@tonyw

Quote:

The worst problem is that when it crashes, it takes down the network with it, so you have to reboot to be able to use the network again.

That has never happened to me. I can pretty much always click Kill in the Reaper, close the window with Exorcist and remove the IBROWSE port with Scout. Then I can start it again and I don't think I've ever seen it crash on startup twice, after killing it once!

Or I can choose to use OWB after IB crashed. The net work always continue working for me.

Since it always crashes at the same address: 0x000003EA suggests that it tries to access an object of some kind that is NULL. So some kind of incorrect startup sequence perhaps. Timing related.

Last edited by Deniil715 on 31-Mar-2011 at 09:44 AM.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@Deniil715

Well, I've never tried Exorcist or Scout. I can imagine that IB wouldn't want to start again if its port is still there.

I must admit that it's so many years since I tried all the combinations, perhaps the network isn't hung up as I remember it. Perhaps it was just that I wanted to use IB because it was the only thing that would do the job, and OWB wasn't suitable for the occasion.

As for the constant DSI address, I agree. But the address varied depending on the page that you are fetching. If you have a Home page, it will always crash at the same offset, but change the Home page and it will always crash somewhere else IIRC.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

Regarding the startup crash, as has been stated before, it is not a bug in IBrowse, but most likely in the 68k emulation (I suspect in the area of the cpu cache flush handling, which we were asked to modify for 2.4 so that the whole cache was not flushed unnecesarily all the time which was a performance hit). We did nothing to fix this for IBrowse 2.5 because there is nothing to fix, and I suspect 2.5 68k will crash in the same way, although I'm in the group of users who cannot get 2.4 to crash on startup at all.

Likewise for any IBNetwork crashes - to my knowledge, nothing was ever done to fix this - it just doesn't crash anymore with the native PPC port, but presumably the 68k 2.5 will be almost as crash happy as 2.4. Of course, there are a few other specific crash inducing bugs which have been fixed for 2.5.

_________________
IBrowse, AmiSSL and Warp Datatype Developer

@Futaura

IB 2.5 will be 68k and PowerPC version?

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@NutsAboutAmiga

Quote:

IB 2.5 will be 68k and PowerPC version?

Yes.

_________________
IBrowse, AmiSSL and Warp Datatype Developer

@Futaura

And when will be availible for the mass ?

To be honest all looks like a bit strange to me, let me explain: IB 2.5 is surely a big effort for 1 man programmer so nothing to say about it, but i can't understand all such politic things behind the project.

- A closed source browser in 2011 when devs ask money for it, somekind of shareware formula seems mutch '80 style for me expecially when you can't offer modern technology with it and more important we have also many modern and opensource alternative now, today they are 4 native browser (already done or in development) on OS4

OWB-Reaction
OWB-MUI
NetSurf
Timberwolf

So, honestly i can't understand at all where to fit it, where is the Amiga "market segment" for such browser. (maybe old 68k machine ?)

Second, if you want to sell a program why not use Paypal or somethings similar ?
It could be quick !

For what i've understand IB 2.5 is done since dunno when, but problem was only about distribution ...

Aniway don't take it as a bad critic, i apprecciate an effort like this, developing a browser from scratch is surely a huge work

Bye

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@samo79

It's not that simple. I'm sure Oliver wants it released yesterday as much as end users do, but Stefan owns the code so it's up to him to make that call and he doesn't seem to be very active on the Amiga front anymore so he might not even share our enthusiasm at all.

There's less of a market segment for it as there would be if it supported CSS but some of us still consider it the best browser ever.
I don't know of any other browser that's as responsive and customizable, but it would turn into a fat piece of sluggish code if it had as many features as Firefox.

