5
llvm-ar - LLVM archiver
9
B<llvm-ar> [-]{dmpqrtx}[Rabfikouz] [relpos] [count] <archive> [files...]
14
The B<llvm-ar> command is similar to the common Unix utility, C<ar>. It
15
archives several files together into a single file. The intent for this is
16
to produce archive libraries by LLVM bitcode that can be linked into an
17
LLVM program. However, the archive can contain any kind of file. By default,
18
B<llvm-ar> generates a symbol table that makes linking faster because
19
only the symbol table needs to be consulted, not each individual file member
22
The B<llvm-ar> command can be used to I<read> both SVR4 and BSD style archive
23
files. However, it cannot be used to write them. While the B<llvm-ar> command
24
produces files that are I<almost> identical to the format used by other C<ar>
25
implementations, it has two significant departures in order to make the
26
archive appropriate for LLVM. The first departure is that B<llvm-ar> only
27
uses BSD4.4 style long path names (stored immediately after the header) and
28
never contains a string table for long names. The second departure is that the
29
symbol table is formated for efficient construction of an in-memory data
30
structure that permits rapid (red-black tree) lookups. Consequently, archives
31
produced with B<llvm-ar> usually won't be readable or editable with any
32
C<ar> implementation or useful for linking. Using the C<f> modifier to flatten
33
file names will make the archive readable by other C<ar> implementations
34
but not for linking because the symbol table format for LLVM is unique. If an
35
SVR4 or BSD style archive is used with the C<r> (replace) or C<q> (quick
36
update) operations, the archive will be reconstructed in LLVM format. This
37
means that the string table will be dropped (in deference to BSD 4.4 long names)
38
and an LLVM symbol table will be added (by default). The system symbol table
41
Here's where B<llvm-ar> departs from previous C<ar> implementations:
47
Since B<llvm-ar> is intended to archive bitcode files, the symbol table
48
won't make much sense to anything but LLVM. Consequently, the symbol table's
49
format has been simplified. It consists simply of a sequence of pairs
50
of a file member index number as an LSB 4byte integer and a null-terminated
55
Some C<ar> implementations (SVR4) use a separate file member to record long
56
path names (> 15 characters). B<llvm-ar> takes the BSD 4.4 and Mac OS X
57
approach which is to simply store the full path name immediately preceding
58
the data for the file. The path name is null terminated and may contain the
63
B<llvm-ar> can compress the members of an archive to save space. The
64
compression used depends on what's available on the platform and what choices
65
the LLVM Compressor utility makes. It generally favors bzip2 but will select
66
between "no compression" or bzip2 depending on what makes sense for the
69
=item I<Directory Recursion>
71
Most C<ar> implementations do not recurse through directories but simply
72
ignore directories if they are presented to the program in the F<files>
73
option. B<llvm-ar>, however, can recurse through directory structures and
74
add all the files under a directory, if requested.
76
=item I<TOC Verbose Output>
78
When B<llvm-ar> prints out the verbose table of contents (C<tv> option), it
79
precedes the usual output with a character indicating the basic kind of
80
content in the file. A blank means the file is a regular file. A 'Z' means
81
the file is compressed. A 'B' means the file is an LLVM bitcode file. An
82
'S' means the file is the symbol table.
88
The options to B<llvm-ar> are compatible with other C<ar> implementations.
89
However, there are a few modifiers (F<zR>) that are not found in other
90
C<ar>s. The options to B<llvm-ar> specify a single basic operation to
91
perform on the archive, a variety of modifiers for that operation, the
92
name of the archive file, and an optional list of file names. These options
93
are used to determine how B<llvm-ar> should process the archive file.
95
The Operations and Modifiers are explained in the sections below. The minimal
96
set of options is at least one operator and the name of the archive. Typically
97
archive files end with a C<.a> suffix, but this is not required. Following
98
the F<archive-name> comes a list of F<files> that indicate the specific members
99
of the archive to operate on. If the F<files> option is not specified, it
100
generally means either "none" or "all" members, depending on the operation.
108
Delete files from the archive. No modifiers are applicable to this operation.
109
The F<files> options specify which members should be removed from the
110
archive. It is not an error if a specified file does not appear in the archive.
111
If no F<files> are specified, the archive is not modified.
115
Move files from one location in the archive to another. The F<a>, F<b>, and
116
F<i> modifiers apply to this operation. The F<files> will all be moved
117
to the location given by the modifiers. If no modifiers are used, the files
118
will be moved to the end of the archive. If no F<files> are specified, the
119
archive is not modified.
123
Print files to the standard output. The F<k> modifier applies to this
124
operation. This operation simply prints the F<files> indicated to the
125
standard output. If no F<files> are specified, the entire archive is printed.
