Rubber Duck Debugging

« first day (208 days earlier) ← previous day next day → last day (30 days later) »

Anonymous

00:40

@Downgoat I have a weird HTML/CSS problem at work and you're the master of both introducing and fixing weird HTML/CSS bugs, so when you have some time I have some questions to ask you

ASCII-only

@Mego wait he is? also what kind of weird problem is it

Anonymous

@ASCII-only Nearly-identical HTML and CSS, but one version has alignment issues for no perceivable reason

Anonymous

I can't share the code because NDA, but I can give some generic details

ASCII-only

@Mego can you explain the difference

or maybe simplify the problem to a small testcase on jsfiddle or something

Anonymous

@ASCII-only Slightly different data being inserted into a template using the same templating engine, slightly fewer template fields.

Anonymous

00:48

We send snail mail through Lob by sending an HTML file generated from a template using Mustache

ASCII-only

@Mego do they have exactly the same indentation/indentation in the same places? and are the node types the same (e.g. does one have a list, which the other one doesn't have)

Anonymous

@ASCII-only Would the indentation actually matter outside of the actual text?

ASCII-only

@Mego Depends

Anonymous

@ASCII-only ...Would having the text indented inside of an indented div do the trick?

Anonymous

The primary difference between the two HTML files is that mine is actually formatted nicely, whereas the other one looks like crap

Destructible Lemon

00:54

fricking frick

why are there so many rounding errors

ASCII-only

@Mego Hmm, not sure actually. Inside a div it shouldn't matter? If it's a text node e.g. p or span I think it matters?

Anonymous

← 13 messages moved from The Nineteenth Byte

Anonymous

← 1 message moved from The Nineteenth Byte

ASCII-only

Another example is on this website, .header-menu's children have no space nor newline between them - if they did, there would be a small gap between them

Anonymous

@ASCII-only Hmm... I don't have the code on-hand, but I can SSH into the server real quick

Anonymous

00:57

Or apparently I can't, because SSH access seems to be internal-only

ASCII-only

Hang on

Maybe I'm misunderstanding what you mean by alignment

Neither of them look like crap right? (as in there's no severe issues with element placement or anything?)

Anonymous

There are some slight alignment differences in the rendered output, but I was willing to chalk that up to the differences in the template data, and the fact that it's being rendered into a PDF on Lob's side

Anonymous

The original, working HTML template file looks like crap, but I can't believe that indenting would change it so much

ASCII-only

i meant the output only has minor alignment issues?

Anonymous

The problem is that part of the output that should be at the bottom of the first page is getting shoved onto the second page. There aren't any glaring alignment issues.

ASCII-only

01:03

wait. so the only issue is some vertical displacement?

Anonymous

Lob sticks some custom formatting in before (or after, not sure) rendering the PDF, and there's no way to see the modified HTML before it gets rendered (if they even modify the HTML and not the PDF)

Anonymous

@ASCII-only Yes, but it's a huge issue, because that bit that's being displaced onto the second page is a tear-off perforation

ASCII-only

@Mego That changes the problem though, the only way I can think of that would shift something down is extra newlines somewhere (maybe around the template fields), or extra elements with vertical padding/margin, e.g. <p><p> instead of <p>

Destructible Lemon

why can't the cylinders just work

Anonymous

@ASCII-only I could try getting rid of any blank lines around the elements. The template data won't be inserting any newlines - if there were newlines in the data, we'd have bigger issues elsewhere

ASCII-only

01:10

@Mego yeah, I meant there could possibly be extra newlines around the template fields (in the template) for formatting

Anonymous

Hmm... They are indented and pretty-formatted, so possibly

Destructible Lemon

these rounding errors have upset me

why

ETHproductions

that's why rounding errors exist of course, to bring programmers to tears

Destructible Lemon

01:26

ok so

now there aren't gaps

I suspect there are still rounding errors but i decide i don't care

rebel against rounding errors

ETHproductions

if it works fine on the surface then it must be alright

- the well of comfortable lies

Destructible Lemon

@ETHproductions like, if something should be not quite touching it will appear to be touching

but i guess i made choices like this when i decided on having aliased stuff

ETHproductions

oh ok

yeah, you can't expect a <1 pixel gap to show up if it's not anti-aliased

Destructible Lemon

it should be a 1 pixel gap though

especially since unzoomed, 1 pixel corresponds to 1 unit

well, without foreshortening anyway

now to implement outlines, i guess

except doing outlines normally doesn't work

they end up looking crappy

dangit pygame

ASCII-only

@DestructibleLemon if it corresponds to 1 unit then there shouldn't be rounding at all...

Destructible Lemon

01:34

@ASCII-only well i mean

without foreshortening

it's very confusing

why it can't work

ok so outline ellipses look like crap in pygame

so i need to make them myself

1 hour later…

ETHproductions

02:40

@Christopher If you hadn't solved your issue yet, I figured out the problem

In collideXr:

var x = round((trans[0]-width/2) / 50 + dx);

should be

var x = round((trans[0]+width/2) / 50 + dx);

it's actually been detecting the player hitting the wall at x=0 even though you're halfway across the board

Destructible Lemon

haha classic

Destructible Lemon

03:09

working fine

ETHproductions

yay

what's it supposed to be?

Destructible Lemon

protagonist with some other blocks for testing

... minor bugs remain

Downgoat

@Mego yes

Destructible Lemon

maybe i can...

make it so this angle never occurs

seems like a healthy approach to bugfixing

wait i just remembered i can't do that super effectively

oh wait yeah i can

Anonymous

03:45

@Downgoat Read through the transcript in for a brief summary

Destructible Lemon

04:26

it gets really laggy with a lot of cylinders uh oh

how could i fix this...

oof pygame is buggy

I could probably speed it up with culling and stuff

or i guess i could use open gl for the graphics if i wanted to give up

12 hours later…

Anonymous

16:31

Solved my html problem by changing my perspective. I was thinking that magic was causing my code to break. Turns out it was magic causing the other code to work.

« first day (208 days earlier) ← previous day next day → last day (30 days later) »

all rooms

Transcript for

Feb27

Feb '1828

Mar14

Rubber Duck Debugging

For people who find rubber duck debugging to be a very effecti...

ducks

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC