vaiyebe / heirloom Goto Github PK
View Code? Open in Web Editor NEWThis project forked from kallisti5/heirloom
Heirloom UNIX tools, from grandma's attic to your /bin
This project forked from kallisti5/heirloom
Heirloom UNIX tools, from grandma's attic to your /bin
README file for the Heirloom Toolchest ====================================== For a description what the Heirloom Toolchest is, see the file intro.txt, the online introduction, or, if the tools are already installed, the corresponding manual page intro(1). Note that the path names mentioned in these documents may differ from those you select when building the tools as described below. It is assumed that you know the contents of the introduction as you read on. For licensing terms, see the files in the LICENSE subdirectory. How to install ============== System environments that are known to support the Heirloom Toolchest are: Linux Kernel 2.4 and above; glibc 2.2 and above, diet libc, uClibc; gcc, Intel C Solaris 8 and above; Sun C, gcc Open UNIX 8.0.0; native (cc), LKP (gcc) HP-UX B.11.11, B.11.22, B.11.23; HP-UX ANSI C, gcc AIX 5.1; VisualAge C, gcc FreeBSD 4.9, 5.2.1, 6.1 NetBSD 1.6, 2.0, 3.0 OpenBSD 3.4, 3.8 DragonFlyBSD 1.3.7-DEVELOPMENT Mac OS X 10.4.3 and above The following prerequisites are necessary: yacc and lex Several of the included yacc and lex programs need to access internals of the yacc and lex implementations beyond the respective POSIX specifications. This has been made to work with Unix yacc and lex as well as with many versions of Berkeley yacc, bison, and flex. However, recent changes in these programs may result in incompatibilities. If this happens, install the "Heirloom Development Tools" package first and use the yacc and lex versions provided with it. You could also fix the problem on your own, but now that portable Open Source versions of Unix yacc and lex are available, such changes can no longer be considered important, and will not be made as part of the regular maintenance of this project. The following prerequisites are optional: Bourne shell It is recommended that the Heirloom Bourne shell from <http://heirloom.sourceforge.net/sh.html> is installed first. The tools will work with any shell, but use of the Bourne shell ensures maximum compatibility with traditional Unix behavior. gzip <http://www.gzip.org/>. Invoked by cpio/pax, file, and tar when a gzip compressed file is encountered. bzip2 <http://sources.redhat.com/bzip2>. Invoked by cpio/pax, file, and tar when a bzip2 compressed file is encountered. libz <http://www.gzip.org/zlib>. It is needed if cpio/pax should be able to create deflate compressed entries in zip files; reading of such entries works without it. libbz2 <http://sources.redhat.com/bzip2>. It is needed if cpio/pax should be able to read or create bzip2 compressed entries in zip files. Unless you are a developer, it is strongly recommended that you use the RPM packaging system to build and install the Heirloom Toolchest on Linux. This requires the "heirloom-devtools" package to be installed; its source code is available from <http://heirloom.sourceforge.net/devtools.html>. Then simply execute rpmbuild -tb heirloom-<version>.tar.bz2 wait a few minutes and install the resulting rpm that is found in /usr/src/redhat/RPMS/i386/heirloom-<version>-1.i386.rpm (or wherever your vendor's setup puts such files). Then access the tools by prefixing your PATH with /usr/5bin and skip the rest of this section. (With older versions of rpm you need to use 'rpm' instead of 'rpmbuild'.) The tools in the package will automatically use the Heirloom Bourne shell if it is installed at compilation time; the Heirloom Toolchest package then requires the Heirloom Bourne shell package to be also installed. If your system lacks rpm or if you want to make changes to the Heirloom Toolchest, you have to build it manually. The first thing to understand is the build system. This is actually quite simple: Every directory contains a file named Makefile.mk that includes the directory-specific make instructions. To generate the real Makefile, configuration settings are prepended to the directory-specific file. You have to edit these configuration settings before you start compiling; they are located in the file build/mk.config which is also in make syntax. Follow the descriptions in this file and select appropriate values for your system. Be very careful at this step! If you overwrite your system utilities or put the Heirloom Toolchest at an early place in your system's PATH, some of your shell scripts or of those of your vendor may cease to work. This is because utilities from the Heirloom Toolchest may behave differently than the ones supplied by your vendor. You may wish to correct this later, but the first build is definitively not the right time. - If you want to stay with the default /usr/5bin path on Solaris, be sure to remove the existing symbolic link to /usr/bin. If you don't know what path names to select, get a competent person to assist you. I won't assist at this point and do, as described in the license terms, not warrant for any problems caused by anyone. You have been warned. After finishing the configuration, type 'make' and 'make install', then use the tools found at the location you selected before. The Heirloom Toolchest can also be compiled using dietlibc <http://www.fefe.de/dietlibc/>. This will result in relatively small statically linked executables; the tools will not have multibyte functionality, though. Note that you have to compile ncurses, optionally also zlib and libbz2, and to install the libraries in an appropriate location for 'diet' before you can start to build the Heirloom tools. After installation ================== Some vendors or system administrators predefine an alias or a shell function for 'ls' in the shell startup files. If you type 'ls' and only get something like 'ls: illegal option -- -' from the Heirloom 'ls' command, try 'unalias ls' or 'unset ls'. The Heirloom 'man' command will display the Heirloom manual pages first, as configured in /etc/default/man. If there is a 'MANPATH' environment variable present however, it will