Sitepatching updates

tweaking the broken code until it works, one site at a time

Facebook, GDocs, Apple, Shopping.com, Sears

By Ola P. Kleivenolak. Monday, September 10, 2012 3:31:26 PM

Added patches

PATCH-852, facebook: avoid unwanted chat box scroll. Small regression from a Core fix causes chat box to jump. 12.50 only.

PATCH-851, Fix event object detection in old DynAPI code.

PATCH-850, ieee.org - postpone insertion of JSONP data source until we've parsed the element the data is meant to be inserted into. Core issue.

PATCH-847, sears.com - fix moving product thumbnail images. Core issue with collapsing margins on hover.

PATCH-846, apple.com/jobs: don't reload from within unload handler.
window.onunload = function(){ location.reload(true);};

OK, some browsers have exceptions to handle such things, but why would you write it in the first place?

PATCH-844, clarkhoward.com: abouse of CSS content property.

PATCH-836, shopping.com - work around browser sniff to see help info.

PATCH-833, help.sap.com : fool sniffing to make frameset complete. Old "Netscape" sniffer.

Changed patches

PATCH-382, Google Spreadsheets cell highlight mismatch and key event workaround. For 12.50 only, now spreadsheets work better than ever in Opera thanks to Mr. Byberg's heroic dive into obfuscated browser sniff branches.

PATCH-176, Allow upload of workspace resources in Salesforce. Add try/catch to avoid x-domain error messages.

GDocs, .NetCtrip, Garmin, udemy

Comments

Martin KadlecBS-Harou Monday, September 10, 2012 3:37:37 PM

PATCH-852, facebook: avoid unwanted chat box scroll. Small regression from a Core fix causes chat box to jump. 12.50 only.


You should have an automated tests for this. It must be one of the most reappearing facebook bugs.

And as always, thank you very much for all the patches up

Ola P. Kleivenolak Monday, September 10, 2012 3:55:16 PM

Sure, but it is never the same trigger. We always add automated tests for bugs that are fixed and this one was a never-before-seen variant that we only managed to track down today. Once that was done, adding a patch was easy enough. However, this is in an area known to be fragile in Opera (generated content+positioning+overflow) and a rewrite is on the shortlist - should free up some debugging time p

d4rkn1ght Monday, September 10, 2012 4:14:50 PM

up

Spadar ShutSpShut Monday, September 10, 2012 4:19:01 PM

Glad to see spreadsheets fixed!

Galileo Monday, September 10, 2012 4:45:50 PM

PATCH-852, facebook: avoid unwanted chat box scroll. Small regression from a Core fix causes chat box to jump. 12.50 only.

Was trying to figure out if an extension was causing that, now i know up
Target version and time stamp of the active browser.js file is Opera Desktop 12.50 core 2.12.378, September 10, 2012. Active patches: 209 .

stwart Monday, September 10, 2012 9:52:03 PM

orkut.com high cpu usage after login(last opera 12.50 x64).

Çağlar Yeşilyurtgreench Monday, September 10, 2012 10:20:57 PM

This is a bit different content but it is related with Open Web Standards. I wonder why browsers handle border color(except solid type) with different color. To demonstrate, http://jsfiddle.net/kg4JC/8/ If you can check border-color on chrome and opera, both gives different border color. Chrome gives more blackish and Opera gives more gray. Which one is correct? I should miss something, because I didn't notice any information about border color on the spec(http://www.w3.org/TR/CSS2/box.html#value-def-inset)

Thank you...

Charles SchlossChas4 Tuesday, September 11, 2012 12:16:45 AM

Target version and time stamp of the active browser.js file is Opera Desktop 12.01 core 2.10.289, September 10, 2012. Active patches: 241 .


remove the fix on Apple's site as you know now why Apple.com is for resolution higher than a HDTV wink

Ola P. Kleivenolak Tuesday, September 11, 2012 9:05:16 AM

Çağlar: there is no "correct" as the standard only says "...but UAs may choose their own algorithm to calculate the actual colors used." There is a summary of browser behavior at http://stackoverflow.com/questions/4147940/how-do-browsers-determine-which-exact-colors-to-use-for-border-inset-or-outset - all slightly different smile

Çağlar Yeşilyurtgreench Tuesday, September 11, 2012 9:25:59 AM

@ola Thank you so much. smile

Christophchristoph142 Tuesday, September 11, 2012 10:14:26 AM

Do you intend to patch that broken bit of CSS-gradients (bottom <-> to bottom) in Facebook as well?

Martin KadlecBS-Harou Tuesday, September 11, 2012 10:42:50 AM

Originally posted by christoph142:

Do you intend to patch that broken bit of CSS-gradients (bottom <-> to bottom) in Facebook as well?


actually it is bottom <-> to top. I hate it, CSS WG should have left the original one-word constants p

Çağlar Yeşilyurtgreench Thursday, September 13, 2012 8:17:49 PM

Hello Ola,
Again me smile

This time, I bring you an old but but remarkable and very popular web site in Turkey, memurlar.net. It holds many problems that are related with Microsoft - W3C standards complexity. However, Opera cannot render width of a table properly as other browser's do properly.

Opera SS: http://imageshack.us/photo/my-images/525/opera2g.jpg/ IE and other browsers SS: http://imageshack.us/photo/my-images/856/ie2.jpg/

Black area has one colgroup and two col elements which one has 150 of width. Opera should give 150px width to left column. However, other col element has "*" of width which doesn't make sense. If you change table layout to auto, table is displayed fine.

URL : http://uye.memurlar.net/profile.aspx
I will send(to your myopera mail) login information.

Charles SchlossChas4 Friday, September 14, 2012 5:55:15 PM

Posted in the open the web forum about a Kronos site that blocks many browsers and masking in Opera and Safari does not work (even as IE 7 thru 9)

Wonder if anyone can find the browser sniffers and maybe offer a possible fix

rseiler Friday, September 14, 2012 7:50:54 PM

Hi there. It looks like something has gone off with MS forums again. Like before, the Submit button is dead (also, you don't see a cursor until you start typing, but that's relatively minor). I'm using Opera Desktop 12.10 core 2.12.378, September 10, 2012 with 12.10.1594 x86 on Win8 x64 (clean Opera profile). Sorry, but I can't narrow down when it started beyond the last couple weeks or so. Thanks.

http://social.microsoft.com/Forums/en-US/categories

While I'm here, I thought I'd give a shout out to two issues with other forums, both still present and longstanding:
http://my.opera.com/community/forums/topic.dml?id=1515552
http://my.opera.com/community/forums/topic.dml?id=1299932

Jimtoyotabedzrock Sunday, September 16, 2012 12:23:06 AM

The political site http://www.politico.com/ blanks itself in Opera Next.

It is a very high traffic site this time of year.

metude Sunday, September 16, 2012 11:29:27 AM

http://jsfiddle.net/jYsB2/
Why Opera deosn't have left border? Other browsers have.

rseiler Monday, September 17, 2012 4:18:21 PM

@Jim, what do you mean? Maybe it's your ad blocker doing something that it shouldn't? I'm not seeing an issue here, though we probably use different ad blocker lists.

Ola P. Kleivenolak Wednesday, September 19, 2012 7:24:00 AM

rseiler: MS forums posting seems to be working in 12.10 here. Just posted in the OT forum.

Jim: works here as well, could be some temp broken ad that does a document.write after onload or similar.

Çağlar: haven't had the time yet, but will get to it smile

metude Wednesday, September 19, 2012 3:24:20 PM

@olak Could you check this: http://jsfiddle.net/jYsB2/
Other browsers have left border but Opera doesn't have it.

Ola P. Kleivenolak Thursday, September 20, 2012 8:43:21 AM

Paint order bug perhaps, the lack of left border on the child nullifies the ul border or something.

metude Thursday, September 20, 2012 12:36:10 PM

I thought bad coding or something related. Thanks, this is not related with your department but reported as DSK-374419.

Write a comment

New comments have been disabled for this post.