_________________
AlexC's free OS4 software collection

AmigaOne XE/X1000/X5000/UAE-PPC OS4 laptop/X-10 Home Automation

@samo79

I guess it's all about choice at the end of the day. I still use IBrowse because I enjoy using it. I've not seen OWB-MUI, but OWB-Rection is one of the most crash happy browsers I've used in recent times - sure most sites look like they should, but often I find that the ones that don't work in IBrowse don't work in OWB either (completely unrelated to page layout). And then there are times when it looks fine in OWB and then it crashes, but works perfectly in IBrowse despite looking "bad". And concerning Timberwolf, if I want to use Firefox I'll just boot up the PC - if OWB is anything to go by, running Firefox via VNC will be a far more pleasurable and faster experience, but after using it at work all day I prefer a change at home.

IBrowse 2.4 was the most stable browser when it was released, IMHO, and remains so (68k emulation related crashes not withstanding) - not many OS3.x users have crashes with 2.4. If we can get IBrowse 2.5 released, then you'll find it pretty much rock solid. The weak link remains flash files, as has been mentioned, but Mike Steed has done a great job at ironing out all the many bugs in libflash (this I might add is why the flash plugins released by the IB team were deemed "experimental" ) whilst improving what flash files it can decipher. One thing I'd like to do is port his plugin to OS4, although finding time is a problem - it's a shame at the lack of Amiga developers these days, because probably in years gone by somebody else would have already ported it especially since the source code is on Aminet.

_________________
IBrowse, AmiSSL and Warp Datatype Developer

@Futaura

Bug in the cache flushing of the CPU? What CPU? Petunia? The emulator doesn't emulate any CPU chip and certainly not CPU caches. It also doesn't make a difference if the JIT or interpretative emulator is used I've heard. I personally think it's a race condition during the setup of the first connections or something that only becomes apparent when running under this emulator.

Anyway. Me and others are still waiting for the release of IB2.5 (PPC). I don't expect CSS to be part of it, just want an update with some bugfixes, like the Network task that you say doesn't exist in the PPC version. Maybe the startup bug will be gone in the PPC version as well. I'd like the string and texteditor gadgets to respect Home/End/PgUp/PgDown instead of inserting 4 garbage characters each. I also hope for a few JS fixes. That's all and I think you've done it all already. So just release, please

Plus what samo79 said.

Last edited by Deniil715 on 01-Apr-2011 at 11:14 AM.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@Futaura

You already know where I stand on this.
Frankly the release of V2.5 (fixed) just makes -sense- when you speak also of an impending classic release of OS4.x.
It's not only more of an attraction for classic users %wise vs NG users, but it just makes -sense- due the performance issues of classic accelerators vs NG.
Again, if you think there is anything the community can do to facilitate this release, please let us know. Public or private...either way.

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@Deniil715

It does make difference because JIT caches generated code. When ever new 68k code is generated (a library base is one) you must flush caches. For performance reasons you dont wish to flush all caches but just relevant parts of it.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@Futaura

Quote:

OWB-Rection is one of the most crash happy browsers I've used in recent times

Well, i don't see all that crash here and i'm quite sure that most users found OWB 3.31 quite stable already, but if you had such crashes why you didn't ask devs to fix it ?

Quote:

I find that the ones that don't work in IBrowse don't work in OWB either (completely unrelated to page layout)

OWB is WebKit based browser, i really doubt you can have many layout issues, it's impossible !
Maybe certain rare page can be problematic, but eventually they are very rare issue/implementation bugs in OWB.

How we can compare IBrowse core with one that support fully HTML 4/5, CSS 2.1/3.0, Styleable form controls, Enhanced Rich Text Editing, XML technologies, SVG, XSLTProcessor, JavaScript API for XSLT, MathML etc .. ?

The OWB core is the same used in Chrome, Safari and others, the rendered pages is like that ...

Quote:

And concerning Timberwolf, if I want to use Firefox I'll just boot up the PC - if OWB is anything to go by, running Firefox via VNC will be a far more pleasurable and faster experience, but after using it at work all day I prefer a change at home.

