don@umd5 (Chris Sylvain) (11/21/86)
[] I'd like to know just what could possibly be causing the following error message in response to a call to system() with a string that if typed by hand to a shell works just fine -- "00: is not an identifier". I'm totally mystified, and invite wild speculation on the part of you net.people (or comp.people if you've been renamed already) on what the cause might be... -- --==---==---==-- .. And hast thou slain the Jabberwock? .. ARPA: don@umd5.UMD.EDU BITNET: don%umd5@umd2 UUCP: ..!seismo!umd5.umd.edu!don
m5d@bobkat.UUCP (Mike McNally ) (11/22/86)
In article <1369@umd5> don@umd5 (Chris Sylvain) writes: >[] >I'd like to know just what could possibly be causing the following error >message in response to a call to system() with a string that if typed >by hand to a shell works just fine -- "00: is not an identifier". >I'm totally mystified, and invite wild speculation on the part of you >net.people (or comp.people if you've been renamed already) on what the cause >might be... > >-- >--==---==---==-- >.. And hast thou slain the Jabberwock? .. > >ARPA: don@umd5.UMD.EDU BITNET: don%umd5@umd2 >UUCP: ..!seismo!umd5.umd.edu!don Could be that the shell you type the command into is not the same shell used by system(). If I recall, system() uses /bin/sh. Perhaps if you provided more details concerning the shell command someone could be more helpful to you. Sort of like telling your doctor, "when I execute a certain body function, I see a large rabbit and hear Frank Sinatra singing; what could be wrong?" -- **** **** **** At Digital Lynx, we're almost in Garland, but not quite **** **** **** Mike McNally Digital Lynx Inc. Software (not hardware) Person Dallas TX 75243 uucp: ...convex!ctvax!bobkat!m5 (214) 238-7474
jerryp@tektools.UUCP (11/22/86)
In article <1369@umd5> don@umd5 (Chris Sylvain) writes: > [] > I'd like to know just what could possibly be causing the following error > message in response to a call to system() with a string that if typed > by hand to a shell works just fine -- "00: is not an identifier". [I've redirected followups to comp.unix.questions only.] I haven't checked into *why* your message is happening, but I can reproduce it. If your environment is trashed (has an environment variable with a bad name), you can get the message. Here's an example from 4.3BSD... the Bourne shell never starts: Script started on Sat Nov 22 05:49:49 1986 $ /bin/csh -f % setenv "FOO bar" trash % sh FOO bar=trash: is not an identifier % --Jerry Peek, Tektronix, Inc. US Mail: MS 74-900, P.O. Box 500, Beaverton, OR 97077 uucp: {allegra,decvax,hplabs,ihnp4,ucbvax}!tektronix!tektools!jerryp CS,ARPAnet: jerryp%tektools@tektronix.csnet Phone: +1 503 627-1603
don@umd5.UUCP (11/23/86)
In article <1903@tektools.UUCP> jerryp@tektools.TEK.COM (Jerry Peek) writes: >In article <1369@umd5> don@umd5 (Chris Sylvain) writes: >> ... >> error message in response to a call to system() >> -- "00: is not an identifier". > >I haven't checked into *why* your message is happening, but I can reproduce >it. If your environment is trashed (has an environment variable with a bad >name), you can get the message. Here's an example from 4.3BSD... the Bourne >shell never starts: > > Script started on Sat Nov 22 05:49:49 1986 > $ /bin/csh -f > % setenv "FOO bar" trash > % sh > FOO bar=trash: is not an identifier > % I never touched **envp, honest! Also, "FOO bar=trash" is not the same as "00" <- read "zero,zero". I peeked at the sh sources and can see where the "notid" error message is generated, but I don't know why or how. For me, all it takes is a plain-vanilla system call like system("printenv") to generate the error message. Really mysterious. -- --==---==---==-- .. And hast thou slain the Jabberwock? .. ARPA: don@umd5.UMD.EDU BITNET: don%umd5@umd2 UUCP: ..!seismo!umd5.umd.edu!don
levy@ttrdc.UUCP (11/23/86)
