From: Digestifier To: Linux-Activists@senator-bedfellow.mit.edu Reply-To: Linux-Activists@senator-bedfellow.mit.edu Date: Sat, 14 Aug 93 00:13:05 EDT Subject: Linux-Activists Digest #112 Linux-Activists Digest #112, Volume #6 Sat, 14 Aug 93 00:13:05 EDT Contents: Re: Bug in bash 1.12 - not! (Brandon S. Allbery) [PATCH] non-blinking software cursor for pl12 (& pl11) (Just GNU it.) Re: More annoyance on the DMA problem (Andrew Gillham) Tape Drives: Suggestions [summary] (Robert W. Brewer) Re: Linux and a 3com Etherlink III 3C579 EISA ethernet card (Donald J. Becker) Re: what software for PC-NFS (pc-end) (Donald J. Becker) ---------------------------------------------------------------------------- From: bsa@kf8nh.wariat.org (Brandon S. Allbery) Subject: Re: Bug in bash 1.12 - not! Date: Sat, 14 Aug 1993 02:53:03 GMT In article <1993Aug12.221302.3511@desaster.hanse.de> michaelw@desaster.hanse.de (Michael Will) writes: >bsa@kf8nh.wariat.org (Brandon S. Allbery) writes: >>*Variable* expansion takes place in double quotes. Globbing, whether against >>files or against case expressions (*except* *in* *bash*), does not. Try: > >>$ echo "/*" > >>It had better respond with > >>/* > >$ echo "/*" >/* >$ bash -version >GNU bash, version 1.12.4 I didn't say bash had a problem with quoted file glob characters, only case patterns. I was replying to a comment that bash expands variables in double quotes by specifying the difference between variable expansion and filename expansion and suggesting an example to show the difference. (Do I *really* have to write every message as if everybody needs *all* of the background from the beginnings of Unix on??? Sheesh.) ++Brandon -- Brandon S. Allbery kf8nh@kf8nh.ampr.org bsa@kf8nh.wariat.org ------------------------------ From: ghod@drycas.club.cc.cmu.edu (Just GNU it.) Crossposted-To: comp.os.linux.misc Subject: [PATCH] non-blinking software cursor for pl12 (& pl11) Date: 13 Aug 93 23:35:37 -0400 [Please note: I'm posting this to comp.os.linux and comp.os.linux.misc for the sole reason that some people may not have the new newsgroups available to them yet. In the future, I will only post updates to this patch (if they become necessary) to comp.os.linux.misc (it doesn't seem to fit in anywhere else).] Greetings: Appended to this file is an updated and much improved version of my software cursor patch for Linux 0.99 pl12 (I was going to hold off on releasing this until pl12 became official, but I wanted to get it out over c.o.l before it bit the dust). This version incorporates a fix for a really nasty bug and a few major overall improvements. This patch should also apply cleanly to pl11 (although the hunks may be offset slightly). Users of pl10 and earlier can also use this patch, but you will have to edit several pieces in by hand. Changes: ======= - Finally eliminated the 'stray cursor left hanging around after screen unblanks' bug. For real this time. Honest. :) - Completely re-did the software cursor code in set_cursor(). - Added a #define called SCURSCOLOR to set the software cursor color at compile time. Features: ======== - With this patch, the standard hardware-generated 'blinking underline' cursor is replaced with a software-generated non-blinking block cursor. - You can switch back and forth from software to hardware generated cursors at any time using a VT-200 style escape code. - Each VC's cursor is maintained seperately: you