firefox

[vinceASPECT]

Hi,

is it possible to have the latest firefox inside TCL 32 bit........think it's about
version 64 now?

i also have another question...i see TCL does not specify that a cpu
must have SSE2 instruction sets...

486 is the cpu it starts at....

i wonder how computers of this type deal with applications that are compiled
with SSE2 or does non of this topic apply to linux archictectures?

V

[Rich]

Hi vinceASPECT

is it possible to have the latest firefox inside TCL 32 bit........think it's about
version 64 now?

That's what  firefox_getLatest.tcz  is for.

i also have another question...i see TCL does not specify that a cpu
must have SSE2 instruction sets...

486 is the cpu it starts at....

i wonder how computers of this type deal with applications that are compiled
with SSE2 or does non of this topic apply to linux archictectures?

They don't. If you try to run an SSE2 enabled application on a 486, it will give you an  "Illegal instruction"  message and exit.

[vinceASPECT]

Right Rich,

thanks for replying fast.....


so while a 486 cpu with math coprocessor will
run TCL.....the repo for the version of TCL used would likely not show any extensions that the processor could not run......right

So any extension that was compiled with SSE2 would not show up
on  a TCL machine containing a lower grade CPU than that.....say SSE1 cpu.

hmmm

thx

V

[Rich]

Hi vinceASPECT
While all extensions are supposed to be 486 compatible, there have been occasions where something compiled for a higher
processor has slipped through.

[vinceASPECT]

i am confused perhaps with compiling and CPU features.

are we saying here, for example, that the latest version of firefox will
still run on a 486 because it was comiled for that lowest cpu and up ?

you seem to be saying that TCL and it's extensions don't need a cpu
with any SSE or SSE2 capability........right.

that is a nice thing if true...

thanks

Vince.

[Rich]

Hi vinceASPECT
That's correct, all extensions are supposed to be 486 compatible.

[Juanito]

Using "-march=i486 -mtune=i686" in CC/CCX or CFLAGS/CXXFLAGS will usually take care of compiling for i486 processors when using configure/make.

When using cmake, more care needs to be taken.

This being said, some source code (for example, gmp) runs its own checks and inserts the actual processor used to compile instead of respecting the flags used. Still other source code (I can't think of an example at the moment) has the processor hard coded.

Finally, some source code (for example, mesa/Xorg-7.7-3d) will not compile for i486, so a different minimum processor has to be used.

[coreplayer2]

Hello

The latest fast Firefox v57.0.3

Code: [Select]

/tmp/tcloop/firefox/usr/local/firefox/firefox: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux.so.2, for GNU/Linux 2.6.18, BuildID[sha1]=60d514e7a6d8f1901cf6848d5f2c33d2692b10cd, stripped
AIUI Firefox is compatible with 386 and up, but I don't have anything less than a 686 machine to test on

Given Juanito's previous input i'm wondering now if file is just reporting the host compile machine arch. (though i doubt a 386 was used to compile firefox on)

[Juanito]

I believe that file uses 80386 to mean x86 in this context - you'd need to use an actual i486 or something like bochs to test.