Running Firefox natively isn't the same that run it emulated, almost for me.

Quote:

IBrowse 2.4 was the most stable browser when it was released, IMHO, and remains so (68k emulation related crashes not withstanding) - not many OS3.x users have crashes with 2.4. If we can get IBrowse 2.5 released, then you'll find it pretty much rock solid

I believe it but i can't use it for browse the web cause 99% of web pages can't work normally without fully HTML/CSS support (not to mention tons of modern features missing), so even if stable it can't work for most purpouse

Quote:

Mike Steed has done a great job at ironing out all the many bugs in libflash

I think all such discussion is similar to "OWB versus IBrowse", for Flash would be a lot better to start with somethings modern rathen then start a new implementation from scratch.

There is already a Gnash port on OS4, why not start from it ?

Again i appreciate the efforts but compare IBrowse/AmigaFlash with OWB/Gnash looks quite strange for me ...

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@samo79

Quote:

Well, i don't see all that crash here and i'm quite sure that most users found OWB 3.31 quite stable already, but if you had such crashes why you didn't ask devs to fix it ?

Because that's not my job? Many of the sites OWB has problems with require secure logins, etc, so these will be hard to debug unless the developer has the same bank account, broadband provider, etc, etc. Actually, OWB-ReAction feels too AWeb-y to me, so I rarely use it - it is just way too slow for me.

Quote:

Quote: I find that the ones that don't work in IBrowse don't work in OWB either (completely unrelated to page layout) OWB is WebKit based browser, i really doubt you can have many layout issues, it's impossible !Maybe certain rare page can be problematic, but eventually they are very rare issue/implementation bugs in OWB.

That's not what I meant - I have found some sites that look fine in OWB, but when you try to perform certain actions on a site (e.g. submit a form, click a link), it comes back with an error or simply does nothing at all. But, when you try the same page in IBrowse, the site works properly - once you've figured out the layout of the form, and/or hidden elements, due to the bad layout .

Quote:

Running Firefox natively isn't the same that run it emulated, almost for me.

Of course - me too. What I was saying is that running Firefox via VNC is miles faster than running OWB natively on my A1XE. I prefer to do that than use OWB.

Quote:

I think all such discussion is similar to "OWB versus IBrowse", for Flash would be a lot better to start with somethings modern rathen then start a new implementation from scratch. There is already a Gnash port on OS4, why not start from it ?

The IBrowse development team has always been fully co-operative with any developer who wishes to develop a plugin for IBrowse - it was the biggest new feature added to 2.4 from a development point of view. IIRC, I already provided the developer of the Gnash port with the necessary PluginAPI files to build a plugin for IBrowse, but it seems he is not interested.

_________________
IBrowse, AmiSSL and Warp Datatype Developer

@Futaura

Quote:

Many of the sites OWB has problems with require secure logins, etc, so these will be hard to debug unless the developer has the same bank account, broadband provider, etc

Quote:

That's not what I meant - I have found some sites that look fine in OWB, but when you try to perform certain actions on a site (e.g. submit a form, click a link), it comes back with an error or simply does nothing at all. But, when you try the same page in IBrowse, the site works properly - once you've figured out the layout of the form, and/or hidden elements, due to the bad layou

This is more probably caused by bad DNS/connection issue, not a specific WebKit/layout problem, see this bug report for an example:

http://bugs.os4depot.net/?function=viewissue&issueid=469

Quote:

The IBrowse development team has always been fully co-operative with any developer who wishes to develop a plugin for IBrowse - it was the biggest new feature added to 2.4 from a development point of view. IIRC, I already provided the developer of the Gnash port with the necessary PluginAPI files to build a plugin for IBrowse, but it seems he is not interested.

Yep, That's not what I meant.
A plugin support in IBrowse would be ok, what is useless IMHO is create a new Flash plugin implementation rathen than create one based to Gnash or SWFdec

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@RacerX

Your crash log for 68k will be useless. Turn off JIT and try again. No one will be able to find an errors when it crashes in a RTS.