can have half your VCs using the software generated cursor, half of them using the hardware cursor, or any combination in between. New Features: ============ - The cursor color can be chosen at compile-time. Possible colors are white, yellow, magenta, red, green, cyan, blue and black. The default is white, which should accomodate monochrome display adapters as well as VGA or SVGA color ones. White seems to be a good, all-around default and I recommend using it, but if you have a color video adapter feel free to experiment. Whatever color is chosen will be retained across all VCs. - The cursor color remains the same regardless of what foreground/background colors are selected. The only exception is if the cursor color and the background color are chosen to be the same: in that case the cursor will appear black. (This is a feature, not a bug: the code in set_cursor() will never allow the cursor color and background color to be the same since that would make the cursor 'invisible', not to mention useless.) Before anyone asks: I am *not* planning to make the cursor color switchable at run time unless an angry mob comes after me and demands it. It would be feasible to add yet another escape code (or codes) to handle this, but I can't see any real advantage to it myself. If enough people show interest in the idea, however, I might be persuaded to give it a shot. Configuration: ============= There are now 3 #defines that control cursor behavior: #define NOBLINK 1 #define HCURSTYPE 0 #define SCURSCOLOR 112 If NOBLINK is #defined, all VCs will default to software generated cursors at boot time. If it is #undef'ed, they'll default to the usual hardware cursors. HCURSTYPE controls the size of the hardware cursor. 0 is a full-sized blinking block. 13 is the standard blinking underline. 15 is off (invisible). (Note that 'invisible' mode won't do you much good unless you're psychic. :) I have it set to a block instead of the usual underline just because it looks neater when switching back & forth from a blinking to non-blinking cursor. I never liked the stupid little flashing underline anyway, which is why I got into this mess in the first place. ;) SCURSCOLOR sets the software cursor color. The default is white. Legal values are: 0 16 32 48 64 80 96 112 black blue green cyan red magenta yellow white Note that setting the cursor color to black will actually work, only it will appear white on a black background since the code is smart enough to prevent the cursor form accidentally becoming 'invisible' due to a bad choice of screen colors. You can switch from software to hardware generated cursors on the fly using a VT-200 style escape code as follows: [?33h switch to hardware cursor [?33l switch to software cursor I have learned (thanks to nation@rocket.sanders.lockheed.com) that there actually aren't any escape codes defined for this sort of function in the VT-200 command set, however the codes I have selected do fit in the pattern. This makes them non-standard but harmless so long as nothing trips them by mistake. Note that even if you #undef'ed NOBLINK, you can still switch a VC to software cursor mode with these escapes. Installation: ============ This patch is a gzipped and uuencoded context diff against /linux/kernel/chr_drv/console.c. To apply it, extract the attached uuencoded file to /linux/kernel/chr_drv/noblink.uu and do the following: # uudecode noblink.uu # gunzip