Public SourceMage Discussion List

[Sergey A Lipnevich <sergeyli AT pisem.net>]

Correct me if I'm wrong but if it mentions "insn" stuff it's an old 
bison problem. Search the lists for "bison", or simply try the latest 
version of bison and its dependencies.

BTW, why don't we have a memtest86 spell? Any takers? I'm ready to 
commit it to `util' section :-).

Sergey.

Duane Malcolm wrote:

> Hi,
>
> Thanks for the help. I have had moderate success.
>
> Re: random errors in compilation, I did suspect the swap space wasn't 
> been mounted correctly but I didn't suspect the ram. I used memtest, 
> you are right it's a "sweet program". Thanks to you my system is now a 
> P4 1.4GHz, 256MB RAM, 80GB HD. I'm not sure if the 512MB stick is bad 
> or if my motherboard can't handle PC2700 RAM. I'm taking it back to 
> the shop today.
>
> Got the system-update and rebuild to work OK apart from xfsprogs (no 
> source), glibc (caught in a loop) and bzip2 (not sure).
>
> XFREE86 - what I have found.
> It still did not compile under gcc3.2. The error is below.
> I created a USEGCC2 file in xfree86 codex. But on compilation it used 
> the flags for gcc3.2 i.e. -march=pentium4 etc..., which caused 
> problems. I changed the defaultgcc2 flags in local/host.def file and 
> it still uses gcc3 flags. I had to change the flags in a "pentium4" 
> file under the var tree (I think it was in a archtype folder) for it 
> to start compiling. USEGCC2 didn't seem to pick up the gcc2 flags nor 
> read the host.def file I keep saving. Another perculiarity was that in 
> the gcc3 compilation of xfree86 I noticed the gcc3 flags were 
> repeated. Anyway I got xfree86 to compile using gcc2.
>
> The error from gcc3 was:
> mipmap.c: In function 'halveDimage_float':
> mipmap.c: 1270: unable to find a register to spill in class 'FLOAT_REGS'
> mipmap.c: 1270: this is the insn: (insn 111 109 114 (set (subreg:SF 
> (reg /v:DI 22 rxmm1 [73]) 0) (float : SF (reg:DI 0 rax [89]))) 170 
> {*floatdisf2_i387_only}(insn_list 108 (nil))
> (expr_list : REG_DEAD (reg:DI 0 rax [89])
> (nil))
> mipmap.c: 1270: confused by earlier errors, bailing out.
>
> Not being a c/c++ programmer I'm lost.
>
> Again thanks for your help, sorry I have no spare RAM to lend it's 
> broken, I did notice some kernel patch someone was writing to boot 
> with Badram. But this ram is one month old so I'm returning it. 
> Hopefully we can get this problem fixed, I am happy to do some 
> debugging/testing of the iso's if need be.
>
> Regards,
> Duane.
>
> Dufflebunk wrote:
>
> > Hello
> >
> > On Wed, 2002-11-20 at 14:48, Duane Malcolm wrote:
> >  
> >
> > > Hi
> > > I've been using Sorcerer since November 2001 and have been using the 
> > > same ISO from 2001. When the beta2 iso was release I tried to 
> > > upgrade my system but had a number of issue.
> > >
> > > To date I have I have managed to install SGML beta2 and rebuild it 
> > > (with only 3 problems left)
> > >
> > > Install problems: I tried to install XFS with the kernel, the kernel 
> > > would be patched but the kernel compilation would fail at random 
> > > places. At first I would remove the module that caused the compile 
> > > problem but it would fail somewhere else. So in the end I compiled 
> > > the kernel with the same configuration twice and found they failed 
> > > at different places. Finally I found that if I didn't compile XFS I 
> > > was fine. Now I'm using ext3.
> > >   
> >
> > You tried compiling twice with the same settings, and it failed in two
> > different places?? This is the hallmark of bad RAM. I would suggest you
> > check your ram with memtest86 (memtest86.com, sweet program).
> >
> >  
> >
> > > Rebuild problems:
> > > xfsprogs - can't find source version 2.2.2 - understandable, I 
> > > couldn't either. Didn't need it since I couldn't install XFS in the 
> > > kernel.
> > > ncurse - I understand this has been discussed - I will look through 
> > > the previous posts and fix it
> > >   
> >
> > Yeah, nasty little bugger. Basicly rm the ncurses.h in the gcc
> > directory.
> >
> >  
> >
> > > bzip2 - not sure why, will look into it.
> > >   
> >
> > This was a problem, I remember having it with a rebuild. I believe it
> > has been fixed in the latest test grimoire.
> >
> >  
> >
> > > But my biggest problem, the one that annoys me and the only one I 
> > > can't start using beta2, is installing xfree86. It has a problem 
> > > compiling 'mipmap.c'. I have searched the web for this issue, I have 
> > > downloaded the xfree86 source again and I tried to compile it after 
> > > my lastest rebuild, all with no luck. I did attempt to compile it 
> > > using gcc2 but I don't know how to set xfree86 to be compiled with 
> > > gcc2, can someone help me here? Unfortunately, I am unable to get 
> > > the compile log to my working computer, I'll write down the error 
> > > and post it if needed. I am really posting this to see if anyone 
> > > else has had problems or if anyone could help.
> > >   
> >
> > To force a program to be compiled with gcc2, you can add an empty file
> > called USEGCC2 to the spell's directory. Be sure to check your ram
> > before trying again... that could again be the culprit.
> >
> >  
> >
> > > I did reset my system setting in sorcery configuration after 
> > > installing beta2.
> > >   
> >
> > There is a mino bug with the iso which causes the optimization info not
> > to be save. After your reboot, you should probable set it again.
> >
> >  
> >
> > > Also just a question, are we able to get a xdelta patch between iso 
> > > releases?, I have a 56k modem.
> > >   
> >
> > I can't speak for the ISO, but I hope xdelta will make it's way back
> > into some of the large spells like mozilla, gcc, and X.
> >
> >  
> >
> > > My system is: Asus P4S333-VM motherboard, P4 1.4GHz CPU, 750MB RAM, 
> > > 80GB HD, Asus V8170 Graphics card.
> > >   
> >
> > Wanna lend me some of that RAM?
> >
> >  
> >
> > > Regards, Duane.
> > >
> > >
> > > _______________________________________________
> > > SM-Discuss mailing list
> > > SM-Discuss AT lists.ibiblio.org
> > > http://lists.ibiblio.org/mailman/listinfo/sm-discuss
> > >   
>
>
> _______________________________________________
> SM-Discuss mailing list
> SM-Discuss AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/sm-discuss