![]() ![]()
The fix for IE5 on MacOS9 though, did not work, just worsened it. I believe these new issues are related to the fix you made for IE5 on MacOS9. But the fix for Classilla works, I will apply the fix to the master branch to continue testing. I used the branch disable-mac-ie-native-navigation for this test.Ĭlassilla started to crash a few times, notably when I always tried the same site -, and Browservice also crashed a few times in the proccess, I had to restart it once. However, several unrelated stability issues surfaced. It recognized the user agent used by Classilla. I can confirm that a8bf950 fixes the issue for Classilla. Maybe friday night I can have a head start on it. I still haven't tried your modifications but I plan to do so this weekend, it's when I have more time to do these things. Yup, apart of the redirect loop, IE5 works well.I tried the same tweaks I did for Classilla, but no luck. Netscape 7.0.2 though, the last version to work on OS 9, did not work. You can, though, treat it just like another mozilla browser, (Classila is based on Mozilla 1.3x code) and tell users to spoof one of the available user agent strings on Classilla if they want to use it. It can be changed to other user agent strings like Firefox 35 ESR as far as I can remember now. Classilla initially spoofs a mobile nokia based user agent string by default, out of the box. IE5 is probably unrelated to this, as it worked without any fixes (apart from the redirect loop thing), right? #Best web browser for mac os 9 portableSo if there is no other portable way to detect Classilla, the only option that comes to mind would be to have command line arguments for enabling width -= 1 height -= 2. The fix a8bf950 tries to detect only Classilla based on the user agent string, but based on what you say it probably won't work, as Classilla spoofs its user agent. #Best web browser for mac os 9 windowsI haven’t tested yet with Windows For Workgroups 3.11 VM. Current version works great for for my Win 9x browsers (IE6 and Netscape 9). Does this new version assume that it will always be these values for both browsers on Mac OS 9? Because these are the only browsers that needed this change so far. This is where Browservice becomes a great addition. ![]() Although I think that old browsers with TLS 1.2 support are still “useless” when their rendering engines are now useless for Web 2.0 (Netscape 9 and Firefox > 35 ESR are still quite decent though). (But this rarely happens, Mac Garden always has everything). Do you want me to show which other user agents are selectable from Classila’s preference panes?Ĭlassilla can use Browservice too (I have to enable global JS for it to work), but I currently had set Classilla to use Cameron Kaiser’s Crypto Ancienne to give it TLS 1.2 support as a backup browser, for some stuff like https downloads when I cannot find what I want from Macintosh Garden, for example. I can test a8bf950 later, but Classilla uses a nokia user agent, because Cameron Kaiser thinks it works best for old macs to deliver mobile web pages. I don’t know if this change affected my other Windows VMs, but apparently Win 9x IE6 is still working OK. ![]() So good that I haven’t changed again for more tests and set IE 5.1.7 to not show any toolbars, apart of back, forward, home and reload (and Browservice’s toolbar). Setting them to width -= 1 and height -= 2 as you previously mentioned, works for both IE 5.1.7 and Classilla 9.3.4b perfectly under Mac OS 9.2.2 (QEMU VM). #Best web browser for mac os 9 updateNote that you need to update the Browservice executable after each change by running make, because the HTML templates are built into the executable. If this works, try smaller numbers instead of 20, for example width -= 1 height -= 2. To do this, add the line width -= 20 height -= 20 before line 198 in viceplugins/retrojsvice/html/main.html. You could try whether this works in your case as well. On some browsers (IE4 and IE5 on Win3.x), I was able to hack around this problem by reducing the detected width and height by 1 and 2 pixels, respectively (this way, the image does not overflow the page normally the image overflows for signaling purposes). Detecting the browser window size such that no scrollbars are created.Browservice tries to do this using CSS body, which is not supported on all browsers. As you said, it is caused by both Browservice and the browser reacting to image size and scroll bar visibility changes in a feedback loop. I remember coming across this resizing issue on various browsers probably at least on classic Mozilla on Debian Sarge. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |