Set auto as default @page size

[coorasse]

See:

• Print tweaks #25164
• Page overlaps header/footer when printing in chrome #27888
• https://stackoverflow.com/questions/49785039/printing-bootstrap-4-html-page-in-chrome-cuts-off-chrome-header-footers/50400421

I think any option would be better than "a3"

[MartijnCuppens]

I've changed the value to null, since this won't generate any additional css. Need an additional approval from @twbs/css-review to be sure.

[patrickhlauke]

does this cause any additional issues a la #26598, or does it solve them?

We'll need to investigate this more, this seems to break the Bootstrap site for example (macOS, Chrome)

[MartijnCuppens]

I did some research about this and it seems browsers behave pretty inconsistent and there are a lot of factors we can not control:

• We never know what the size of the paper will be, so whenever we try to fix this for one paper size with fixed widths, it will break for other cases.
• Media queries seem to behave very buggy for print. I've used this technique to determine the width of a page, but the results I got were pretty inaccurate:
  • IE/Edge: 648px
  • Chrome/ New Edge: 540px
  • Safari: The window width
  • Firefox: 793px
• Safari seems to print the page with the width of the window, while chrome has a width around ~1260px (A4 paper size), for Firefox it's pretty hard to determine, since it scales all content whenever its wider than, I don't know what happens with Edge or IE, I gave up after FF.
• Browsers rescale the page "when needed". This "when needed" is pretty opinionated by the browsers since every browser seems to do this in a different way.

I think we should just ditch the page size and $print-body-min-width. Yes, browsers will generate different print layouts, but that's better than the behaviour we have now where parts of the pages are cut of.

Maybe we should also mention the problems with print layouts in our docs.

[XhmikosR]

Just a reminder, we need to be extra careful with this since it's for v4. If we can't be 100% sure about any side effects, I say we leave it as is...

[patrickhlauke]

historically the print styles have been broken/problematic in many tiny aspects, so if this fixes at least some of the common situations even though it may not get to 100% perfection every time, it's probably an improvement though

[MartijnCuppens]

I agree with Patrick. #25164 broke more than it fixed I'm afraid. But this PR shouldn't be merged without removing (or null-ifying) the $print-body-min-width thing.

[MartijnCuppens]

@mdo, what do you think, should we remove the size and min-width or leave it as is?

[MartijnCuppens]

Ping @mdo

[XhmikosR]

Ping @mdo for a decision.

[XhmikosR]

I'm still not sure about this change @mdo. It can and will introduce change in behavior. See the above comments #28753 (comment)

[mdo]

Hmm, re-reading again, I could be convinced to not do this and just drop it entirely like we have in v5. The change could be seen as a breaking one for others.

Could be a breaking change

[MartijnCuppens]

Fine by me, close it than?