345531 - Request for addressing various Java problems

Status: VERIFIED INVALID

(Core Graveyard :: Java: OJI, defect)

Description

[ChrisX]

User-Agent:       Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20060419 SeaMonkey/1.5a
Build Identifier: SeaMonkey 1.5a (Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20060711 SeaMonkey/1.5a)

I would like to file this bug in order to politely request that the way Java applets are displayed be fixed. I am aware of the fact that I'm filing a summary of some other bugs, but I think so many issues have accumulated which are all related to Java that it has now reached a state where SeaMonkey is not only just lagging behind other browsers, but is pretty much unusable for displaying Java. My opinion is that Java display should be given a higher priority in development, like Flash must have gotten because its performance and compatibility have considerably improved during the last months.

These issues make Java in SeaMonkey unbearable:

1) Java is responsible for the vast majority of crashes I experience with SeaMonkey 1.5a (Gecko 1.9) builds. If anybody is interested in a crash log, I can provide one:

http://chtrusch.ch.funpic.de/Verschiedenes/Crashlog.txt

2) Java applets are displayed improperly and with unexpected results. Sometimes they work, sometimes the don't. Sometimes you have to switch to another tab, scroll down and up again, or click somewhere within the page to get them displayed at all. Here are two screenshots of a website (http://www.elektron-berlin.de) using three Java applets (top left portion of the page) in a row. Be careful, it made SeaMonkey crash when trying to re-load. This one

http://img377.imageshack.us/img377/8655/javasm15afp2.jpg

is SeaMonkey 1.5a (Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20060711 SeaMonkey/1.5a), and this one 

http://img110.imageshack.us/img110/9755/javamoz12pu1.jpg

is Mozilla 1.2.1 on Mac OS 9 (!) (Mozilla/5.0 (Macintosh; U; PPC; en-US; rv:1.2.1) Gecko/20021130). I see a clear and concerning regression here in comparison to an almost 4 year-old Gecko engine which got this rendered perfectly right. The latest Camino and Firefox (official) builds which still have Gecko 1.8 don't have a problem with this website, interestingly, although FF crashed once in three times trying to display the website. Also, Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20060419 SeaMonkey/1.5a doesn't have this problem (despite Gecko 1.9), but it also still has JEP 0.9.5+d which may be related to that. 

3) After Java has loaded, all keyboard commands are executed twice until the browser is re-started. I'm not going to harp on this one since it's being addressed by Steven Michaud.

4) Java applets *still* draw in the wrong tabs. dto.

My general question to the people developing Java support in Gecko would be why have we seen (at least from a user perspective) a constant line of long-standing bugs, regressions (especially in the last months), decreasing application stability, workarounds and emergency fixes in this field - the situation for the user getting worse and worse - while other engines like KHTML or Presto manage to display Java stuff more or less lean and clean, even on the Mac with Apple's obviously problematic Java implementation. I really do care about SeaMonkey and Gecko (that's why I'm filing this bug), but all this makes me once again look at Opera and even Safari with envy. 

Reproducible: Always

Comment 1

[timeless]

so. first, mozilla.org doesn't really have a magical pool from which we draw developers. if you want to make a request, throw $$$$ or $$$$$ into a pool or at Mozilla Corporation and ask them to magically invent a developer. who knows, it might happen. and finally, bugzilla isn't the place for that problem.

most distressingly, java for osx has one active developer who's been working busily for the past couple of years.

Comment 2

[ChrisX]

That was the *exact* snobby comment I was waiting for as an answer to testing, reporting bugs and asking some general and honest questions in a maybe just slightly unusual way. And this bug is *not* resolved.

Comment 3

[Steven Michaud [:smichaud] (Retired)]

If you want to help make Mozilla.org browsers better (and the programs
bundled with them, such as the Java Embedding Plugin), please open
detailed, informative reports of reproduceable problems.  Please also
open a new bug for each different problem -- they're _much_ easier to
deal with that way.

Reproduceability is the key -- without being able to reproduce a
problem, neither I nor any other programmer can fix it.  So if I (or
some other programmer) can't reproduce a particular problem, you will
often have to think particularly carefully how _you_ are able to
reproduce it, and follow suggestions about (say) temporarily disabling
extensions, and so forth.

If you merely want to make vague accusations, as you have here, please
save your time and ours -- don't bother.

Comment 4

[ChrisX]

@ Steven Michaud: That was not meant as an accusation, I'm sorry if you understood it that way. I'm German, and so a) blunt by nature, and b) can't stand if something is not in order :-) You know that I'm also filing detailed bugs elsewhere, I just wanted to pronounce what I felt needs to be said. I know very well that without your work, the situation would be much worse. Points 2) and 3) from my initial report are actually fixed in JEP 0.9.5+f+, so thank you! I still think that Mozilla needs to allot more rescources to Java development for the Mac instead of letting just one or two people struggle with this. As soon as I get money, Mozilla is on top of my donation list.