-
NewTobinParadigm
ok since when did windows start keeping shit on the GPU it started on and not move execution across whatever gpu is driving taht display?
-
therube
Tools | Translate does look to still work in SeaMonkey 2.49.5
-
therube
though a useragent change (to 2.49.5) does not get things working in SM 2.53.xx
-
therube
beta page download links are returning 404,
seamonkey-project.org/releases/2.53.15b1
-
therube
-
frg_Away
does an older 2.53 still work maybe 253.1^0 or so with translate?
-
frg_Away
Copy step for thew binaries were not doe it seems. They are still in tmp
-
therube
frg: Ah, 2.53.1 does still work. that uses "60" (but then a UA change did not make any difference in current)
-
therube
sample URL,
uwe-sieber.de/deutsch.html (for a de to en)
-
therube
works: seamonkey-2.53.11.en-US.win64.installer.exe
-
therube
broken: seamonkey-2.53.12.en-US.win64.installer.exe
-
therube
11:
-
therube
Use of attributesí nodeValue attribute is deprecated. Use value instead.
-
therube
m=el_main:368:329
-
therube
12:
-
therube
SyntaxError: An invalid or illegal string was specified
-
therube
m=el_main:468
-
njsg
I had no luck toggling UA strings, even setting IE6 led to failure
-
njsg
perhaps I forgot to clear some cache?
-
njsg
I think it was visibly serving a different layout to "IE6", but it still failed with that message
-
therube
well, in one case, 11, you get a warning, in the other, 12, there is an actual Error, so... ?
-
njsg
oh, I mean luck getting it to actually translate
-
njsg
the code seems to be calling performance.measure if it exists, or am I getting misled by the line and column numbers and the spaghetti
-
njsg
?
-
» njsg wonders what would EWD have to say on this spaghetti-like javascript code format
-
njsg
therube: that syntaxerror sounds like it happens when the page is initializing a "google translate" instance, setting window.performance=null and then doing window.google.translate.TranslateService(); unbreaks the translator
-
njsg
-
njsg
therube: setting dom.enable_performance_navigation_timing to false might work as a workaround too
-
njsg
now I had a... hiccup in google maps, I wonder if it's related to dom.enable_performance_navigation_timing=false, or if it was just a change on google's side
-
njsg
the hiccup itself was unrelated, but setting that to false does give me a slightly different UI, at least in incognito mode
-
therube
-
therube
now, no idea what the Pref is for... but it works.
-
frg_Away
therube will break other sites. Not recommended
-
therube
and, it doesn't seem to be working - always, anyhow.
-
therube
i'm confused at this point, but gotta go now...
-
njsg
frg_Away: I wonder if what's needed here for translate is
bug 1444150, or will that break things too? or not easy to apply?
-
njsg
therube: I think it's toggling certain kinds of measurements in the performance measurement code
-
njsg
therube: it's not hiding window.performance (although setting it to null allows google translate to initialize, because it will skip the code throwing the SyntaxError)
-
therube
it does work here,
welt.de, but not on uwe's page that i posted earlier
-
frg_Away
njsg thought it was in but it is not. Exported it and will see if it applies easy.
-
njsg
therube: there's «performance.getEntriesByType("navigation")» in the google translate code, so I wonder if that's what makes the preference "unbreak" google translate
-
frg_Away
njsg applies clean
-
frg_Away
15 minutes to get it in :)
-
frg_Away
put it in. Will be in the next pre build
-
njsg
therube: it's a bit... challenging to follow this code, but it operates on the first entry coming from that .getEntriesByType() call, and that will be empty if the page is loaded after the setting has been set to false. Given that the SyntaxError code is doing performance measurements, it's possible that that code runs on the result of getEntriesByType.
-
njsg
oh, after looking at the code in the bug: privacy.resistFingerprinting=true avoids the SyntaxError too, no idea if it causes issues elsewhere
-
frg_Away
njsg it broke other stuff including the reported version I think. You can't win with google and idiotic development apis in production code.
-
frg_Away
gitlab 2.53 updated. reshuffle for 2.53.15 patches and 2 py3 backports. No need to stop current builds. Just housekeeping.
-
tomman
Real Men Test In Production™
-
hardys
-
frg_Away
hardys yes. Mailed ewong. Have no rights. Should be fixed by tomorrow. If not I will backout the website changes.
-
hardys
no problem for me,I always the latest pre beta
-
hardys
always USE the latest pre beta :)
-
frg_Away
Bug 1444150 does not help with the traslate page problem.
-
frg_Away
^translate