126
Printing bitcode files is ill-advised as they might confuse your terminal
127
settings. The F<p> operation never modifies the archive.
131
Quickly append files to the end of the archive. The F<R>, F<f>, and F<z>
132
modifiers apply to this operation. This operation quickly adds the
133
F<files> to the archive without checking for duplicates that should be
134
removed first. If no F<files> are specified, the archive is not modified.
135
Because of the way that B<llvm-ar> constructs the archive file, its dubious
136
whether the F<q> operation is any faster than the F<r> operation.
140
Replace or insert file members. The F<R>, F<a>, F<b>, F<f>, F<u>, and F<z>
141
modifiers apply to this operation. This operation will replace existing
142
F<files> or insert them at the end of the archive if they do not exist. If no
143
F<files> are specified, the archive is not modified.
147
Print the table of contents. Without any modifiers, this operation just prints
148
the names of the members to the standard output. With the F<v> modifier,
149
B<llvm-ar> also prints out the file type (B=bitcode, Z=compressed, S=symbol
150
table, blank=regular file), the permission mode, the owner and group, the
151
size, and the date. If any F<files> are specified, the listing is only for
152
those files. If no F<files> are specified, the table of contents for the
153
whole archive is printed.
157
Extract archive members back to files. The F<o> modifier applies to this
158
operation. This operation retrieves the indicated F<files> from the archive
159
and writes them back to the operating system's file system. If no
160
F<files> are specified, the entire archive is extract.
164
=head2 Modifiers (operation specific)
166
The modifiers below are specific to certain operations. See the Operations
167
section (above) to determine which modifiers are applicable to which operations.
173
When inserting or moving member files, this option specifies the destination of
174
the new files as being C<a>fter the F<relpos> member. If F<relpos> is not found,
175
the files are placed at the end of the archive.
179
When inserting or moving member files, this option specifies the destination of
180
the new files as being C<b>efore the F<relpos> member. If F<relpos> is not
181
found, the files are placed at the end of the archive. This modifier is
182
identical to the the F<i> modifier.
186
Normally, B<llvm-ar> stores the full path name to a file as presented to it on
187
the command line. With this option, truncated (15 characters max) names are
188
used. This ensures name compatibility with older versions of C<ar> but may also
189
thwart correct extraction of the files (duplicates may overwrite). If used with
190
the F<R> option, the directory recursion will be performed but the file names
191
will all be C<f>lattened to simple file names.
195
A synonym for the F<b> option.
199
Normally, B<llvm-ar> will not print the contents of bitcode files when the
200
F<p> operation is used. This modifier defeats the default and allows the
201
bitcode members to be printed.
205
This option is ignored by B<llvm-ar> but provided for compatibility.
209
When extracting files, this option will cause B<llvm-ar> to preserve the
210
original modification times of the files it writes.
214
use full path names when matching
218
This modifier instructions the F<r> option to recursively process directories.
219
Without F<R>, directories are ignored and only those F<files> that refer to
220
files will be added to the archive. When F<R> is used, any directories specified
221
with F<files> will be scanned (recursively) to find files to be added to the
222
archive. Any file whose name begins with a dot will not be added.
226
When replacing existing files in the archive, only replace those files that have
227
a time stamp than the time stamp of the member in the archive.
231
When inserting or replacing any file in the archive, compress the file first.
233
modifier is safe to use when (previously) compressed bitcode files are added to
234
the archive; the compressed bitcode files will not be doubly compressed.
238
=head2 Modifiers (generic)
240
The modifiers below may be applied to any operation.
246
For all operations, B<llvm-ar> will always create the archive if it doesn't
247
exist. Normally, B<llvm-ar> will print a warning message indicating that the
248
archive is being created. Using this modifier turns off that warning.
252
This modifier requests that an archive index (or symbol table) be added to the
253
archive. This is the default mode of operation. The symbol table will contain
254
all the externally visible functions and global variables defined by all the
255
bitcode files in the archive. Using this modifier is more efficient that using
256
L<llvm-ranlib|llvm-ranlib> which also creates the symbol table.
260
This modifier is the opposite of the F<s> modifier. It instructs B<llvm-ar> to
261
not build the symbol table. If both F<s> and F<S> are used, the last modifier to
262
occur in the options will prevail.
