Re: Internal Error, crib

Re: Internal Error, crib

[Gerald Pfeifer]

On Fri, 16 Jul 1999, V Lakshman wrote:
> I tried to send this with a nice preamble and MIME-encoded attachment
> from my regular mailing system (pine on enterprise.nssl.noaa.gov) and
> found that the egcs mailer wouldn't accept my mail, telling me to
> contact my system administrator. My sys-admin said the ORBS database
> was nothing more than net terrorism and flatly refused to do anything
> about it.

Fire him. He is completely clueless.

That system is not configured correctly, that's all.

> I had to try several machines until I chanced on one (doplight) that
> works.

doplight.nssl.noaa.gov is not configured correcty, either.
It will shortly be listed by ORBS as well.

> God knows how many bugs don't get reported because the egcs group has
> chosen to make it so hard.

Again, it's your systems that are not configured correctly.

We, a bunch of volunteers, try to fix bugs reported to us as quickly as
possible. Is it too much to ask that people configure their mail servers
correctly, at least once they have been notified of the problem, instead
of blaming us?

Gerald
-- 
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/

Re: Internal Error, crib

[craig]

>>#> This file decoded to complete and utter garbage on my system, rather
>>#> than text.
>
>Sorry about this. The file was uuencoded from a gzipped-file.
>The internal compiler error originated from egcs, C++ compiler v. 2.91.6
>and originated with an assignment operator in algo_base.h

Thanks for the clarification!  Knowing the system might help as well.

If you uuencode a gzipped file in the future, you might try naming
it `bug_report.gz', as the .gz suffix warns the reader ahead of time
what is coming.

We're still learning just what to recommend to users regarding how
to package up bug reports, and how to accommodate large bug reports,
so please be patient with us.

In the meantime, I just thought you'd like to know how to increase
the likelihood that your bug reports will be seen by people who can
actually do something about it.

>>#> Further, your subject header gave no indication as to *what* gave an
>>#> internal error, what kind of system you were using, etc.
>
>Well, there is a reason for this.  I tried to send this with a nice
>preamble and MIME-encoded attachment from my regular
>mailing system (pine on enterprise.nssl.noaa.gov) and found that the
>egcs mailer wouldn't accept my mail, telling me to contact my system
>administrator. My sys-admin said the ORBS database was nothing more than
>net terrorism and flatly refused to do anything about it.

MIME is a bit of a hassle for a few of us (well, maybe just me, and
hopefully that's temporary), but it does seem to be preferred by many
here.  I get the impression some anti-spam/anti-virus filters reject
MIME attachments they don't recognize, but I don't think Cygnus uses
those for egcs.cygnus.com.

ORBS is an interesting situation.  My ISP (world.std.com) said something
similar about it earlier this year when I found my email bouncing from
egcs.cygnus.com mailing lists.  I had researched ORBS only a bit myself,
but learned it *had* a history of being badly managed but, at the time,
presented itself as being much more sanely handled, predictable, etc.

So I asked my ISP to look into the issue again, and consider contacting
Cygnus Support, which is hardly a clueless organization when it comes
to such issues -- suggesting that, if ORBS was indeed problematic,
Cygnus would like to know how, and if it wasn't, Cygnus might help my
ISP why not.

Within a week or so, my ISP had fixed its open relays, got quickly delisted
from ORBS, and I was back in business.

>By this time, I had lost patience and was just trying to somehow get
>through.

I can certainly understand that!

>God knows how many bugs don't get reported because the egcs group has
>chosen to make it so hard.

We're (well, Cygnus is) mainly trying to make it hard for spammers to get
through to thousands of productive people.  The success rate is quite
high -- seems like I see only one or two spams via egcs.cygnus.com per
week, compared to several times that from other sources per day.  Much
better than back when egcs@cygnus.com (basically spam-filter-free) was
the list.

