ah Wikipedia, you always find a way to suck

what did they do now?

they now popup a sort of dialog box to setup page appearance (font size, text width, etc)

very handy

...if only the settings sticked

even worse, the popup comes on every page you open, while the settings reset to defsault

--default

Developer console logs no script errors...

cookie?

dom storage?

bad coding?

tomman:

I liked Wikipedia how it was about 15 years ago

Basic HTML and it just worked

and every language version had the exact same layout

well Sompi_ want to create a clone of olderpedia with php?

i mean once you got the wikitext handled the hardest part is .. SECURITY

Sompi_: these days every language version have pretty much the exact same layout

there was a time not long ago where some languages did diverged

but I guess no more

nsITobin: Wikipedia uses both cookies and Local Storage, it seems

FWIW, I do not have an account there, I always read it as guest

Also, after checking with devtools, it seems the appearance control panel there (that pops up by default) doesn't save the settings at all

it does apply them successfully to the currently displayed page, but settings don't seem to be persistent

no, wait, a cookie changes

enwikimwclientpreferences:"vector-feature-custom-font-size-clientpref-0%2Cvector-feature-limited-width-clientpref-0%2Cvector-feature-appearance-pinned-clientpref-0"

enwikimwclientpreferences:"vector-feature-custom-font-size-clientpref-0%2Cvector-feature-limited-width-clientpref-0%2Cvector-feature-appearance-pinned-clientpref-0"

is there a difference?

hmm, the cookie flashes orange on devtools when changed, but it seems every time it gets saved... nothing changes at all

(I did changed text size between those two cookies, notice nothing really changed)

maybe the cookie being read is where the issue is

but it sets fine

perhaps a typo somewhere

yeah, no mater what I change, the correct event fires up, the cookie gets saved properly, but the actual cookie value is not set to the expected user-selected value...

maybe some subtle type difference or whatever?

what like oatmeal rasin?

nsITobin: Nowadays I would probably just write the backend in C... because if I write it in PHP, the same code would not work after ten years

I hate it when programming languages change so that the old code stops working

Sompi_: without updates likely not but I do have an old ass mozilla rdf parser from 2006 or something that I brought back to life

in php

few bc updates and a couple enhancements but otherwise same code

I enjoy php overall

My idealized webapp is a nice piece of php server software and only as much JS as is required for specific tasks.. like syntax highlighting or whatever.. with as reasonable a graceful fallback as possible on the js to no js and css featuresets

I guess I have just always been more of the server does the processing client does the presenting .. thing

including most html output

I just shut down an old IBM PC 300GL that was running non-stop for ten years

First 7 or 8 years as a server, then as an IRC shell

ha, a 300GL

haven't seen one of those since I graduated from college... in 2009

didn't liked that case design that much

fortunately it was succeeded with the much nicer looking and better serviceable Netvista/Thinkcentre

I just installed ST-DOS to a Thinkcentre

tomman: I've been seeing slight differences between enwiki and ptwiki, but it might be both different settings and some userContent.css rule

0731|01:51:25 < Sompi_> I hate it when programming languages change so that the old code stops working <-- you must love Python, then. :-P

tomman: there are apparently GET parameters to set the skin, you can always try that with wikipedia. I get no pop-ups, but besides being logged in with a skin set, there's also NoScript

njsg: on the rust compiler changing front. I con;t consider the crates part of the language. tey are currnt best practice of how to perform funcitons securely within the rust language

the crates are morle like the C runtime library then part of the actual language definition

changes are more like you need to use snprintf now instead of sprintf in you c code because sprintf has no way to prevent overflowing the buffer size

differnce is they did not remove the unsafe to use fuction from the c library so still really easy for users to generate unsafe code

What's crackin #SeaMonkey

hi nsITobin

What's new frg_Away

nsITobin nothing much. Hopefully more time in 2 weeks.

I am this ibb.co/NFJPqGd

push pop reorder change refresh

that was all one patch originally

I'll likely switch gears in a little bit and reset my env for 2.53 and work some php and setting my gogs instance back up

nsITobin yes mq is great for this.

any good hg forges out there lol

See if I were Mozilla .. and I might be.. tho many will contest that especially Mozilla.. I would maintain hg internally and use git cinnabar for the public git version accepting PRs but not actually merging them.. a hook would take the PR and reformat it for hg and add it to a patchqueue for testing that way you get the ease of git for little more overhead than you already have for an established and well oiled hg scheme.. even if half of it is

objectively dumb cause lolmodernzilla

I haven't managed for form any opinions about hg yet. I have very strong positive feelings for git, but that may be because I'm comparing it to cp -r verison control and/or CVS (and frankly between the two of those, I prefer the former).

*to form

I suppose if I ever am involved in a dev project that uses hg, I'll form opinions one way or the other about it.

I've heard good things about DARCS but haven't had occasion to work with it either.

My biggest complaint about git is that it's rather more line-based than I would prefer, same as tools like diff and patch.

maybe we need a new vcs

I have no idea whether there's an existing one that moves away from being line based. For me that issue isn't serious enough to motivate me to look into alternatives.

not sure there is much more I can really do on central without disterbing the landscape too much aside from hook up the toolbar buttons in Buoy .. i.ibb.co/FBVYqc2/image.png

esmification i suppose as far as their scripts went ignoring functionality

the mechanical conversion

that's something for another time tho

and that is a maybe anyway

god scripting is locked down to hell from a xul perspective

clrh

njsg: so this window i been creating looks pretty really similar to Netscape 0.9

which I GUESS makes sense seeing as the absolute basic bare bones to do any in-content diag with would need the most primitive interface.. eg netscape

actually mosiac but we don't talk about that

lol

traditionally

I guess I should buy pre-printed post-its saying "IRC commands require slash; numlock disabled disables numpad slash"

just for fun, the order will also include a "password is 12345" batch

WORKSFORME

njsg: numpad / soeks for me for irc irrespective on numlock enabled state

so seems might be a hardwareor OS related issue

my numeric keypad / key lsit no other character so haiving it workind differently depending on nulock state would seem to me to be incorrect behavior

I doubt I will ever adopt that _Away tradition