headers
Laurens Vets | 2 Jun 2005 11:39
Picon

Re: Browsers on NetBSD sparc

>>>See above. But since mozilla fails with a bus error here I guess
>>>it's still 64bit-related.
>>>
>>
>>Well, it could still be a sparc64 related issue but I can say for
>>certain that it runs fine on amd64 so I don't think it is a generic
>>64-bit problem (maybe an endian bug though :)
> 
> 
> Ok, I meant 64bit alignment constraint violations - SPARCs require
> natural alignment for all memory accesses, so the popular game of
> passing an int* for a size_t* is pretty likely to blow up with a bus
> error on sparc64, but these won't bite you ;)
> On the other hand, lots of gtk-applications like to cast ints to
> pointers and the other way around - this would likely bite all 64bit
> architectures, at least when running 64bit binaries.
> 
> I don't think there are many endianness bugs left in gecko - it runs
> fine on MacOS X for ages, same goes for Solaris. The PowerPC-specific
> code seems to blow up on macppc ( /weird/ rendering errors but no more
> crashes than on MacOS X ), on sparc64 I don't see any of these rendering
> problems. 
> 
> And about libfontconfig and friends - putting something like this into
> your mk.conf may help:
> USE_BUILTIN.fontconfig?=yes
> PREFER_fontconfig?=native
> 
> this should keep pkgsrc from building its own libfontconfig and make it
> use the one that comes with X. I guess the crashes come from compiling
(Continue reading)

Permalink | Reply | 14 comments |
headers
Michael Lorenz | 2 Jun 2005 13:49
Picon

Re: Browsers on NetBSD sparc

Hello,

> > And about libfontconfig and friends - putting something like this
> > into your mk.conf may help:
> > USE_BUILTIN.fontconfig?=yes
> > PREFER_fontconfig?=native
> > 
> > this should keep pkgsrc from building its own libfontconfig and make
> > it use the one that comes with X. I guess the crashes come from
> > compiling with the headers from one libfontconfig but linking
> > against the other for whatever reason. Of course you'd have to
> > recompile everything using libfontconfig so its usefulness may be
> > limited. Something similar for Xft2 and freetype2 would probably be
> > a good idea too. since these libraries depend on each other.
> 
> I added the above 2 variables to mk.conf. I'm now trying to rebuild 
> mozilla-gtk2.  However, upon installing Xft2, I get the following
> error now:
> 
> <...>
> checking for freetype-config... freetype-config
> checking for fontconfig >= 2.2... Requested 'fontconfig >= 2.2' but 
> version of F
> ontconfig is 1.0.1
> configure: error: Library requirements (fontconfig >= 2.2) not met; 
> consider adj
> usting the PKG_CONFIG_PATH environment variable if your libraries are
> in a nonstandard prefix so pkg-config can find them.
> *** Error code 1
>
(Continue reading)

Permalink | Reply | 13 comments |
headers
Laurens Vets | 2 Jun 2005 14:05
Picon

Re: Browsers on NetBSD sparc

>>>And about libfontconfig and friends - putting something like this
>>>into your mk.conf may help:
>>>USE_BUILTIN.fontconfig?=yes
>>>PREFER_fontconfig?=native
>>>
>>>this should keep pkgsrc from building its own libfontconfig and make
>>>it use the one that comes with X. I guess the crashes come from
>>>compiling with the headers from one libfontconfig but linking
>>>against the other for whatever reason. Of course you'd have to
>>>recompile everything using libfontconfig so its usefulness may be
>>>limited. Something similar for Xft2 and freetype2 would probably be
>>>a good idea too. since these libraries depend on each other.
>>
>>I added the above 2 variables to mk.conf. I'm now trying to rebuild 
>>mozilla-gtk2.  However, upon installing Xft2, I get the following
>>error now:
>>
>><...>
>>checking for freetype-config... freetype-config
>>checking for fontconfig >= 2.2... Requested 'fontconfig >= 2.2' but 
>>version of F
>>ontconfig is 1.0.1
>>configure: error: Library requirements (fontconfig >= 2.2) not met; 
>>consider adj
>>usting the PKG_CONFIG_PATH environment variable if your libraries are
>>in a nonstandard prefix so pkg-config can find them.
>>*** Error code 1
>>
>>Stop.
>>make: stopped in /usr/pkgsrc/fonts/Xft2
(Continue reading)

