I can't confirm on 61.0.3163.100, running on Ubuntu 14.04 LTS, the bottom of the page looks the same before and after re-load.

Could you mind to explain what "- the bottom map - Openlayers" is? Maybe it looks the same because I don't understand what it means?

Thank you for providing more feedback. Adding requester "kojii@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Hello,

I join pictures that explain the "How to reproduce?".

Say me if it's OK.

Thanks,

Eric

PS : if you inspect the element by clicking on the map, you will see (maybe ?) width and height of the map div, looping .... without throw onLoad.

Deleted: initial_load_of_page_no_zoom_100percent.jpg 221 KB

	initial_load_of_page_no_zoom_100percent.jpg 221 KB View Download

Deleted: reduce_the_page_zoom_to_90_percent.jpg 187 KB

	reduce_the_page_zoom_to_90_percent.jpg 187 KB View Download

Deleted: reload_page_map_is_not_displayed_infinite_loop.jpg 89.0 KB

	reload_page_map_is_not_displayed_infinite_loop.jpg 89.0 KB View Download

Deleted: reset_the_page_zoom_to_100_percent_maps_are_displayed.jpg 274 KB

	reset_the_page_zoom_to_100_percent_maps_are_displayed.jpg 274 KB View Download

Able to reproduce this issue on reported version 61.0.3163.100 and on latest dev 63.0.3239.7 using Ubuntu 14.04,Mac 10.12.6 and Windows 10 with attached link given in comment#0.

Good Build: 58.0.3011.0 
Bad Build: 58.0.3012.0 

You are probably looking for a change made after 449942 (known good), but no later than 449943 (first known bad).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/d6df828820be4b0f22b9bed959fb2bd93f84b331..934becac5daa91ea979fb66e4ae21761ca11ebc9

Review-Url: https://codereview.chromium.org/2640143005

Suspecting same from above changelog.

@karlo: Please confirm the bug and help in re-assigning if it is has nothing to do with your change.

Thanks!

Sorry for the long delay in looking at this. I opened the page, and this looks to me like a problem with the website. A script "GeoportalExtended.js" is infinitely looping, calling a function called "setInformationPanelVisibility" through setTimeout. Unfortunately the contents of the file are minified, so I can't really tell what's going on, but it looks like a line like

a = a || !this.isMapReady(l) || c.h != g + j || q.w == 0;

is setting the condition "a", which controls whether setInformationPanelVisibility sets a timer to recursively call itself. It also looks like "!this.isMapReady(l)" returns true, but also:

c.h (appears to be the size of this.infoCntrl) == 50, but g + j == 49.
g = this.infoCntrl.div.offsetHeight - this.infoCntrl.div.clientHeight = 50 - 44
j = Geoportal.Util.getComputedStyle(this.infoCntrl.div, "height", true) = 43


It sounds to me like there's a problem somewhere with rounding, so you should probably file a bug with OpenLayers instead.

Hello,

The js script is in place since 2012/2013 and is used in many website. Nor Firefox or Chromium have never shown this behavior before. And this only occure now on Chromium.

As nothing has been changed in the script ? I guess it's have to do with chromium change ?

Thanks,

Eric

I'm not sure how you're including OpenLayers, so it's possible something changed on their side too. 

I suppose it's possible that the CL to do with sub-pixel rendering changes the returned value for height from getComputedStyle at lower zoom levels.

I'll hand this to the Layout team who were doing the subpixel rendering things. PTAL

Without minimizing this script down by hand it's hard to say what exactly is going on. I can imagine there is some javascript-visible change from https://codereview.chromium.org/2640143005 where the javascript code was relying on snapped border values that are no longer snapped.

I'm not aware of other bugs reported in this area so I'd lean towards this being a bug or assumption in the javascript that has been broken by this change.

Are you about to upgrade the OpenLayers scripts?

Hello,

>Are you about to upgrade the OpenLayers scripts?

Not for the moment, but I will do in next months, switch to openlayers 3.

I have posted this thread here : https://www.developpez.net/forums/d1772870/applications/sig-systeme-d-information-geographique/ign-api-geoportail/passage-l-api-v2-api-v3/#post9793791

But I'm not sure it's related.

Thanks,

Eric

PS : Test page has moved here : http://www.opalesurfcasting.net/faune_et_flore/le_bar_commun_ou_loup_-_dicentrarchus_labrax_article690.html

Am I correctly reading https://www.developpez.net/forums/d1772870/applications/sig-systeme-d-information-geographique/ign-api-geoportail/passage-l-api-v2-api-v3/#post9793791 (and the other linked thread) that this is a bug in the javascript + chrome 62 and not necessarily chrome itself? Could you see if someone can create a small testcase without the javascript library that demonstrates the bug?

Closing due to lack of feedback and activity.