gcc 2.8.1

gcc 2.8.1

Post by Matthias Halfman » Wed, 21 Oct 1998 04:00:00



 I have heard that gcc 2.8.1 has quit a few bugs and I have seen some of
these (i think anyway). For instance when ever I recompile my kernel
with gcc 2.8.1 I can't get X to work, also with 2.8.1 I can't compile
the X source. Are there patches for these bugs? Thanks

matt

 
 
 

gcc 2.8.1

Post by Nathan Mye » Wed, 21 Oct 1998 04:00:00



> I have heard that gcc 2.8.1 has quit a few bugs and I have seen some of
>these (i think anyway). For instance when ever I recompile my kernel
>with gcc 2.8.1 I can't get X to work, also with 2.8.1 I can't compile
>the X source. Are there patches for these bugs?

Yes, gcc-2.8.1 is buggy, but that's not why you can't get X to work.  
Those are kernel bugs.  The fix is twofold:

  1. Get rid of gcc-2.8.x.  It's dead.  
     Egcs-1.x is the actively maintained GNU compiler line.

  2. Switch to the 2.1.x series kernels, which have been adapted
     to work with the Egcs compiler (as well as with gcc-2.7).

Alternatively:

  1. Go back to gcc-2.7.2.3, and stay with the 2.0.x kernel.
     The kernel bugs have avoided tickling it.

--
Nathan Myers


 
 
 

gcc 2.8.1

Post by Alain.Bo.. » Thu, 22 Oct 1998 04:00:00




>> I have heard that gcc 2.8.1 has quit a few bugs and I have seen some of
>>these (i think anyway). For instance when ever I recompile my kernel
>>with gcc 2.8.1 I can't get X to work, also with 2.8.1 I can't compile
>>the X source. Are there patches for these bugs?
> Yes, gcc-2.8.1 is buggy, but that's not why you can't get X to work.  
> Those are kernel bugs.  The fix is twofold:

Not THAT buggy...
The problem is in the kernel: a code block was written as a turn around
for a bug in gcc 2.7.x; with 2.8 the bug is gone and now the code in the
kernel no longer works properly! Ain't life fun?

Quote:>   1. Get rid of gcc-2.8.x.  It's dead.  

No it's not; it's just progressing at the usual gcc speed: no publicity,
cathedral development.

Quote:>      Egcs-1.x is the actively maintained GNU compiler line.

Egcs is a independent branch of gcc started by Cygnus Corp. using a
bazaar development model (thus faster, but many people start to
use unstable snapshots with unpredictable results; stable releases
are fine).

Quote:>   2. Switch to the 2.1.x series kernels, which have been adapted
>      to work with the Egcs compiler (as well as with gcc-2.7).

What makes you say that 2.1.x kernels are less buggy than gcc 2.8.1?
2.1.x are development versions whereas gcc 2.8.1 is supposed to be
a stable well-tested release.

Quote:> Alternatively:
>   1. Go back to gcc-2.7.2.3, and stay with the 2.0.x kernel.
>      The kernel bugs have avoided tickling it.

Or use kernel 2.0.35; I guess the problem should be cured there and
you can stick with stable kernels...

Hope this helps!

--
Alain Borel

 
 
 

gcc 2.8.1

Post by Frank Sweetse » Thu, 22 Oct 1998 04:00:00



> >   2. Switch to the 2.1.x series kernels, which have been adapted
> >      to work with the Egcs compiler (as well as with gcc-2.7).

> What makes you say that 2.1.x kernels are less buggy than gcc 2.8.1?
> 2.1.x are development versions whereas gcc 2.8.1 is supposed to be
> a stable well-tested release.

actually, this is a good idea.  along with it's bugs, gcc 2.8 also got
stricter about certain optimizations in ways that broke fragile or outright
buggy code in the kernel.  while linus has been unwilling to accept
signifigant changes to fix these in 2.0, since 2.0 is supposed to be the
stable line, he's been much more willing to accept fixes for the 2.1
kernels.  for example, the ioport.c problem that caused X to stop working
has been fixed in the kernel in 2.1

--
Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
paramount.ind.wpi.edu RedHat 5.1   kernel 2.1.125   i586 | at public servers
That wouldn't be good enough.

 
 
 

gcc 2.8.1

Post by Rudolf Leitg » Thu, 22 Oct 1998 04:00:00




Quote:>>   1. Go back to gcc-2.7.2.3, and stay with the 2.0.x kernel.
>>      The kernel bugs have avoided tickling it.
> Or use kernel 2.0.35; I guess the problem should be cured there and
> you can stick with stable kernels...