Permalink | Reply | 8 comments |
headers
Michael Lorenz | 2 Jun 2005 14:28
Picon

Re: Browsers on NetBSD sparc

Hello,

> Hmm, so this won't work on sparc?  Hence, me trying to rebuild 
> mozilla-gtk2 is actually fruitless? :)

It would probably work if you had XFree86. 
The fontconfig package should fall back to install itself when it
doesn't find a native version supplied by XFree86. These weren't
intended for sparc anyway, sorry if I didn't make that clear enough.

> Not being able to run a 'decent' browser on sparc kinda sucks :/

Mozilla should build and work without these hacks, although I never
tried because the only 32bit sparc I have here is a 110MHz microSPARC
with 64MB RAM - not enough for a beast like Mozilla - it would crawl,
swap like hell and even if the box had 256MB or so it would still not
be fast. Compiling would take days.
If you have loads of RAM and fast Hyper- or TurboSPARC CPU(s) it may be
worth a try though.

have fun
Michael

PS: for the records - the mozilla-gtk2 built with native Xft2, xfree2
and fontconfig on macppc doesn't seem to show any of the weird rendering
problems I had before. No idea if that's because of newer patches
though.
Permalink | Reply | 7 comments |
headers
Laurens Vets | 2 Jun 2005 14:40
Picon

Re: Browsers on NetBSD sparc

Hi,

>>Hmm, so this won't work on sparc?  Hence, me trying to rebuild 
>>mozilla-gtk2 is actually fruitless? :)
> 
> It would probably work if you had XFree86. 
> The fontconfig package should fall back to install itself when it
> doesn't find a native version supplied by XFree86. These weren't
> intended for sparc anyway, sorry if I didn't make that clear enough.

I'm new to NetBSD, it seems to be the only open source OS running 
decently on my system (SS20, Dual HyperSparc).  Didn't know it uses a 
sparc specific X.

>>Not being able to run a 'decent' browser on sparc kinda sucks :/
> 
> Mozilla should build and work without these hacks, although I never
> tried because the only 32bit sparc I have here is a 110MHz microSPARC
> with 64MB RAM - not enough for a beast like Mozilla - it would crawl,
> swap like hell and even if the box had 256MB or so it would still not
> be fast. Compiling would take days.

Compiling indeed takes days ;)  It compiles and installs perfectly, 
however, when it starts, I get a segmentation fault.  That's what 
started this whole discussion.

> If you have loads of RAM and fast Hyper- or TurboSPARC CPU(s) it may be
> worth a try though.

I'm trying, it doesn't work though :)
(Continue reading)

Permalink | Reply | 6 comments |
headers
Bruce O'Neel | 2 Jun 2005 14:43
Picon
Favicon

Re: Browsers on NetBSD sparc

Hi,

You might try adding

/usr/pkg/lib

to the beginning of LD_LIBRARY_PATH.

Also make sure that you rm -rf ~/.mozilla before you start if you've run
firefox and/or mozilla sometime in the past.

cheers

bruce

On Thu, Jun 02, 2005 at 02:40:25PM +0200, Laurens Vets wrote:
> Hi,
> 
> >>Hmm, so this won't work on sparc?  Hence, me trying to rebuild 
> >>mozilla-gtk2 is actually fruitless? :)
> >
> >It would probably work if you had XFree86. 
> >The fontconfig package should fall back to install itself when it
> >doesn't find a native version supplied by XFree86. These weren't
> >intended for sparc anyway, sorry if I didn't make that clear enough.
> 
> I'm new to NetBSD, it seems to be the only open source OS running 
> decently on my system (SS20, Dual HyperSparc).  Didn't know it uses a 
> sparc specific X.
>
(Continue reading)

Permalink | Reply | 2 comments |
headers
Michael Lorenz | 2 Jun 2005 14:52
Picon

Re: Browsers on NetBSD sparc

Hello,

> I'm new to NetBSD, it seems to be the only open source OS running 
> decently on my system (SS20, Dual HyperSparc).  Didn't know it uses a 
> sparc specific X.

