A) 1) Good news for your non mobile use of saving/"printing" as pdf file (chrome browser)
a) No distorsion at all, pdf printing perfect.
all the mini-boards, in all pages that contain them
lichess entry page; "parties en cours" tab, the tv, and "today's problem" (as displayed on that page).
profile page; lichess.org/@/username/playing, any sub-tab, some are long pages.
true for all margins settings, default or not.
These are the ones i tried, feel free to amend this statement, upon your experience or knowledge.
A) For the user: for the pages with mini-boards, no interactive boards.
b) Page-breaks aspect: (from big format, A1 Lp, to smaller, A3 Pt, hitting a first grid break-point-?)
Useful:
A1 Portrait: whole entry page fits in 1 page, page breaks may split mini-boards, for longer pages.
format seems in the outer grid range
Less useful (1), more page breaks as the image gets bigger for a given pdf reading zoom:
A1 Landscape: page breaks appear sooner, index page fill the vertical space, but horizontal wasted.
A2 Portrait: entry page has 1 break for 1 element, redundant with A1, but closer image (1)
A2 Landscape: also outer RWD rectangle zone, but with more page breaks, even for index page
A3 Landscape: also outer RWD rectangle zone, but with more page breaks, even for index page.
Interesting:
A3 Portrait: looks like an inner grid rectangle (2), narrower menu.
(1) Opinion: Bigger aspect at a given zoom level is not interesting when you can fit-to-width or -page in your snappy pdf viewer.
(2) Note: Grid in mind: from the biggest page format one can simulate all the smaller ones, in theory, by adjusting the margin settings. If attempting empirically to map the screen grid breakpoints (CSS pixels) to the pdf page format (distance or proportion?), the above, tells you to start above A3 landscape (in the width direction). A3 Lp might contains all cases. but not A3 Pt. See posts below in the dev questions.
how to indent?
A) For the user:
1) Good news for the non-mobile usage (last post was a set-up for this post)
c) Pages with distorsion/dislocation: (1 post) now 3
(i) Clarifying problem and extent (previous post)
(ii) Two Easy Temp-Fixes:
Important: Lichess Zoom level (slider) = 50 % here, for the whole thread. (see future post on zoom level variations adjustments or upon request if i forget).
Consider the two extremes in browser window size that the chrome allows, in terms of WIDTH (keeping height constant at max, on my 1920x1080 monitor, with bookmark bar present):
1) Full-Screen (F11) on maximum end
2) "Mobile-like" on the minimum width end.
1) go full screen (F11) before pdf print. looks like one end of the CSS grid, elements spread out.
Quick Fix: Pdf print preview window settings: A3, Landscape, default margins. A3, Lp
Pdf File: you get page breaks but the board and move list are not split, the minor elements can be. No distorsion.
2): go "mobile"(1): width to minimum before pdf print, other end of the CSS grid. one column disposition.
Quick Fix: A5 Portrait, default margins. A5, Pt.
Pdf File: you get the one column layout, as in the web view, page breaks may split elements anywhere though.
(1) Note: although tempting, don't try to resize the previewing window, as the css update (or full page?) will loose the alignment. but you can use the zoom there, to check the alignement on a few squares, but its is very small viewing area.
(iw) Non obvious temp-fixes (next post).
changing the order.
C) Questions for the Dev(s) or dev-like reader, stemming from this bottoms-up learning effort.
1) Narrowing down to the html piece position code: mini-board vs interactive board.
from the same game position, same piece.
A) mini-board: (pdf-printing friendly)
<piece class="black queen" style="left: 50%; top: 87.5%;"></piece>,
B) interactive-board: (pdf-printing distorsion, simple geometrical transformation family though)
<piece class="black queen" style="transform: translate(248px, 434px);"></piece>
Questions:
any idea how this pixel/proportion difference might explain the difference in pdf-printing treatment?
how does pdf-print interpret pixels in terms of distances. grid dimensions in terms of paper sizes?
it seems that pdf-print can handle relative or proportional dimensions, agreed?
i assume that translate( x px, y px) is with respect to an origin, top left case where the ghost rook lives...
correct me, if feasible. don't know where that's going. more interested in the image-board, size definition.
mini-board proportions:
is the image-board the reference for the proportions, or is it a common ancestor?
would make sense that the tables of mini-boards dictates dimensions to both image and pieces.
The table, then, is it defined in pixels or proportions?
if it is automatically filling some space allotted by the design grid, then we can perhaps conclude that pdf-print handles relative and proportional dimensions rather well? not?