But, as with any war, there can be collateral damage, ranging from
friendly fire (accidentally bouncing email due to a misconfigured
spam filter) to the inevitable perils from civilians refusing to
distinguish themselves from enemy soldiers (a sysadmin who refuses to
find out whether he can get his systems off of the ORBS listing --
though that's more akin to a school teacher refusing to lead his
classes out of a military barracks on the grounds that targeting
bombs at such barracks without first verifying there are *only*
enemy soldiers inside constitutes "terrorism" on the part of the
bombers).

        tq vm, (burley)
>From ronis@ronispc.chem.mcgill.ca Fri Jul 16 13:23:00 1999
From: David Ronis <ronis@ronispc.chem.mcgill.ca>
To: egcs-bugs@egcs.cygnus.com
Subject: ICE in bootstrap of gcc version 2.96 19990716 (experimental)
Date: Fri, 16 Jul 1999 13:23:00 -0000
Message-id: <199907162023.QAA04813@ronispc.chem.mcgill.ca>
X-SW-Source: 1999-07/msg00630.html
Content-length: 6561

I get the following ICE when doing a full bootstrap build of the latest CVS:

/home/ronis/Project/objdir/gcc/xgcc -B/home/ronis/Project/objdir/gcc/
-B/usr/i586-pc-linux-gnulibc1/bin/ -fgnu-runtime -c -o linking.o
-I. -I../../../egcs/libobjc -O3 -march=pentium -funroll-loops
-fomit-frame-pointer -fstrict-aliasing -DIN_GCC
-I../../../egcs/libobjc/objc -I../../../egcs/libobjc/../gcc
-I../../../egcs/libobjc/../gcc/config -I../../gcc
-I../../../egcs/libobjc/../include ../../../egcs/libobjc/linking.m
xgcc: Internal compiler error: program cc1obj got fatal signal 11
{ronispc:158} /home/ronis/Project/objdir/gcc/xgcc
-B/home/ronis/Project/objdir/gcc/ -B/usr/i586-pc-linux-gnulibc1/bin/
-fgnu-runtime -c -o linking.o -I. -I../../../egcs/libobjc -O3
-march=pentium -funroll-loops -fomit-frame-pointer -fstrict-aliasing
-DIN_GCC -I../../../egcs/libobjc/objc -I../../../egcs/libobjc/../gcc
-I../../../egcs/libobjc/../gcc/config -I../../gcc
-I../../../egcs/libobjc/../include --save-temps
../../../egcs/libobjc/linking.m xgcc: Internal compiler error: program
cc1obj got fatal signal 11


The --save-temps was added by me.  A linking.mi file was created, as
was a linking.s file (which is empty).  A uuencoded version of the
gzipped linking.mi is appended below.

I'm running on an i586-pc-linux-gnulibc (linux-2.2.10) and built
gcc-2.96 19990707 successfully.

