rothen@unix.cis.pitt.edu (Seth B Rothenberg) (03/19/90)
I am trying to convert a fortran code-generator into a
C code-generator. I think I have worked out a scheme for
commons, but I was recently told something that casts some
doubt on my idea.
I was thinking that a common could be converted into a structure
e.g.,
COMMON TEST /A/B/C/
INTEGER A,B,C
becomes
struct st_com_TEST {
int A,B,C;
};
extern struct st_com_TEST *com_TEST;
I have actually tried this on the vax and it worked.
But someone told me recently that the order of the variables in
memory is not always the same as the order of the variables in the
structure. Is this true?
Based on the way structures can be initialized, I think the
person I spoke with is mistaken.
I would appreciate any comments.
Thanks
Seth Rothenberg
rothen@unix.cis.pitt.edunews@awdprime.UUCP (USENET News) (03/20/90)
In article <22944@unix.cis.pitt.edu> rothen@unix.cis.pittsburgh.edu (Seth B Rothenberg) writes: >I have actually tried this on the vax and it worked. >But someone told me recently that the order of the variables in >memory is not always the same as the order of the variables in the >structure. Is this true? I don't understand why the rest of the code would depend on the order of the variables in memory. Are you sure that the order matters to the rest of the generated code??? It would be legal for a compiler to change the in memory order of structure elements (though I don't think anyone does this), as it is also legal for compilers to insert unused bytes to pad elements to fit onto memory boundries (really nasty if you write out structures on machine A and try and read them on machine B, a big no no). -- sanders The 11th commandment: "Thou shalt use lint" For every message of the day, a new improved message will arise to overcome it. Reply-To: cs.utexas.edu!ibmaus!auschs!sanders.austin.ibm.com!sanders (ugh!)
henry@utzoo.uucp (Henry Spencer) (03/20/90)
In article <22944@unix.cis.pitt.edu> rothen@unix.cis.pittsburgh.edu (Seth B Rothenberg) writes: >... someone told me recently that the order of the variables in >memory is not always the same as the order of the variables in the >structure. Is this true? Both K&R1 and ANSI C say the order must be preserved. There are probably compilers that don't do this, however. >Based on the way structures can be initialized, I think the >person I spoke with is mistaken. Structure initialization has no bearing on the question, since the compiler need only rearrange the order of elements in the initializer to match. -- MSDOS, abbrev: Maybe SomeDay | Henry Spencer at U of Toronto Zoology an Operating System. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
dwg@bpdsun1.uucp (David W. Glessner) (03/21/90)
In article <1990Mar19.213645.1962@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: in structures >Both K&R1 and ANSI C say the order ^ must be preserved. Are any guarantees made about the order (i.e. addresses) of external variables? -- David quintro!bpdsun1!dwg@lll-winken.llnl.gov uunet!tiamat!quintro!bpdsun1!dwg
meissner@osf.org (Michael Meissner) (03/22/90)
In article <1990Mar21.031726.18192@bpdsun1.uucp> dwg@bpdsun1.uucp (David W. Glessner) writes: | In article <1990Mar19.213645.1962@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: | in structures | >Both K&R1 and ANSI C say the order ^ must be preserved. | | Are any guarantees made about the order (i.e. addresses) of external | variables? Nope. The Data General MV linker sorts global variables by alignment, so that the items with the largest alignment come first (the C compiler also sorts static data by alignment and size). The VMS linker sorts global variables alphabetically. The System V.3 linker allows the use of a runtime script to determine where things go. -- Michael Meissner email: meissner@osf.org phone: 617-621-8861 Open Software Foundation, 11 Cambridge Center, Cambridge, MA Catproof is an oxymoron, Childproof is nearly so
henry@utzoo.uucp (Henry Spencer) (03/22/90)
In article <1990Mar21.031726.18192@bpdsun1.uucp> quintro!bpdsun1!dwg@lll-winken.llnl.gov (David W. Glessner) writes: > in structures >>Both K&R1 and ANSI C say the order ^ must be preserved. > >Are any guarantees made about the order (i.e. addresses) of external >variables? None whatsoever. -- Never recompute what you | Henry Spencer at U of Toronto Zoology can precompute. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
wallis@labc.dec.com (Barry L. Wallis) (03/22/90)
In article <1990Mar21.031726.18192@bpdsun1.uucp>, dwg@bpdsun1.uucp (David W. Glessner) writes... >In article <1990Mar19.213645.1962@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes: > in structures >>Both K&R1 and ANSI C say the order ^ must be preserved. > >Are any guarantees made about the order (i.e. addresses) of external >variables? The order of variables in a structure statement *must* be preserved (UNION wouldn't be much use otherwise). However variables which are independent from each other have no such guarantee. The following is from the _VAX C Run-Time Library Reference Manual (V3.1)_: "The C language does not gurantee any memory order for the variables in a declaration. For example, int a, b, c; "The VMS Linker (linker) usually places VAX C extern variables in program sections (psects) of the same name as the variable. The linker then alphabetically links the psects by name. If you are porting a C program from another operating system to a VMS system, you may find that the order of items in the program has been allocated differently in virtual memory. This causes existing programs with hidden bugs to fail" --- Barry L. Wallis USENET: wallis@labc.dec.com Database Consultant Prodigy (don't laugh): DNMX41A U.S. DECtp Resource Center DECUServe: EISNER::WALLIS (not on the net yet) Los Angeles, CA "No one voted for me, I represent myself" ---