As far as I know, the kernel maintainers in charge have repeatedly
stated that they will not adapt the 2.0.x kernel line for gcc 2.8.x
or egcs. They basically say that you are welcome to try it but don't
bother them with bug reports. However, I have found a web page with
a list of patches to the 2.0.x kernels, so you can compile them with
gcc 2.8.x or egcs. The author claims that his patched kernels are
rock solid. Check it out at:

http://www.suse.de/~florian/kernel+egcs.html

Cheers

Rudi

--

         | | | | |
       \   _____   /      
          /     \                      B O R N
      -- | o   o |  --                   T O
      -- |       |  --                S L E E P
      -- | \___/ |  --                   I N
          \_____/                   T H E   S U N
        /          \    
         | | | | |

 
 
 

gcc 2.8.1

Post by Nathan Mye » Thu, 22 Oct 1998 04:00:00




>>   Egcs-1.x is the actively maintained GNU compiler line.
>>   Get rid of gcc-2.8.x.  It's dead.  

>No it's not; it's just progressing at the usual gcc speed: no publicity,
>cathedral development.

Is this meant ironically?  It has no developers, they have all
jumped ship and are working on Egcs.  Kenner might fold some
code back into gcc from egcs, but that's becoming increasingly
difficult as the code bases diverge.  If there is ever a gcc-3
it will be based directly on the egcs codebase.

--
Nathan Myers

 
 
 

gcc 2.8.1

Post by Christopher Brow » Fri, 23 Oct 1998 04:00:00





>>   1. Get rid of gcc-2.8.x.  It's dead.  

>No it's not; it's just progressing at the usual gcc speed: no publicity,
>cathedral development.

>>      Egcs-1.x is the actively maintained GNU compiler line.

>Egcs is a independent branch of gcc started by Cygnus Corp. using a
>bazaar development model (thus faster, but many people start to
>use unstable snapshots with unpredictable results; stable releases
>are fine).

Consider that various people that used to work on GCC are now working on
EGCS.

It is not beyond the realm of possibility that the progression of
development of GCC 2.8.x has diminished as a result...

--
"Problem solving under linux has never been the circus that it is under
AIX." (By Pete Ehlke in comp.unix.aix)

 
 
 

gcc 2.8.1

Post by Christopher Brow » Fri, 23 Oct 1998 04:00:00





>>   1. Get rid of gcc-2.8.x.  It's dead.  

>No it's not; it's just progressing at the usual gcc speed: no publicity,
>cathedral development.

>>      Egcs-1.x is the actively maintained GNU compiler line.

>Egcs is a independent branch of gcc started by Cygnus Corp. using a
>bazaar development model (thus faster, but many people start to
>use unstable snapshots with unpredictable results; stable releases
>are fine).

Consider that various people that used to work on GCC are now working on
EGCS.

It is not beyond the realm of possibility that the progression of
development of GCC 2.8.x has diminished as a result...

--
"Problem solving under linux has never been the circus that it is under
AIX." (By Pete Ehlke in comp.unix.aix)

 
 
 

gcc 2.8.1

Post by Alain.Bo.. » Fri, 23 Oct 1998 04:00:00



> Consider that various people that used to work on GCC are now working on
> EGCS.
> It is not beyond the realm of possibility that the progression of
> development of GCC 2.8.x has diminished as a result...

Point taken. That's quite possible indeed; by the way I use egcs myself
and did not intend to start a gcc/egcs flamewar. Funny how one can
disregard data that does not fit one's world view ;-)
Regarding the kernel issue, I just meant that for some people it would
be safer to use stable 2.0.x than to try the development series (even
if the late 2.1.12x seem pretty good to me).
Thanks to all those who helped me tear off the veils of Illusion!
--
Alain Borel

 
 
 

gcc 2.8.1

Post by Ronald Col » Fri, 23 Oct 1998 04:00:00






> >>   1. Get rid of gcc-2.8.x.  It's dead.  

> >No it's not; it's just progressing at the usual gcc speed: no publicity,
> >cathedral development.

> >>      Egcs-1.x is the actively maintained GNU compiler line.

> >Egcs is a independent branch of gcc started by Cygnus Corp. using a
> >bazaar development model (thus faster, but many people start to
> >use unstable snapshots with unpredictable results; stable releases
> >are fine).

> Consider that various people that used to work on GCC are now working on
> EGCS.

> It is not beyond the realm of possibility that the progression of
> development of GCC 2.8.x has diminished as a result...

You can say that again!!  Whoops, I think you did...  ;)

--
Forte International, P.O. Box 1412, Ridgecrest, CA  93556-1412

President, CEO                             Fax: (760) 499-9152
My PGP fingerprint: 15 6E C7 91 5F AF 17 C4  24 93 CB 6B EB 38 B5 E5

 
 
 

gcc 2.8.1

Post by Christopher Brow » Fri, 23 Oct 1998 04:00:00





>> Consider that various people that used to work on GCC are now working on
>> EGCS.

