RU/2: Форум. Общение пользователей и разработчиков OS/2 (eCS). : firefox3


Список сообщений | Написать новое | Ответить на сообщение | Домой Поиск:
Предыдущее сообщение | Следующее сообщение
From : Alexander Conon, http://alexander-conon.narod.ru/
To : Alexander Conon
Subj : firefox3
>>Ну возьми и хотя бы спроси об этом портера, если так интересно. weilbacher.org
>произвёл кое-какие замеры и отправил вопрос по поводу
>os/2 firefox 3.0.3 vs windows firefox 3.0.1
поимел переписку (или она меня?). короче лучше сами почитайте
ясно одно - с такими портерами осевой субж нормально работать не будет. никогда.
прям хоть сам садись и пиши свой браузер
... если б голова не была в другую сторону заточена :(
-------
> there are little differences:
> win has two plugins more
> and win on fat32 but os/2 on hpfs
^^^^ OK, that explains it then. Use JFS!

>>> memory usage with one empty page
>>> win 26660 K ph + 94264 K virt
>>> os/2
>>> Private Object Summary for 'FIREFOX':
>> [...]
>> I simply don't know what ph+virt means in the Windows case,
> physical memory + virtual memory
Ah, sure, but that doesn't really map on any of the values that you
list for OS/2.

>> and in
>> general it's _really_ hard to compare memory usage across systems.
>> That said, I would expect the OS/2 version to use a bit more memory.
>> It's not using the JEMalloc allocator that FF is using on the other
>> platforms (nobody has ported that), and our cairo port is not using
>> memory as well as on Windows. I don't think the latter is fixable
>> without _huge_ amounts of work.
> it is 10% only and it isnt screen of death.
?

> but os/2 version not swap ! it is wholly in the phisical memory !
> and it is too big programm for this :(
It _will_ swap if the system is so configured. That it doesn't show up
as such in your Theseus may be because Theseus doesn't handle high
memory correctly.

>>> and folks says that os/2 version has empty cycles
>> Not sure what this means...
> sorry my english - I dont know how to translate it.
> it means that there are fragments
> when firefox take CPU for a time but do _nothing_.
> mail to "Slavik Gnatenko" <>
Hasn't really happened to me, at least not while loading a page. If
you have a way to reproduce this, please file a bug in Bugzilla.

> and one more bug:
> when trouble with my ISP - firefox suspend
> and block system queue of communication.
> need reboot :(
That's an old problem and not OS/2 specific as far as I know. But you
shouldn't need to reboot, just wait for a timeout (30s or 60s). I don't
see how it can block the system event queue, since OS/2 has the SIQ fix
for years now. I have always been able to switch away from it using
Ctrl-Esc or similar.
-------


Tue 04 Nov 2008 23:41 XZ/77.0 (WARP; i386)



Programmed by Dmitri Maximovich, Dmitry I. Platonoff, Eugen Kuleshov.
25.09.99 (c) 1999, RU/2. All rights reserved.
Rewritten by Dmitry Ban. All rights ignored.