noblink.dif.gz # patch < noblink.dif Check for complaints from 'patch'. If it bitches about some hunks being offset by a few lines, that's fine. If it rejects some of the patch, you'll have to apply the rejected hunks by hand. Once you are sure the patch has been applied correctly, edit console.c and set the #defines as you prefer, then build your kernel. As always, questions, comments, complaints, flames, suggestions, bug reports, bug fixes, or large sums of cash are welcome and encouraged. Share and enjoy! -Bill Paul "A black hole is what happens ghod@drycas.club.cc.cmu.edu when God divides by zero." PS: Special thanks to Michael Haardt (u31b3hs@pool.Informatik.RWTH-Aachen.DE) for testing this on a mono display. ==================================CUT HERE==================================== begin 644 noblink.dif.gz M'XL("'0^:"P``VYO8FQI;FLN9&EF`-U:>W/;-A+_F_T4ZW3J4C8MBY9E27;M MUDG+!ERDYS^O=>1))!!:+Q6\? MV`6XM;4%7ARE<_U=%&-;,I[&_@]^LO36UOELD9ML4,^(FS?!U$V.\S^"X,HNQN M-_6FS&]/3QJZ>#!CR8HNOJ2.)@CZMC6P)01R0I^-@XC!QU$U@JR7&XK\L M2H-)Q'P(XVABW`8^BYTHFR&',)M%Z9%A[&[!138;L03B,7"<#E2?@3!MK&2` M2V<-PU$"T6,84HT`*7=YX!5\2$#%AR_G3+*XQE_$P`_2>>@N8<2":`)9RGS) MI\ZD),R,S9R1FRI&C_$7,1)]@'UQLA2RD`7V#RR[T],FV*R.PNZ>JK[+Q\_/ M+YX!CR%E''YYDM+/*(YV1KC2&Y)3&2;:)/)#@^1MP4&9+P[Z^'6#3.6`HY4`I*$^&R"X8$DL??4D%'H M>C?XF3&8)(Q%X"W="$W6@'?+YW(/Q7ZY`I22[U%:` MWBFUEK"0LOR?.>#=?ZL#-H5'NV.C1QX4>X2]U[/L[KYV4:,VG^?<6?BQ/&KL MX_&<>D1SZDL;E M/&D6+9J["8F/7V\H'KUM)BLPQ^D2?D287]&O`O17IR^4]M<-9Y$O!I_)$)0/ M!1-#`7%KK=3:_L"R>]V2UBBP#O+`JEWMSC!,P@;=X@TN/J$?;]NDQ%:):KF* M:EFF$MBNHA2=K9*3*UVNHE?=Y1&%@E>,*0C*PPJ]KQA6$)2'5-,W42\A^=ZH! MJ'5R,FQ5AN-&51EN]SZ)N]W,??6J,6%8N^A@#*9NA(UC&$_TS@)__*'3;@=W MRN@&?1B;;CW,?AD<'\.SI\Y/KTY?_GS^Y*J%NX1A)(QG272DV3JX.C=TXB28 M!!'Q=M1OFMPP2#+Y7*7$!2$S+PS,5L[*9YXW:Z'$-,TZ/97Z%93S.'U/)95Y M-VGI/7DWJ@C'O@,6IDQP*9MBKA%!E'*U[*I&I2J]))M7M&@5NR$U\]KSJ-`R M/E&&ARQB+%F.P3Z2!=&@;_6Z0QTV/\9YMLL)Z_FW,XQ\'-(,4SI47,"_30&U M.N64=F)I@_D6I7TJVQ-1"<:A.X$I2U@;SKGDDS(VTP/#>!)0T%M:,,JXF(*6 M$B[!U7TP9\@O:NNQ=&`6_#:\P(R9I&#`C!!58*R'.,YM104$A6"U=43#R>3$)9]F"9 M@B;#B"/^UC:98-)/;D:U`"HXX$$/Y?'DHN7-*7XDQ_:=Y0BS#LA1\K+\2=Y%*9IM41=$C+`(^E4MP9[@_HTF[ M-+^T>ESE`CV$ENG)Y6^=7EV]?G&VA4-<#FJ=BLJ/]=S*]Z;TK-8IFS#%]G`5 M+^)4I&U"SF!&XRNLY.SDTFZX<)G2E(I'ZA\C$DN2QTUB7#NX8RQG-.I9)+::M%R^7E`%QM`4;E'C#:&# M5,]<;XJ%L02#8]66$M18MV%S#J5DA])A`PZ:M>%G[!MG(K8(Z1=TH'/#V)Q& MIVY$(2D"U_<#J0JE2LE(VA"VXS;I,V&J(UJ#'Z.ANXB64"EQP(+-%2GN/S.$ MFS9)6G\L]%'F)9A;M%BJ\VFRW/30;H*$>3Q<*M-#:TJ5.47!F"^EDI55A<%- M#I@H7C5L.*>KL18&Y\KZ=X25\$3HH[TR@G[F;(%X;>AP3&7`4W0VW+3<(`0, M/T*M]>,!48X:GS73:$PQ:CN%F)/:,"C@,PF+QA$@Z')+TV)I?/(-!`G_QJ15 MT`:'OW!5['LY0(TA;K^R]!#F&1=ZH"VE4'R[2KQV+])$M3U)MJJ=)9]4_Q(> M=H@+NF75F:T2!7FO\2YKM: MA)P.13!+)SXG)[#?:A7=Z_19\+C'X2CORY/*![AM%NS,/TL8Y)PJC':E493! MT`JIZ.$!1911Z-QU^J4&U,1]7#X,EF+XA\-10J-!?+AG2$`+&/S8@C^@`;QW MZGM7'K;&65(<:(K33K%3I84E:<:55-#6[%;GCZ*?9JO^Y7ECO4/_J7)J5??Z MDFKMF(\HL];+T%QZ?;0,*\JQYC]5?MV#-V]9#?3*FNX^JW?2@1HR8['C-.?& M&R!"-L;B4Y7UQG)/P5Q`9PMJRY)I$>8YO[,D!C,>CUO@1KXFDUSJ!_Y3ILNT MXM3:%$D;8&]^U-^2HV5ZL`%Y-"@0T4)0:9,WUNQ/Z[ICK52U(LF%L58J,5?: M:ER53`\A2[Y;QY"2.3..T(IF\2V[!QN"*],IR8%@3IB?>:@7S,\*4$D5[0_` MSGX0NW6^^BG%87V>=?[X*67BA]L"BO*0$:BR<_L+GYTT7K3VNE:_-]0'D13' M1YBZW!S)!X]N)?9ZAV*3>**S\5VT'WG=3/0B!Z4SAL"S%K M+\2CLZ)A9V`-]^S\8N2SK:]2".@-MWG1&^5I-_)IN]W*M(+3KO:E2G(N9\YG MJ42%>_.4,-^QQKV>Z7+$+N[9]GV?J$K`?'9BY?7OYJ;G"]W M3A8)%H7.;RUUXB4+.DS)4A8Z^GP8!0J9FSC8AH4GUKG2P;Y&%T=^!/?EQ8_G M/SE79\_/GER?7U[D=Z*5^DF<;C6K4[BI@04JUH_FAA`LY?