@jumblerg: If I clone hoplon/hoplon master right now, would I get the most compatible version of hoplon for IE 8?

yes.

although, i need to look into why the tests are failing more deeply, i'm going to take care of that this afternoon.

the tests may have actually caught something i missed with my minimalist app in firefox

allright good! Thanks!

it would be terrific if you gave it a try and let me know of any issues you find.

of course, just ping me when ready!

@jumblerg: can we back out the protocol stuff?

the ->node thing

@micha: sure, i was thinking about stripping ->node, both as a fn and a protocol, from the source altogether based our last convo, given that the string and number cases are no longer necessary

how are the string and number cases no longer necessary?

also need to figure out how to do the ui tie-in

(div "hello") i think should continue to be supported

i don't see any problem with how it was, can't we just leave it like that for a while, until we have a reason to change it?

so we don't break stuff

on the go, will address when i get behind my laptop in a bit

right now ui depends on it, but i'm open to diff implementation as discussed

wrt string, number, i thought you were suggesting we didn't need them a few days ago, but i don't recall the particulars, probably misunderstood.

something to do with prev HLisp implementation

oh yeah

i meant that the ->node function was introduced to support strings originally

there may be a better way to do it that eliminates that, but it doesn't seem worthwhile at this point

at some point it got feature creep to add numbers

changing that now doesn't seem very attractive, because it could break a lot of things

and there isn't really any pressing need to do so

@jumblerg on IE 8 (not 9) I still got a problem

do you have the dev tools removed? it works for me.

or run with simple, advanced opts

I am on simple

should I check out how to remove devtools on browserstack?

@jumblerg: ^^^

do you see hoplonKids="[object Object]" yourself?

I don't get this with IE9+

only IE8

this cause some parts of my pages not to load

the one that depends on castra cells I believe

checkout head~ and build against that

i'm going to revert the node thing anyway

hmm

git clone --branch head~

git checkout HEAD~

that will take your working directory back a commit

i’m back at my desk, looking into this stuff now

is this correct: HEAD is now at 17bd97b... ie-8-proof ->node extensibility, purge hoplon kids atom in body on page reload

you should see something like

$ git checkout HEAD~
Previous HEAD position was 17bd97b... ie-8-proof ->node extensibility, purge hoplon kids atom in body on page reload
HEAD is now at 98adfce... add ie 8 special sauce to -head property

@jumblerg

sorry I am not good at git

should I checkout again?

oh ok, I did git checkout origin/head

doing git checkout HEAD~ I see exactly what you printed

no worries, i’m not great at in either. sledgehammer your local hoplon repo via

git reset --hard origin/master

cool

i’m looking into this stuff now, give me a few mins

ok

@jumblerg

IE 9+ works with this branch

IE 8+ doesn't work. I don't see anything in my page

I know you are fixing stuff

just wanted to let you know

k,thanks

@micha: i switched over to thedavidmeister’s more elaborate test app and started using this for x-browser testing; we’ve been broken on ie 8 for a long time.

bummer, what was broken?

most of the commits, including the most recent ones, work on minimal hoplon apps, as well as the apps i’m creating internally, but this thing seems to be catching a number of deficiencies. i’m not sure what the deltas are yet between those apps that pass and the test app, but i’m working to find out.

it is a time-consuming process

could be something about how the test harness works maybe

like if the tests raise some bug themselves

that’s what i was initially suspicious of. but i created a separate test app using his index.cljs.hl page, and i’m using that for functional tests.

ahh…..

https://developer.mozilla.org/en-US/docs/Web/HTML/Element/article

many of the element contstructors in hoplon are html 5

one way to deal with this could be to associate some metadata on each one indicating which browser versions they work with, then warn when they are used

@leontalbot: if you use any html 5 elements in your app, it will break because they’re not supported by ie8 in the first place

Oh i see

Is css 3 still ok ? (can't test right now)

i think this could be the issue anyway… this is my hypothesis

Ok will test tonight

don’t make any changes until i do more research

Ok

@jumblerg: ie8 should still support those elements, they just have a default implementation

you might not be able to css them very well

etc

unless you do some other stuff before you use them

but you should be able to do body.appendChild(document.createElement("section")) in ie8

for example

right

apparently, you can create an element named anything

so that’s not it

yeah still can

which… come to think of it, could be quite useful in ui

or in hoplon in general… when new elements are defined, we could make them appear that way in the dom, could be useful for debugging

have you considered or attempted this yet?

yeah, i didn't find it so helpful really

i think it could be handy when not using ids and selectors

make it easier to spot things, doesn’t have any production value though, only useful for debugging maybe

it's complicated to track the names of things

doesn't work very well with higher order constructor things

you can also make dev tools extensions in chrome

which is kind of a nea idea

neat

kind of a mission to maintain maybe

could make some pretty cool things possible though

bbiab

@micha: so one of the culprits is the object ctor

the object ctor?

new Object()?

ah, the html-object situation perhaps

makes sense to rename that guy

that’s what i thought, but that didn’t fix it either

i think that is the IFn implementation breaking down on the object Element for some reason