David

-------------------------------------------------------------
begin 644 linking.mi.gz
M'XL("*Z3CS<"`VQI;FMI;F<N;6D`G5MM<]RV$?XL_@HDGG;NSB=+.KE.8B:9
MI&[2>AK':9S)=,:3<B@2=\<<CV0(4HKR\M^[V`5(``1YE#0>'X'=?;!8++`+
M$'S"KMB'SYY=T#^^2\1%GMV4-S\G\%L<LF+W[/AAP`+?GZI],@Z!_[V]^9DG
MS;/]A^S*#]17GH*2_\T"ZL`N]N617]1ED8F+[^I2:B)1TJR^V"7)158D>9OR
M"]&D*=\B-+MVT9A^9GTMPP9>/*:%:QMEI!,L&+%ZT-Q7'*!85C2L:NHTVVZC
M)G0!QN6]8&TALEW!4T05V6_<@O0H>`*<S6@S+XL=MG>7[.-ZO,$1+=D=_*^E
MJ+M/V/751X\<DH'.K'>DZ^>SW'+32PTTEEUD[.]OWWX3]NB:*RD+`69OZC9I
MF$2+!,]!Z[)FP>\!8[=EEK(5U$59&D*9^!$2:R6."(,_V>K=5Q(_*V#R<@",
MFRS!1@/)QG]A"V!@XFK-\'>S1/ALRQ;BBGWV&;MD?_P!U?BX!`IC-6_:NF!$
M%AO9.L\%MVGGGY-NQ*-+H)#JJ.ZGV<,2EP4FVS>KDSP68@5=E+]15<(8\QJ[
M)@%MLT&#B]7K-]\M%UF*/5JS9\^>+</.]SO[HJ7>_?!]R*;_+"?P:8UJL=4;
MWL2OY&,XQ:4XAI3?J3'&.AQ#![OGMG)>4XFVXC650J<WEJ^L.D(1'[G7%,"/
ML]+YN^6UR,HB]/%W'FX(9L6V#$?LZVT@`QWC(N&17'KL/KN]SFYAK<@ST:R8
M?!3AL(U@*'7DS;Y,E1RCD@A=6RD)$==U?,_07FD3W^1>8P7XSS,@HKVA,92M
MA8Z(CS^[@>FZ4R/H:E[595,F98YH;*6+(CSK0>7ZL&([/:EP/8`Y$V(0?'XY
M9_$*@B_(-[]3#83&#"!\F.ZW,2PVX=D9LTB+55Q5^7W4+!>RN`P->EN`X^`L
MITD8U[L(HE:H*V2YYCOQ7HY\"0L1>NKR)YSR0)2]QB;/Y!#!9%?C*7917I:'
MMH*Y#YHE/`,GI66MK)9AOW1?SDQ.-J?3#^P3C`V/C_.RD%.`^UCL'X#TJ&1&
M1J_-XV*B#*>S]-^8VCH+8A(G>QX59<H#8ZWO:]FJX+\VTF6ZQ8K"W8'?AS3J
M700$]VLY^0)XAY0F7S)-Y<T5%BNI:K1MBP0G!SDJ6ZVM%I?AL!<DG93'*JZY
M"6`*#F'</,#L-MI!3EO=`=45O=($:J$<Y&2A6J8L0BMX&MHKL:P^QN(0$LX0
M#"9Z$]VTR8$W)H]M([FV=#6A4GE@"*L&)RT-;3<N,+@<DH2NEAW+M,UYA-#8
MY;4*>7T-+H)!T,L@J>!W;#$PRCHX.[-4);6ZHJ9/:[[$#OI#8A#@N")B"GE9
MPV&5ZE3#I]YS>M8X34T^LDNU=GU\;?JV5@/!>JB:'\O;8:L#+-.!@Z!S+V6]
M7QL/0L<D'Y0\X:$4:A5MRSH"^-D*R(R36LV$%(RR`D?W@3W`!Y7$JI36'/X`
M6Y!8IV%QSJE\=;&@#<Y2=NF"Z<"C)N^2_95`SC_'.:23V'$]&"D"4QRB^%Q=
M+`!0C$'&';KU"8"H^B[6=P]W^RP'GUBIB"F1,<YB/]G_/F,&Y>E3]NFG$HZ@
M"'<A?YZRJR7[BVL$R?9GT)M,(GK-8EL%C:&G%5A`,&N9/%PYQMA8X_+!XG#%
MSAG4SD`G:\]L0&YP#KB#D37&UN5*[V<,%MP#'3Q[H$O?YN<#4"0Y5E)XW:M^
M.G0ZN83./BY:49O!."MU')^7(9CB6QZ#BEQ,(LP[)K"X[/DY.,I@\^`\W;T7
M%PD$RU,*/ZZK('3]8D(,\YW-\RF=-MW9D'D:](3][7)*ZEK%5Y]&EZ<D/2:W
MCD_2LH5XR:)(EB(JX9D(@&\^/@7./.-I6^SYA,64.<C5_8[;L;@<,!=.N':O
MT%`Z^F<$TWR;[3H$=OJD20KE97*0YADDB6)?UHUD2?FMR="=40%I6Y5B6GB7