Now, amongst all these suggestions, check your cache volume space. If the volume you have your cache on runs out of space IBrowse will crash! Check my thread as just started crashing instantly for me as well:
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=33365&forum=32

All fixed now.

@Futaura

Quote:

IBrowse 2.4 was the most stable browser when it was released, IMHO, and remains so (68k emulation related crashes not withstanding) - not many OS3.x users have crashes with 2.4.

That's because OS3 on 68K was a lazy OS. A program can do many illegal things by todays standards and you won't see a whimper. Even simple thiings like hitting on location 0, passing null pointers which the OS will still read from regardless and mis-using functions like OpenFromLock() (by freeing a freed lock) will go by unhitched.

On OS4 this will break instantly and crash it doesn't let any of this crap go through as it is a strict OS!

And BTW, hate to say it but IBrowse does have at least one bug. The cache code is buggy, if it runs out of space IBrowse will crash.

@AlexC

Quote:

There's less of a market segment for it as there would be if it supported CSS

IBrowse 2.5 was suppossed to have CSS. That was one of the main features.

@Hypex

What? I've read that 3.0 would have CSS support lots and lots of times for many years but never that 2.5 would have support for it.

What I've read about 2.5 is that it wasn't planed to be released at all but it will be anyway as a bug-fix release of 2.4.

@hotrod

You're kinda right. CSS was never targetted for 2.5. 3.0 was the next intended release after 2.4 and was to have CSS support. However, it was subsequently decided to release an OS4 native build soon after 2.4, partly as a surprise, and that would be 2.5. Obviously "soon" has slipped somewhat . Pretty much the first thing that happened after 2.4 was released was backporting the OS4 changes from the 3.0 source tree into the 2.x tree (for 2.5) as some work had already been done on porting IBrowse to OS4 before 2.4 was released, and then work on the port continued until it was up and running (all done on my A1200T at the time - quite a chore).

@Hypex

Quote:

That's because OS3 on 68K was a lazy OS. A program can do many illegal things by todays standards and you won't see a whimper. Even simple thiings like hitting on location 0, passing null pointers which the OS will still read from regardless and mis-using functions like OpenFromLock() (by freeing a freed lock) will go by unhitched.On OS4 this will break instantly and crash it doesn't let any of this crap go through as it is a strict OS!

I agree in general, but what I wrote above is why I insist that many of the IB crashes, particularly the unexplained ones, are due to the 68k emulation because when the 2.5 OS4 build was first working, it was pretty much the same as 2.4 - there was no other development other than the porting work. Yet, the native version did not crash like the 68k version did - I don't think it fixed itself by magic (would be handy sometimes though ).

Of course, I'm not saying 2.4 doesn't have bugs, because it does, and no doubt 2.5 does too. I'll take a look at your cache disk full problem - it's not something I've heard about before, and my cache volume often runs out of space (IBrowse betas write lots of large debug logs there related to cache operation and I sometimes forget to clear them out).

_________________
IBrowse, AmiSSL and Warp Datatype Developer

@Futaura

Quote:

I agree in general, but what I wrote above is why I insist that many of the IB crashes, particularly the unexplained ones, are due to the 68k emulation because when the 2.5 OS4 build was first working, it was pretty much the same as 2.4

I agree 68K emulation is a problem, from crash logs, I put the following files on the black list, IB crashes are very rare now.

IBrowse
gif.codec
imagedecode.module
IBCacheBrowser.mcc
IBPreferences.ibcc
IBTextEditField.ibcc

I also believe os4 has a bug writing to disk with at least 2 apps for me, IB & Transmission. Writing to SFS & JXFS partitions cause GRs, writing to a FFS2 partition gives no GRs. Given that, I set my disk cache up in RAM:T.

_________________
Hate tends to make you look stupid...