In article <1369@umd5>, don@umd5 (Chris Sylvain) writes: >[] >I'd like to know just what could possibly be causing the following error >message in response to a call to system() with a string that if typed >by hand to a shell works just fine -- "00: is not an identifier". >I'm totally mystified, and invite wild speculation on the part of you >net.people (or comp.people if you've been renamed already) on what the cause >might be... My speculation: garbage being somehow put in the environment. An example: main() { putenv("00"); system("echo hello world"); } $ a.out 00: is not an identifier -- ------------------------------- Disclaimer: The views contained herein are | dan levy | yvel nad | my own and are not at all those of my em- | an engihacker @ | ployer or the administrator of any computer | at&t computer systems division | upon which I may hack. | skokie, illinois | -------------------------------- Path: ..!{akgua,homxb,ihnp4,ltuxa,mvuxa, go for it! allegra,ulysses,vax135}!ttrdc!levy
mouse@mcgill-vision.UUCP (11/24/86)
In article <1369@umd5>, don@umd5 (Chris Sylvain) writes: > I'd like to know just what could possibly be causing the following error > message in response to a call to system() with a string that if typed > by hand to a shell works just fine -- "00: is not an identifier". As you have probably already been informed, this is because the Bourne shell, which is the shell system() uses, doesn't like environment entries not conforming to a specific syntax. [This is a BUG, do you hear me, a BUG....] Might I suggest a patch to sh (line numbers are 4.3BSD version, but this should work on the 4.2 version if you adjust the line numbers): diff -c -r sh/name.c /usr/src/bin/sh/name.c *** sh/name.c Sun Nov 23 22:28:13 1986 --- /usr/src/bin/sh/name.c Sun Nov 10 18:32:05 1985 *************** *** 55,61 **** { WHILE arg DO REG STRING s=mactrim(arg->argval); ! setname(s, xp, 0); arg=arg->argnxt; IF flags&execpr THEN prs(s); --- 55,61 ---- { WHILE arg DO REG STRING s=mactrim(arg->argval); ! setname(s, xp); arg=arg->argnxt; IF flags&execpr THEN prs(s); *************** *** 64,73 **** OD } ! VOID setname(argi, xp, nofail) STRING argi; INT xp; - int nofail; { REG STRING argscan=argi; REG NAMPTR n; --- 64,72 ---- OD } ! VOID setname(argi, xp) STRING argi; INT xp; { REG STRING argscan=argi; REG NAMPTR n; *************** *** 98,107 **** return; FI FI ! if (nofail) ! { return; ! } ! failed(argi,notid); } replace(a, v) --- 97,103 ---- return; FI FI ! failed(argi,notid); } replace(a, v) *************** *** 306,312 **** REG STRING *e=environ; WHILE *e ! DO setname(*e++, N_ENVNAM, 1) OD } LOCAL INT namec; --- 302,308 ---- REG STRING *e=environ; WHILE *e ! DO setname(*e++, N_ENVNAM) OD } LOCAL INT namec; der Mouse USA: {ihnp4,decvax,akgua,utzoo,etc}!utcsri!mcgill-vision!mouse think!mosart!mcgill-vision!mouse Europe: mcvax!decvax!utcsri!mcgill-vision!mouse ARPAnet: think!mosart!mcgill-vision!mouse@harvard.harvard.edu [USA NSA food: terrorist, cryptography, DES, drugs, CIA, secret, decode]
henry@utzoo.UUCP (Henry Spencer) (11/25/86)
As I recall, the "is not an identifier" message means that the shell
found an environment string which was not in the "variable=value" form
that pre-V8 shells insist on. This means either somebody has mangled
the environment, or else dear old init isn't setting up a proper
environment before forking children and a shell is eventually getting
run without any of its ancestors doing anything about an environment.
--
Henry Spencer @ U of Toronto Zoology
{allegra,ihnp4,decvax,pyramid}!utzoo!henryjsdy@hadron.UUCP (Joseph S. D. Yao) (12/12/86)
In article <1369@umd5> don@umd5 (Chris Sylvain) writes: >message in response to a call to system() with a string that if typed >by hand to a shell works just fine -- "00: is not an identifier". Got the same message by using Bell's scheme for logging into a chroot'ed shell in System V. In that case, the environment var list was trashed to put a funny SUB-LOGIN symbol. The Bourne shell saw this, and had fits. System() also uses the Bourne shell, perforce (unless you've modified it the way I did mine). Somehow, you're getting a weird environment string in that says "00" something. You may be trashing your stack. -- Joe Yao hadron!jsdy@seismo.{CSS.GOV,ARPA,UUCP} jsdy@hadron.COM (not yet domainised)