MI5X&W7Y6E-,`1\P#ISCD4G@8:`B$DHX!S;P5:BO2R$EH):$_.1Q5%XJ"-XZ\
M%:"0Z3?NTZ?)CB>ZTMK6,DW=G0@Z6@NW-3P(1`EU&&CN2V$'#Y3;F#;BED<8
M)P\*`!X^CJ"1*&X@M-RT#8^BQ2*B(8G@X3^OHVBY7(8CIY'8H0=CV'J0&E<O
M)C'^-:F'4N.A&#X]KC>3&._FZ/%0#)\>+YY/8OQCCAYS,1[_1T<-_D7.#AP4
M,T97Q&MZM7`]11^/C)]\,B5W.D@S?Y7:OT:OWT:TAK+^7%H=Y]"\EC;?9CQ/
MKSQU&SJ48A;>S^VQBN0^U*C[^O4W7X46US&N#[PVCXN-VA7N;L)@@`$4<=-N
M0[W#E7DS>@4LY:$,+!]/#<.?86!W'2%_[PU#6-L\W@D\JE`9NCK;6J$,9'@I
M[H;///6\2+WU-['@^DBDH]S5&3BN)@TIGD:(X&D%C.)%DO4=.^M.[21)Q+>J
M=8=P$R>'MM(DCY`$M(ZOK:&C!Q%:YZG&``)85B@JV1OV/449ZN)-?L##D8!U
M81">X!=B%XH%SD&U"D!)6T,6D;?'HCN,C-H"CU.Z,K*"3=Y?_:0T<+NA9H-^
M)`=WCWVZWAB^[?I55.4M&$$=E_B(6)*I51&M]7/9-GV!UW74+V$*"6T$'0/J
M-B_O&%MHT-62L=!A.\U2W@Z!UI*&K/[UQ`0`F=NM2.)B:R,8[RG61MAFB#W4
M`V$JV`,V,W%,A#Z38(A4Q6EAH##JT-KD<\4-:;'CC2N.)^9K@Y.LTP-0,JD`
M.#^`O]K]Z'Q9Z:(M/`X!Q0&$XC$&2,GCTBUYMS7G>@K#QCH>C+X=$NAEZ&2V
M/Q)8Y);OHU'9ZXE90S-&4XV>ZXS<#%RLZR*-!4Z8-?["9*$'F"BT1;K<3.GC
M`J+-DIR#F>H:[.2=(5O89`@^2I6G+*.TNB['<;>P!.Q'J>"#B2(R+Y6\`X7!
M;FME.XC>I3#9*11(`<FNIA$ZX*!=A556'+T_LF>>61SHTTU<[:DF-T#6Q[BA
ME[D#R1;[Z==(<PA''SGH]5HYQ+8R1=3TW,JPR^C\"_2AVOYWK.\UG]'[,47U
M(JC5>H@9Y'RWAK/;/5'@6]/%BCTO$NX1'GB#;GG$*SKT;</SW)#TF1*3#L,F
M<VTJ\2?=6!+Q&@6]9AQXK>FT)A7G;:5GE^T8;AM]2)D[%JY'>C!;K;>SB"NB
MUUU=/O4.P%%LR"7?[0_4C\H\==U+OD\96*GFL)%.QU:9+FS/L0WB@:M!^M1/
M=/_8`)<\X34X1X!@B72P!H$1]>3-[;!9FA-RSZ>=$'\&PIT+^";SE"<(OWWH
MJ'HF#DVLYEC)+'?4TX$.X^SS=7PE6*AIA.\0ULPW627IUEZ':0VBQ%>INMH>
MFTLSC1I`6,:RY";%'".?L)$#%-B=,-=19[C8:EH+SWB=$AF5F<9PUK&[L2D&
MR\'=,+XYH\_U\-N-1E&:U:XF451M?^US<]-=[P6L`3(5,M!A[8%JJ)6:O__)
M,78T(A.Y0LQX]4O+K]Q\Q;ERRDB>+0XU%>8>PI.K@^%2GF?:\=FJ"RMLF*4,
MNEO8OFJ$)'OD/#/[=K;PQ+#'PL%8S6GYM-1@=M@)\#95.0IM:2:C'VVM?;Y)
M:TCUB&2O&B;$@74S`#5)X/F8I5TO(WD4H?<?-B=LT6N#DQFL$[M>9U,QO;-U
MF8W]TQ.VN7Y^>M?0O1RZWCSP/:!S]02O.V$5'0WA;06=IQHA,!R7HU,@6]`<
MOA/BN/$PQ:>887_CL$Y?J"0I,D!WI@B.MK\761+G>!33O5#7U]OPC@B=7'4T
M@O#3Z'Z)EZ0@:[X51-(Y`ETSHC+>IS/*^CZD425M+"_QU6791)54VS=P3#XY
MQ'YT$(0[9&U0!@\.J1\8AH_R^LT/DO(.;1'Z[ID[+_MPQR`120N\%[,P,%;*
MKI;'H#W,-^TKEL9-#&/=P6&/'X1FX?CT$K.A\!*P1Y>'(KA:]+[":[FQ&H&#
MM8DX324(`;U#W7P^+1S@M@O%(?*I24+71"/E@B,H=*<8.?O;:(XN=,EUID74
M)2KR]:1LBV;&^#^J!:.!29>8;45;IQ-B*\/Z>HY8K'1W':*@O7*I#;%>MOJM