%\CCZXN6F8'IK* M!`=B!UT`9*PF?`M.<)-I23<5,W4[^Q9^#$O8[B/:^#$LW>%+>.<)1@IWI$IAM+"I5\-HD(MH_5 M.863!K\SY6=TT^8EHVQ+]!YH>JZ7Q$>YKZ&8CJLKE*8Z*%%,W'$M*03A0A).$+36) MB"P.9\DLB-S0+':]JAGI,W`%C#2ZU3L.;:Y&$2_*YS\*PA6K[^)2]_=[^>8X MB7E\MRS$(GTY=_)K*>?*YCZJQI%&8!8SR=["VW3=)%INS$=/)-DA?).*?_[= M-[Z%7VAG"<\P)"@^Z3^B1Y8,B?!P;CM MR99[KU%8M?W]"'9D)\34,Q`6Z**W:1UPBC.2`W/.CF-]U? M`$D2O&P3(C_5*?(U24:U1D,BW(93CQ"FX]0%'<1BE4+';>&2BA6L4U1FG-<@ M>XM9'L+N,!'.<*$C% M*:PX2HW8+4N0=>9-&P5W)VX0-:;F1IYY&VN35&-M?FKD]8FA"I0/-U4K?QM* M"GW\S7_:>F6#"J7?/]('LSB1QO/1)QMX<^CH]S%T#+I%7ETZ_!6O1#K">M\\ M?GYZ\QR)]]/$2;0R?#&`T@E\Q(FK?H"+*Q?P=1^IGVP+3\%V MY(.Q^[P-0 MHV9[G2%JUBYRA4J241J_#C5X?_V6) jmonroy@netcom.com (Jesus Monroy Jr) writes: > > > More annoyance on the problem with the DMA/RAM refresh > ------------------------------------------------------ > > Some facts from: > "IBM Personal Computer Troubleshooting and & Repair > for the IBM PC, PC/XT, and PC AT" > by Robert C. Benner Ok, Ok, I've got a question. Now this is nothing personal against Mr. Monroy or anything, this article just solidified it. Ok, the question.... Why the hell do I care if my floppy disk drive is 100% optimized as long as it can read the stupid disks? I personally have ZERO interest in using floppies for *ANYTHING* except the bare minimum necessary to boot my stinkin' PC! What's wrong with a SCSI tape drive? These floppy attached ones are nothing but trouble! They may be great in the single-tasking DOS world, but seem like a real pain in the multi-tasking world. Buy a stinkin' SCSI tape drive if you're *serious* about backing up your machine! It seems like a waste of bandwidth to flame/counter-flame over some vapor-code from a blowhard (seemingly) that simply speeds up something that we don't want to do anyway. I'd think all this raw coding talent should be spent on more significant things, for example, Mr. Evans' amazing interrupt/serial code. There is a perfect example of spending that extra effort and yielding *SIGNIFICANT* improvements at the kernel level rather that wasting time on seldom used drivers (as compared to kernel routines). Anyway, just my $.02 (or maybe $5.00) worth. IRETs, -Andrew -- ====================================================== Andrew Gillham gillham@andrews.edu Andrews University ...!aucis!gillham Berrien Springs, MI 49103 ------------------------------ From: rwb114@cac.psu.edu (Robert W. Brewer) Subject: Tape Drives: Suggestions [summary] Date: 14 Aug 1993 03:45:52 GMT Reply-To: rwb114@cac.psu.edu A few days I posted a request for suggestions on tape drives for use under Linux. Here are the responses I received (unfortunately there were only two). I'll just quote their responses directly since they are short: #1: I'm running a Wangtek 5150 ES (150/250 MB) via an Adaptec 1542B. My disks and a CD-ROM are SCSI, too. I backup my Linux and DOS partitions with a speed of 6MByte / minute under Linux 0.99 pl6. Germany.EU.net nearby in Dortmund transferred part of their software to two 250 MB tapes which I could easily read. And I have read an old DC-600 archive. Of course, the Adaptec is not a cheap SCSI controller, but consider buying it or the EISA version now for the next disk(s) (SCSI for performance and other reasons). If you want to have _only_ slow devices on the SCSI bus, any cheap but reliable controller will do. #2: You can either go for SCSI, or shop around for a used QIC-02 