It uses Xsun or Xsun24, you can install XFree86, it should support
whatever graphics adapter you're using ( probably unaccelerated though )
but getting it to work is not quite trivial. I did it for my SPARCbook,
mainly to write a driver for the graphics chip but because one of the
RAM modules has problems I won't be able to do anything about that any
time soon.

> >>Not being able to run a 'decent' browser on sparc kinda sucks :/
> > 
> > Mozilla should build and work without these hacks, although I never
> > tried because the only 32bit sparc I have here is a 110MHz
> > microSPARC with 64MB RAM - not enough for a beast like Mozilla - it
> > would crawl, swap like hell and even if the box had 256MB or so it
> > would still not be fast. Compiling would take days.
> 
> Compiling indeed takes days ;)  It compiles and installs perfectly, 
> however, when it starts, I get a segmentation fault.  That's what 
> started this whole discussion.

Hmm, can you trace it in gdb? That's not for the faint-hearted though. 
( look at the script you run to start mozilla - should be in
/usr/pkg/bin/mozilla. It will set a bunch of environment variables and
then run mozilla-bin. You'd have to set up the variables and then run
mozilla-bin in gdb - this should at least tell you where exactly it
(Continue reading)

Permalink | Reply | 2 comments |
headers
Andy Wallis | 4 Jun 2005 00:50
Picon
Favicon

NVRAM replacement on sun4c machines

This should add a few more datapoints for the NVRAM replacement FAQ and
generail sun4c recovery.

I have put a pair of Maxim (formally Dallas Semiconductor) DS1642-70s into 
a SLC and IPC. Both NVRAM replacments work without any problems in their
respective machines.

This should help remove the problem that the changes made to the SGS-Thompson
(formerly Mostek) M48T02 chips had on the sun4c machines.

-Andy Wallis
Permalink | Reply | 0 comments |
headers
Tom Uban | 4 Jun 2005 01:01

Re: NVRAM replacement on sun4c machines

I've had good luck with removing the battery from the top of a dead NVRAM
and attaching a CR2032 socket so that the battery is replaceable with a
standard PC cell.

--tom

At 06:50 PM 6/3/2005 -0400, Andy Wallis wrote:

>This should add a few more datapoints for the NVRAM replacement FAQ and
>generail sun4c recovery.
>
>I have put a pair of Maxim (formally Dallas Semiconductor) DS1642-70s into
>a SLC and IPC. Both NVRAM replacments work without any problems in their
>respective machines.
>
>This should help remove the problem that the changes made to the SGS-Thompson
>(formerly Mostek) M48T02 chips had on the sun4c machines.
>
>-Andy Wallis
>
Permalink | Reply | 0 comments |
headers
Tom Uban | 4 Jun 2005 01:05

Re: NVRAM replacement on sun4c machines

I've had good luck with removing the battery from the top of a dead NVRAM
and attaching a CR2032 socket so that the battery is replaceable with a
standard PC cell.

--tom

At 06:50 PM 6/3/2005 -0400, Andy Wallis wrote:

>This should add a few more datapoints for the NVRAM replacement FAQ and
>generail sun4c recovery.
>
>I have put a pair of Maxim (formally Dallas Semiconductor) DS1642-70s into
>a SLC and IPC. Both NVRAM replacments work without any problems in their
>respective machines.
>
>This should help remove the problem that the changes made to the SGS-Thompson
>(formerly Mostek) M48T02 chips had on the sun4c machines.
>
>-Andy Wallis
>
Permalink | Reply | 0 comments |

Gmane