M!@#-D,<;5ICQ+ZS\%VHBN:&S\.A^EKI!#*'5T<33,RF$=[)0B!=@ANTLJ;XE
M6FMFM73Z\HMUZY:L_`5>=]_&,*N(2K<DR9//,A'+U1FO).D;4$_!)K#FQCD>
M_I]C*;098(<;PD^<YV4B6>19I_Q-RNI>_HJ])-V]4L64<Q@8_42U!MXY6Z`V
M2W51NROC!?Q7765WJ7]Y[#\3.#='%A(OO()OHPN>;Y'1?(&YE'?+L%8.X#(3
M7_W2QOG+N"`K(46RJ7M10"B;/:\UU59?0;RQU%*5@XK7ZD[^",B_LR)]NWVI
M3$!XADY=4\<;T.8D'Z$A\5NP3/K2ME;<47S@L\0LQU`(JH?B>RZJLDA_*%_*
M[V.6\3N>]^W41!0V-3CKC2'OHL!.%3GT]7EHW+A);[4MOUC13;_!+Q&^+FNG
M9<ES]-&>=LFI^5U#RD529U4C(\)J:91`^K75E-/.`[&&&)9OP#20AK#;&%2R
MNZS96QY\@N7*+FY<C]0?)BP!Y"ZN#?W8RT7W#<$2GKZNM0MU,JKMVY-"UB!6
ML.Y^*4:]FJJ;.B[$L6TX%J5W^-CM-4!].T(>*;*;+,^:>]N@,,-?'ZN<'SDL
MF<Z`BGW9YNFW!L?H8*4E%\#X/4_*70%KJ`V$IZ7N='JG3FD@/7*=6JY!.BV&
MLN#-CU1ZB:3XQYZ&%10N-(\=2[`A2F#/,6,>I9]3RCQ!C^_B@QR\+R#J86B2
MM^).??`ZYV.0;__[2MH&6IK\@L.L-F*<DHZ+AFSZTHQZ:I^K[K@.;@7G,D9!
M*#RW1X<E[Q3_N1U!I,"NV5M&^&2>$4A]>E<8X2*AZ.I@$M5]3ZKC5V7XFOR]
/VSM-`J7_#WG;;C)Q/```
`
end
>From tm@netcom.com Fri Jul 16 16:19:00 1999
From: Toshiyasu Morita <tm@netcom.com>
To: egcs-bugs@egcs.cygnus.com
Subject: new_ia32_branch internal compiler errors
Date: Fri, 16 Jul 1999 16:19:00 -0000
Message-id: <199907162319.QAA18297@netcom15.netcom.com>
X-SW-Source: 1999-07/msg00631.html
Content-length: 954

With the new_ia32_branch compiler from cvs today (gcc-2.96 1990621)
using only the "-O2" compiler option:

./g_phys.i: In function `ClipVelocity':
./g_phys.i:4067: internal error--unrecognizable insn:
(insn 179 23 180 (set (reg:SI 82)
        (and:SI (zero_extract:SI (reg:HI 34)
                (const_int 8 [0x8])
                (const_int 8 [0x8]))
            (const_int 69 [0x45]))) -1 (insn_list 23 (nil))
    (expr_list:REG_DEAD (reg:HI 34)
        (nil)))
...
./q_shared.i: In function `BoxOnPlaneSide2':
./q_shared.i:3067: internal error--unrecognizable insn:
(insn 179 131 180 (set (reg:SI 82)
        (and:SI (zero_extract:SI (reg:HI 72)
                (const_int 8 [0x8])
                (const_int 8 [0x8]))
            (const_int 5 [0x5]))) -1 (insn_list 131 (nil))
    (expr_list:REG_DEAD (reg:HI 72)
        (nil)))

stress-1.12 is at the same location as always, namely:

ftp://shell14.ba.best.com/pub.t/tm2/stress-1.12.tar.gz

Toshi