]> git.saurik.com Git - apt-legacy.git/blame_incremental - COMPILING
Figured out a much simpler fix for the cache rebuild performance.
[apt-legacy.git] / COMPILING
... / ...
CommitLineData
1General Information
2~~~~~~~~~~~~~~~~~~~
3To compile this you need a couple things
4 - A working POSIX system with working POSIX gcc, g++, make (GNU),
5 ar, sh, awk and sed in the path
6 - GNU Make 3.74 or so, -- normal UNIX make will NOT work
7 * Note 3.77 is broken.
8 - A working ANSI C++ compiler, this is not g++ 2.7.*
9 g++ 2.8 works OK and newer egcs work well also. Nobody has tried it
10 on other compilers :< You will need a properly working STL as well.
11 - A C library with the usual POSIX functions and a BSD socket layer.
12 If your OS conforms to the Single Unix Spec then you are fine:
13 http://www.opengroup.org/onlinepubs/7908799/index.html
14 - Refer to the Build-Depends information in debian/control for
15 additional requirements (some of which are Debian-specific)
16
17** NOTICE **
18The C++ global constructors do not link correctly when using non-shared
19libraries. This is probably the correct behavior of the linker, but I have
20not yet had time to devise a work around for it. The correct thing to
21do is add a reference to debSystem in apt-pkg/init.cc,
22assert(&debSystem == 0) would be fine for instance.
23
24Guidelines
25~~~~~~~~~~
26I am not interested in making 'ultra portable code'. I will accept patches
27to make the code that already exists conform more to SUS or POSIX, but
28I don't really care if your not-SUS OS doesn't work. It is simply too
29much work to maintain patches for dysfunctional OSs. I highly suggest you
30contact your vendor and express intrest in a conforming C library.
31
32That said, there are lots of finicky problems that must be dealt with even
33between the supported OS's. Primarily the path I choose to take is to put
34a shim header file in build/include that transparently adds the required
35functionality. Patches to make autoconf detect these cases and generate the
36required shims are OK.
37
38Current shims:
39 * C99 integer types 'inttypes.h'
40 * sys/statvfs.h to convert from BSD/old-glibc statfs to SUS statvfs
41 * rfc2553 hostname resolution (methods/rfc*), shims to normal gethostbyname.
42 The more adventurous could steal the KAME IPv6 enabled resolvers for those
43 OS's with IPv6 support but no rfc2553 (why?)
44 * define _XOPEN_EXTENDED_SOURCE to bring in h_errno on HP-UX
45 * socklen_t shim in netdb.h if the OS does not have socklen_t
46
47The only completely non-shimmed OS is Linux with glibc2.1, glibc2.0 requires
48the first three shims.
49
50Platform Notes
51~~~~~~~~~~~~~~
52Debian GNU Linux 2.1 'slink'
53Debian GNU Linux 'potato'
54Debian GNU Linux 'woody'
55 * All Archs
56 - Works flawlessly
57 - You will want to have debiandoc-sgml and docbook2man installed to get
58 best results.
59 - No IPv6 Support in glibc's < 2.1.
60
61Sun Solaris
62 SunOS cab101 5.7 Generic_106541-04 sun4u sparc
63 SunOS csu201 5.8 Generic_108528-04 sun4u sparc
64 - Works fine
65 - Note, no IPv6 Support, OS lacks RFC 2553 hostname resolution
66
67OpenBSD
68 OpenBSD gsb086 2.5 CMPUT#0 i386 unknown
69 OpenBSD csu101 2.7 CMPUT#1 i386 unknown
70 - OS needs 'ranlib' to generate the symbol table after 'ar'.. (not using
71 GNU ar with the gnu tool chain :<)
72 - '2.5' does not have RFC 2553 hostname resolution, but '2.7' does
73 - Testing on '2.7' suggests the OS has a bug in its handling of
74 ftruncate on files that have been written via mmap. It fills the page
75 that crosses the truncation boundary with 0's.
76
77HP-UX
78 HP-UX nyquist B.10.20 C 9000/780 2016574337 32-user license
79 - Evil OS, does not conform very well to SUS
80 1) snprintf exists but is not prototyped, ignore spurious warnings
81 2) No socklen_t
82 3) Requires -D_XOPEN_SOURCE_EXTENDED for h_errno
83 configure should fix the last two (see above)
84 - Note, no IPv6 Support, OS lacks RFC 2553 hostname resolution
85