@pseudaxos I would be very surprised if Mega Download works at all on an Amiga, given it fully "downloads" to memory (RAM), before it lets you write the download to hard disk.
@sundown I have no idea what versions you were comparing against, but I can assure you I did a binary comparison against the last Odyssey beta version I had (file date 19th January), and it was identical. Kas1e also confirms that it should have been the same, since it was a copy.
@ChrisH I didn't strip it by default because of reassons on which LiveForIt pointed out. Striping will make debug logs useless, as then will have no references to name of functions and so on.
My router developer updated it's firmware and now i can't access the routers webpages anymore and only get this message: Quote:
Willkommen bei Ihrer FRITZ!Box
Die FRITZ!Box-Benutzeroberfläche nutzt moderne Methoden für eine optimale Darstellung auf einer Vielzahl von Geräten. Leider unterstützt Ihr Browser diese Methoden nicht.
Wir empfehlen daher die Aktualisierung des vorhandenen oder die Verwendung eines anderen Browsers.
Mögliche Browser-Alternativen finden Sie hier
Translation: Quote:
Welcome...
The ...GUI uses modern methods for optimized display on a variety of machines. Unfortunately your browser doesn't support these methods.
We recommend updating your browser or using another one.
Unfortunately none of the three (on AmigaOS) available browsers support the new methods.
I checked on a windows machine and the method where it breaks seem to be Window.requestAnimationFrame().
According to to this site the version of WebKit introducing this method was 6.0 (On the linked site scroll down to "Browser compatibility").
Would it be possible to update Odyssey to WebKit 6.0?
Yes but that makes any debug logs from users useless.
While I hate to disagree with you & Kas1e, it certainly does NOT make debug logs useless! You just need to keep the UNstripped executable around (along with the source files that you'd keep anyway), and then you can use addr2line on that unstripped executable, and it will give you the correct information (including function name).
OTOH, yes, having to use addr2line to see function names does make it a little more laborious.
ANYWAY, I never said you should strip Odyssey executable. I merely pointed out that it was not stripped, in case this was an accident. 60MB vs 50MB is hardly worth much worrying!
While I hate to disagree with you & Kas1e, it certainly does NOT make debug logs useless! You just need to keep the UNstripped executable around (along with the source files that you'd keep anyway), and then you can use addr2line on that unstripped executable, and it will give you the correct information (including function name).
@Hans I ended-up using "-g" rather than "-gstabs" (as your blog suggests). Sadly I can't remember the reason now, nor what difference it is supposed to make.
@Hans Yep, that your article is nice, and add2line is welknown, its just i choice to avoid mambo with addr2line, as well as its easy for copy+paste/forwarding/etc crashlogs when they have names of functions insted of reference values, which i need manually translate and loose time for.
@ChrisH -gstabs just recommented by core devs on os4 (dunno why, maybe they have reasons for), but its about the same as -g. On linuxes -gstabs produce debug binaries without GDB extenstions, while -g produce debug binaries with GDB extensions. It can be that our old GDB do not handle any special GDB extensions, so for us its all the same, -g or -gstabs.
@Raziel When i see something swapped around (like mirrored on your image), i always think about endianes differences. But i can be wrong. Its always good to compare morphos version on the same sites if anyone have ability to do so, so to known if it non os4 port specific problem
Browsing to this site makes part of the text display mirrored.
I (seldomly) stumble over such sites, why is that? Is Odyssey interpreting css/html code wrong?
I'd say its a feature of the website. Try clicking on either of the cards, and it will flip to display the text correctly.
Glancing at the HTML code, there's indications that this flip-thing is disabled in certain browsers (probably those that doesn't support the attribute "flip").
I just talked to a MorphOS User about my router problem and (as he has nearly the same setup as me) he is still able to access his router (with the same firmware) but he's using Odyssey 1.24//MorphOS.
He also tried it with Sputnik and gets the same error message.
So, would it be possible to update our port to 1.24, please? I'd really like to be able to access my router again.
I just talked to a MorphOS User about my router problem and (as he has nearly the same setup as me) he is still able to access his router (with the same firmware) but he's using Odyssey 1.24//MorphOS.
He also tried it with Sputnik and gets the same error message.
You may ask him to try 1.23 morphos version.
If it will not works on 1.23, but work on 1.24 => webkit problem. If it will works on 1.23 on mos, then or porting-to-aos4-problem, or something with settings.
Quote:
So, would it be possible to update our port to 1.24, please?
I didn't see sources of 1.24 online, but newer ones (which on github), and which, have welknown js-endian problems.