8%F0%9D%90%A6%F0%9D%90%9A%F0%9D%90%A7%F0%9D%90%9E%F0%9D%90%AC%F0%9D%90%9C&submit-query=&focus=web

Reported results seem to look good with the fix applied:
http://2-review-1031-the-c-3ttjhv.review.metager.de/meta/meta.ger3?eingabe=Chat+Apropo+Romania+-+%F0%9D%90%95%F0%9D%90%9E%F0%9D%90%9C%F0%9D%90%A1%F0%9D%90%A2%F0%9D%90%AE%F0%9D%90%A5+%F0%9D%90%80%F0%9D%90%A9%F0%9D%90%AB%F0%9D%90%A8%F0%9D%90%A9%F0%9D%90%A8+%F0%9D%90%82%F0%9D%90%87%F0%9D%90%80%F0%9D%90%93+%F0%9D%90%91%F0%9D%90%A8%F0%9D%90%A6%F0%9D%90%9A%F0%9D%90%A7%F0%9D%90%9E%F0%9D%90%AC%F0%9D%90%9C&submit-query=&focus=web

Looks good.

closed with merge request !1814 (merged)

mentioned in commit f0b6f677