266
This modifier instructs B<llvm-ar> to be verbose about what it is doing. Each
267
editing operation taken against the archive will produce a line of output saying
274
The B<llvm-ar> utility is intended to provide a superset of the IEEE Std 1003.2
275
(POSIX.2) functionality for C<ar>. B<llvm-ar> can read both SVR4 and BSD4.4 (or
276
Mac OS X) archives. If the C<f> modifier is given to the C<x> or C<r> operations
277
then B<llvm-ar> will write SVR4 compatible archives. Without this modifier,
278
B<llvm-ar> will write BSD4.4 compatible archives that have long names
279
immediately after the header and indicated using the "#1/ddd" notation for the
284
The file format for LLVM Archive files is similar to that of BSD 4.4 or Mac OSX
285
archive files. In fact, except for the symbol table, the C<ar> commands on those
286
operating systems should be able to read LLVM archive files. The details of the
289
Each archive begins with the archive magic number which is the eight printable
290
characters "!<arch>\n" where \n represents the newline character (0x0A).
291
Following the magic number, the file is composed of even length members that
292
begin with an archive header and end with a \n padding character if necessary
293
(to make the length even). Each file member is composed of a header (defined
294
below), an optional newline-terminated "long file name" and the contents of
297
The fields of the header are described in the items below. All fields of the
298
header contain only ASCII characters, are left justified and are right padded
299
with space characters.
303
=item name - char[16]
305
This field of the header provides the name of the archive member. If the name is
306
longer than 15 characters or contains a slash (/) character, then this field
307
contains C<#1/nnn> where C<nnn> provides the length of the name and the C<#1/>
308
is literal. In this case, the actual name of the file is provided in the C<nnn>
309
bytes immediately following the header. If the name is 15 characters or less, it
310
is contained directly in this field and terminated with a slash (/) character.
312
=item date - char[12]
314
This field provides the date of modification of the file in the form of a
315
decimal encoded number that provides the number of seconds since the epoch
316
(since 00:00:00 Jan 1, 1970) per Posix specifications.
320
This field provides the user id of the file encoded as a decimal ASCII string.
321
This field might not make much sense on non-Unix systems. On Unix, it is the
322
same value as the st_uid field of the stat structure returned by the stat(2)
323
operating system call.
327
This field provides the group id of the file encoded as a decimal ASCII string.
328
This field might not make much sense on non-Unix systems. On Unix, it is the
329
same value as the st_gid field of the stat structure returned by the stat(2)
330
operating system call.
334
This field provides the access mode of the file encoded as an octal ASCII
335
string. This field might not make much sense on non-Unix systems. On Unix, it
336
is the same value as the st_mode field of the stat structure returned by the
337
stat(2) operating system call.
339
=item size - char[10]
341
This field provides the size of the file, in bytes, encoded as a decimal ASCII
342
string. If the size field is negative (starts with a minus sign, 0x02D), then
343
the archive member is stored in compressed form. The first byte of the archive
344
member's data indicates the compression type used. A value of 0 (0x30) indicates
345
that no compression was used. A value of 2 (0x32) indicates that bzip2
346
compression was used.
350
This field is the archive file member magic number. Its content is always the
351
two characters back tick (0x60) and newline (0x0A). This provides some measure
352
utility in identifying archive files that have been corrupted.
356
The LLVM symbol table has the special name "#_LLVM_SYM_TAB_#". It is presumed
357
that no regular archive member file will want this name. The LLVM symbol table
358
is simply composed of a sequence of triplets: byte offset, length of symbol,
359
and the symbol itself. Symbols are not null or newline terminated. Here are
360
the details on each of these items:
364
=item offset - vbr encoded 32-bit integer
366
The offset item provides the offset into the archive file where the bitcode
367
member is stored that is associated with the symbol. The offset value is 0
368
based at the start of the first "normal" file member. To derive the actual
369
file offset of the member, you must add the number of bytes occupied by the file
370
signature (8 bytes) and the symbol tables. The value of this item is encoded
371
using variable bit rate encoding to reduce the size of the symbol table.
372
Variable bit rate encoding uses the high bit (0x80) of each byte to indicate
373
if there are more bytes to follow. The remaining 7 bits in each byte carry bits
374
from the value. The final byte does not have the high bit set.
376
=item length - vbr encoded 32-bit integer
378
The length item provides the length of the symbol that follows. Like this
379
I<offset> item, the length is variable bit rate encoded.
381
=item symbol - character array
383
The symbol item provides the text of the symbol that is associated with the
384
I<offset>. The symbol is not terminated by any character. Its length is provided
385
by the I<length> field. Note that is allowed (but unwise) to use non-printing
386
characters (even 0x00) in the symbol. This allows for multiple encodings of
393
If B<llvm-ar> succeeds, it will exit with 0. A usage error, results
394
in an exit code of 1. A hard (file system typically) error results in an
395
exit code of 2. Miscellaneous or unknown errors result in an
400
L<llvm-ranlib|llvm-ranlib>, ar(1)
404
Maintained by the LLVM Team (L<http://llvm.org/>).