tape, I've picked up Wangtek 5150's with controllers for as little as $130, they'll do 250mb uncompressed, and a lot more if you compress. They will use either 6150 (150mb) tapes or 6250 (250mb) tapes. I've used both, and the only problem with SCSI is you'll have to pick up a host adapter, that will normally make the SCSI solution a bit more expensive. [end responses] -Rob -- Robert W. Brewer rwb114@cac.psu.edu ------------------------------ From: becker@super.org (Donald J. Becker) Subject: Re: Linux and a 3com Etherlink III 3C579 EISA ethernet card Date: Fri, 13 Aug 1993 06:52:09 GMT In article <246ji8$cn7@netline-fddi.jpl.nasa.gov> unbelver@brain.jpl.nasa.gov (Carlos Y. Villalpando) writes: >Does anyone have any definite information on whether or not Linux will work with >3Com's Etherlink III 3C579 EISA ethernet card. (I'm looking for more of a "no it >will never work answer", but I'll gladly take a "yes I'm using it right now" answer) The answer is closer to "yes, I'm using it right now". (Actually it's a 3c509, the ISA version.) Before you get too excited: I'm probably the only one running it -- certainly I'm the only one running the latest version ;-). It's not even close to the beta test stage. I seem to be getting some packet corruption (with back-to-back packets?). The driver should work with the EISA version with minimal changes, but I have neither an EISA machine nor a 3c579. -- Donald Becker becker@super.org Supercomputing Research Center 17100 Science Drive, Bowie MD 20715 301-805-7482 ------------------------------ From: becker@super.org (Donald J. Becker) Subject: Re: what software for PC-NFS (pc-end) Date: Fri, 13 Aug 1993 16:38:42 GMT In article <1993Aug10.004825.5627@engr.engr.uark.edu> hbj@engr.engr.uark.edu (Bernard Johnson) writes: >What I want is the DOS end. Doesn't the Net-2 distribution already have a nfs >server? > >And as I remember, the nfs server code is plainly copyrighted, but has no >distribution restrictions. I'd have to check, but if the Net-2 code is >stable I'll probably just stick with that. There is a freely available MS-DOS NFS server called SOS/SOSS. But many people want a free MS-DOS NFS client implementation. That's unlikely to happen: It wouldn't promote "the state of the art" in computing, in fact it would do just the opposite. Some people would summarize this as "it wouldn't be fun to write". The support burden would grossly overwhelm the code writing effort, since MS-DOS applications are notoriously ill-behaved, and people tend to blame free software when something doesn't work. It requires very intimate MS-DOS knowledge to write, which only comes with having money and "major software company" muscle. Every new version of MS-DOS would break it. Acceptable commercial versions are readily available. Someday, perhaps, the Linux "dosemu" virtual machine will be good enough that some people will use it just to replace PC-NFS. -- Donald Becker becker@super.org Supercomputing Research Center 17100 Science Drive, Bowie MD 20715 301-805-7482 ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Activists-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux) via: Internet: Linux-Activists@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux tupac-amaru.informatik.rwth-aachen.de pub/msdos/replace The current version of Linux is 0.99pl9 released on April 23, 1993 End of Linux-Activists Digest ******************************