[ Previous | Next | Table of Contents | Index | Library Home | Legal | Search ]

Files Reference

ar File Format


Combines several files into one. This is the archive file format for AIX 5.1 of the operating system on an Itanium-based.


The ar command is used to combine several files into one. Archives are used mainly as libraries to be searched by the ld command, which is the link editor.

Each archive begins with a unique string identifier called an archive magic string. The following are examples:

    #define  ARMAG   "!<arch>\n"  /* magic string */ 
    #define  SARMAG  8            /* length of magic string */ 

Following the archive magic string are the archive file members. Each file member is preceded by a file member header, which takes the following format:

   #define  ARFMAG    "`\n"      /* header trailer string */ 
     struct  ar_hdr              /* file member header */ 
       char    ar_name[16];      /* '/' terminated file member name */ 
       char    ar_date[12];      /* file member date */ 
       char    ar_uid[6];        /* file member user identification */ 
       char    ar_gid[6];        /* file member group identification */ 
       char    ar_mode[8];       /* file member mode (octal) */ 
       char    ar_size[10];      /* file member size */ 
       char    ar_fmag[2];       /* header trailer string */ 

All information in the file member headers is in printable ASCII code. The numeric information contained in the headers is stored as decimal numbers (except for the ar_mode, field which is in octal). Thus, if the archive contains printable files, the archive itself is printable.

If the file member name fits, the ar_name field contains the name directly and is terminated by a forward slash (/) and padded with white spaces on the right. If the member's name does not fit, the ar_name field contains a forward slash followed by a decimal representation of the name's offset in the archive string table, which is described below.

The ar_date field is the modification date of the file at the time of its insertion into the archive. Common format archives can be moved from system to system as long as the portable archive command ar is used.

Each archive file member begins on an even byte boundary; a new-line character is inserted between files if necessary. Nevertheless, the size given reflects the actual size of the file, not including padding.

Notice there is no provision for empty areas in an archive file.

Each archive that contains object files (see the a.out file) includes an archive symbol table. This symbol table is used by the ld command to determine which archive members must be loaded during the link edit process. If it exists, the archive symbol table is always the first file in the archive (but is never listed) and is automatically created and/or updated by the ar command.

The archive symbol table has a zero length name (that is, ar_name[0] is '/', ar_name[1]==' ', and so on). Each "word" in this symbol table has four bytes, using the machine-independent encoding shown below. (All machines use the encoding described here for the symbol table, even if the machine's "natural" byte order is different.)

  0 1 2 3
0x01020304 01 02 03 04

The contents of the symbol table are as follows:

  1. The number of symbols. Length is 4 bytes.
  2. The array of offsets, one per symbol, into the archive file. Length is 4 bytes * number of symbols.
  3. The name string table. Length is ar_size - 4 bytes * the number of symbols + 1.

As an example, the following symbol table defines four symbols. The archive member at file offset 114 defines name and object. The archive member at file offset 426 defines function and a second version of name.

Offset +0 +1 +2 +3  
0 4 4 offset entries
4 114 name
8 114 object
12 426 function
16 426 name
20 n a m e  
24 \0 o b j  
28 e c t \0  
32 f u n c  
36 t i o n  
40 \0 n a m  
44 e \0      

The number of symbols and the array of offsets are managed with the sgetl and sputl subroutines. The string table contains exactly as many null terminated strings as there are elements in the offsets array. Each offset from the array is associated with the corresponding name from the string table (in order). The names in the string table are all the defined global symbols found in the common object files in the archive. Each offset is the location of the archive header for the associated symbol.

If an archive member's name is more than 15 bytes long, a special archive member contains a table of file names, each followed by a forward slash and a new-line character. This string table member, if present, precedes all "normal" archive members. The special archive symbol table is not a "normal" member and must appear first, if it exists. The ar_name field of the string table's member header holds a zero length name (ar_name[0]=='/'), followed by one trailing forward slash (ar_name[1]=='/'), followed by blanks (ar_name[2]==' ', and so on). Offsets into the string table begin at zero.

The following are example ar_name values for short and long file names:

Offset +0 +1 +2 +3 +4 +5 +6 +7 +8 +9
0 f i l e  _  n a m e  _ 
10 s a m p l e / \n l o
20 n g e r f i l e n a
30 m e x a m p l e / \n
Member Name ar_name Note
short-name short-name/ Not in string table
file_name_sample /0 Offset 0 in string table
longerfilenamexample /18 Offset 18 in string table

Related Information

The a.out file.

The ar command, ld command, strip command.

The sgetl and sputl subroutines.

[ Previous | Next | Table of Contents | Index | Library Home | Legal | Search ]