>> It is not beyond the realm of possibility that the progression of
>> development of GCC 2.8.x has diminished as a result...
>Point taken. That's quite possible indeed; by the way I use egcs myself
>and did not intend to start a gcc/egcs flamewar. Funny how one can
>disregard data that does not fit one's world view ;-)

No flames assumed...

I'll be interested to hear if RMS comments on this matter at the ALS in
Atlanta.  I am very much hoping that EGCS developments will be, at some
level, be "blessed" by the FSF so as to fold the code bases back
together.

RMS made the (possibly "flameworthy") comment not long ago that "There
is only one True GCC." (That was about all his message said, and I don't
think I'm taking it out of context...)  Hopefully that can *become*
true...

Quote:>Regarding the kernel issue, I just meant that for some people it would
>be safer to use stable 2.0.x than to try the development series (even
>if the late 2.1.12x seem pretty good to me).

Quite true.  

For many, many people, there's not forcibly a need to use the "latest
and greatest" when even the 1.0.x kernels provide substantially useful
functionality.  

1.0.x may not support some of the newer "stuff" quite as readily
(ELF/EGCS comes to mind, and further maturing of library/binary
codebases may have smoothed this...), but still represents a pretty
robust "almost-like-UNIX" OS that can satisfy many needs.

--
"Once upon a time there was a DOS user who saw Unix, and saw that it was
good. After typing cp on his DOS machine at home, he downloaded GNU's
unix tools ported to DOS and installed them. He rm'd, cp'd, and mv'd
happily for many days, and upon finding elvis, he vi'd and was happy.
After a long day at work (on a Unix box) he came home, started editing
a file, and couldn't figure out why he couldn't suspend vi (w/ ctrl-z)


 
 
 

gcc 2.8.1

Post by Christopher Brow » Wed, 28 Oct 1998 04:00:00





>>   1. Get rid of gcc-2.8.x.  It's dead.  

>No it's not; it's just progressing at the usual gcc speed: no publicity,
>cathedral development.

>>      Egcs-1.x is the actively maintained GNU compiler line.

>Egcs is a independent branch of gcc started by Cygnus Corp. using a
>bazaar development model (thus faster, but many people start to
>use unstable snapshots with unpredictable results; stable releases
>are fine).

Consider that various people that used to work on GCC are now working on
EGCS.

It is not beyond the realm of possibility that the progression of
development of GCC 2.8.x has diminished as a result...

--
"Problem solving under linux has never been the circus that it is under
AIX." (By Pete Ehlke in comp.unix.aix)

 
 
 

gcc 2.8.1

Post by Petri Kaukasoi » Wed, 28 Oct 1998 04:00:00



>I am very much hoping that EGCS developments will be, at some level, be
>"blessed" by the FSF so as to fold the code bases back together.

>RMS made the (possibly "flameworthy") comment not long ago that "There
>is only one True GCC." (That was about all his message said, and I don't
>think I'm taking it out of context...)  Hopefully that can *become*
>true...

RMS, 13 Oct 1998: "I'm open to having the EGCS team become the GCC
maintainers and making EGCS the official version of GCC."

The message is archived in http://www.cygnus.com/ml/egcs/1998-Oct/0439.html

 
 
 

1. RH6.0 - i386-glibc20-linux-gcc or i386-redhat-linux-gcc

I notice that the RH6.0 has two gcc compilers in

  /usr/bin/i386-glibc20-linux-gcc
  /usr/bin/i386-redhat-linux-gcc

and the default one is /usr/bin/i386-redhat-linux-gcc. Anyone
know what is different between them? Because I have found some
programs fail to be compiled using the redhat one but successful
if using /usr/bin/i386-glibc20-linux-gcc.

Thanks

david

2. Matrox or ATI - Millinia or Mach 64

3. gcc 2.96 vs. gcc 3.2: namespace, template incompatability

4. Has the atapi cd problem been solved

5. gcc 4.1.1 build under Intel Solaris 10 does not build gcc

6. Get rid of slidebutton in enlightenment themes! (DR16)

7. HOWTO build/install GCC 3.1 from SRPM w/o replacing GCC 2.96 ?

8. where is my mitsumi-cdrom ?

9. gcc-2.96/RH7,3 .vs gcc-3.2/RH8.0 portability issues

10. Upgrade gcc 2.96-98 to gcc 3.2-7 on Redhat 7.2

11. gcc 2.8.1 vs gcc 2.95.3 optimization on Sparc V8

12. GCC-2.8.0 or GCC-2.8.1(alpha?)

13. Patch for kernel with new gcc (was Re: GCC-2.7.1...)