Is there a blocking reason for using pixel dimensions in the interactive boards?
and not use the relative dimensions or positions, as in the mini-board?
how are the image-board dimensions defined in both types of boards?
any hints or pointers could help, no need to make it easy. i can search.
more under the hood questions to follow according to plan above. same order within.
This thread is the gift that keeps on giving, huh?
"how to indent?"
Oh, that's easy - Alt-F4.
can't do that.
I'm curious, are you forced to read all the posts? or are you having fun trolling here and there?
just go read elsewhere, i would have hoped that my above tweet threshold verbosity would have deterred you from coming back.
but you are the reader that keeps on coming back.... Alt-F4 to you too.
Some recent experiments, FullScreen web page to pdf-previewing. A3, Lp, lower margin varying from default, to a threshold, 2.04 (in). The quick fix mentioned above is exhibited, as well as another plateau without distorsion past the discontinuity.
near 4.00 (in). read the filenames.
https://ibb.co/DwCpQxC
https://ibb.co/Jzq9c7g
https://ibb.co/yXbmVYF
https://ibb.co/qmx4CrH
https://ibb.co/S5jwQSB
https://ibb.co/0XKwjDp
https://ibb.co/RpZT2cH
https://ibb.co/3zdFvtF
https://ibb.co/fDdkbkX
https://ibb.co/PGHVKYG
https://ibb.co/mGGY2B3
https://ibb.co/3c1Q5DK
https://ibb.co/44v814z
there are two chessboards, one comes from the static image board (e.g. wood3.png), the other one induced by the chess pieces positions. The results here show that we have expansions or dilations along the top-left to bottom-right diagonal, the top-left corner being a fixed point, that rook is always aligned.
the threshold 2.04 is a constant with respect to initial browser window width below that of fullscreen, this is not shown here.
Shown here is the continuous dependence of the distorsion below or above that value. (the distance between boards, jumps at margin 2.04).
There might be enough information here, to inverse problem which variable(s) are the most likely producing those variations, when lichess goes through the pdf-printing layer.
Not sure how I'm the troll here; I'm not the one spewing page after page of utter drivel. It's not so much a bug report as a cry for help, is it? Have you not seen a bug report before? Isn't this not better suited to a blog or set of Facebook posts? What sort of response are you after, exactly?
Brace yourself, or just read each paragraph one at a time, over a week.
it is a brainstorming journal, of my research attempts, that started with a practical problem. You are right, it has gone, into many directions, not very suited now, as a bug report. but it isn't anymore. I should extract something to present to the dev elsewhere, now that i looked at the source code, and am increasingly understanding the front-end, chessground and modern web dev. i will propose something later on, but in better format, not in this thread. need to see how much proof of concept i can do client-side (on my puter).
The bug stems from skia pdf module of chrome's interpretation of pixel as a length unit (what kind of DPI does it assume?) but lichess needs, for now, to keep it pixels for the 3D version of the board (3D pieces and apparent 3D board, z-index is to simulate occlusion through stacking, no 3D space computation). The 3D-board, is not a square, so the current solution involves using pixels for piece coordinates. The proportion coordinates (Rel is the source code prefix or suffix) don't work there.
i will try to understand why, on my own at some point, more precisely, because there may be other ways to place the pieces on a projected 3D-board -that's-not-a-square, ways that keep the coordinates dimensionless, as in FEN's, PGN's or the back-end code, or as in the mini-boards. not here, but for my scientific curiosity (i will put something elsewhere if i find solution).
To answer your question more closely.
i don't expect much of a response here and now, anymore, but it was a place to keep my progress (which may not be apparent to you). So avert your eyes, it sound painful. sorry. i can understand that it is.
so not really a bug report. you are right also, i never did any, formally, me be newbie, in all that team code dev thing, it is the first time i invest myself into a forum. Sorry it mutated into a blog, or scrapbook of a blog. moi pas branché.
Some questions, specially when left with no hint of answer, or guidance, from any knowledgeable reader, need many words to get a handle of. Even the sparse comments in the source code could have cut this thread in half. If you know that some source code in the know person read my posts, give them some of your irritation, if they are not overworked. I don't mind, really. i like finding out on my own, but i could have gone faster, and less annoying to forum readers.
i just hope that the pictures linked are understandable to other users, future users with pdf concern, desktop or mobile. I'd welcome your serious open-minded feedback on that specific question.
Or i should make a summary positive solution thread for the users (no code tweak or dev stuff), without the learning trajectory.
PS: lichess should get rid of the captcha, bots can't drivel as much as i do, can they?, but really, i get too distracted to answer them, and each time i try, i end-up adding a paragraph to my post. i'll try again... wish me luck. so black mate in one....
This topic has been archived and can no longer be replied to.