override this setting and you get whatever MANPATH points to. In this situation, either unset MANPATH or adjust it to include the Heirloom manual path first (normally /usr/share/man/5man, see build/mk.config.) Porting the Heirloom Toolchest ============================== If you want to port the Heirloom Toolchest to a new environment, check that the following conditions are satisfied: - The environment is a Unix-like system, e.g. the file system is case-sensitive, there is a unique dev_t and ino_t combination for each file i-node, there are at least POSIX.1 calls etc. The code heavily relies on this, and I'm not interested in changes for non-Unix environments for a variety of reasons. - The compiler supports 'long long' and distinguished 8-bit, 16-bit, 32-bit, and 64-bit data types. The system stdio library supports printf("%lld"). - The compiler does not insert padding between two structure fields of type 'char' or array of 'char'. - You know how to read directory contents without opendir() and friends, using a file descriptor obtained from open(). On most systems, this involves either the System V function getdents() or the BSD function getdirentries(). It may also be possible to use read(), although this should only be the last resort. Such functionality may or may not be documented; search for it in /usr/include if in doubt. If getdents() or getdirentries() exist, you can probably adjust the existing code in 'libcommon/getdir.c'. - You know how to obtain a process listing in C. This may be as easy as reading /proc (you can probably re-use existing code then), but might also require that you know how to read structures from /dev/kmem in the worst case. Documentation for this is usually sparse; look around in /usr/include or search the web for existing open source code that performs such lookups (e.g. 'lsof', 'libgtop'). - You know how to code tape device control (<sys/mtio.h> or the like). This is usually documented, so if in doubt, just read the manual pages and header files. If all of these are true, you can expect the work to be done within a day or so. The Heirloom Toolchest is coded with The Open Group's 'Single Unix Specification', Version 3, 2001, Base Definitions and Headers, and System Interfaces, in mind. If the system is Unix-like, missing functionality can usually be emulated. For a number of functions, such as those for wide-character support, utmpx access, simplified signal management (sigset() etc.), emulation code is already present in the libcommon and libwchar directories; enable it if necessary. Hacking the Heirloom Toolchest ============================== One of the considerations the Heirloom Toolchest is designed along is accessibility for other developers. Each utility was made largely independent from other ones, to the extent of duplicate functionality at some places. There are just three libraries you have to know about: 1. The 'common' library, found in libcommon, provides the following functionality: - Traditional ('simple') regular expressions, available as full source to #include in the old-style ed-hack regexp.h and as normal library functions in regexpr.h. You should definitively use the latter if you want to use simple REs in new applications. - Traditional file name matching; manually include the prototype 'extern int gmatch(const char *, const char *);' if you want to use it. - SVID4-style error reporting; #include <pfmt.h>. - Locale-independent character class functions in <asciitype.h>. These are necessary e.g. to recognize programming language syntax white space if LC_CTYPE is used otherwise (if you do not want to handle no-break-space characters etc. as word delimiters). - Functions for reading directory entries with a file descriptor obtained from open(), in <getdir.h>. These are only useful if you descend a directory hierarchy with fchdir() and have to care about the limit of open file descriptors, as the 'file' and 'rm' tools; otherwise, stay with the standard readdir(). - A getopt() that has the traditional functionality, and just the traditional functionality. This is here mainly because GNU getopt() is so annoyingly non-standard. - Functions for buffered input and output in <iblok.h> and <oblok.h>, respectively. They can be used as a replacement for stdio in simple cases; the main reason for their existence is the desire to read lines of arbitrary length without the slow-down caused by getc(). 2. A modified version of Caldera's 'UNIX(R) Regular Expression Library' in the libuxre directory. This is used for (mostly) POSIX.2-compatible regular expression matching. It provides a syntax without too much extension fuss, and good performance, since it includes both a DFA and a NFA style matcher and uses either of them depending on the expression and the need for subexpression locations. 3. The widechar portability library in the libwchar directory. This is only used for system environments that do not provide appropriate wide character functionality in their own libc. It currently just does simple copying between bytes and wide characters and thus does not handle anything than plain ASCII well. If other wide character functions than those currently present are used in new code, it will be necessary to add those function to libwchar as well. I make no premises about keeping the interfaces of these libraries compatible in upcoming releases, so if you want to use them in other projects, just fork off your own version and don't rely on mine. To go on, you now only need to know about the sources in the directory of the utility you want to hack. This should make it not too difficult, even though I've generally stayed with good old Unix programming style (few comments, short variable names - but not too much to read either). Extending the Heirloom Toolchest ================================ If you want to add an extension to a utility, just go on and do it - as you've read in the section above, accessibility for hacking is intended to be easy. For my version, though, the concept is to have not too many extensions. I sometimes include extensions because the functionality is really expected these days (such as diff -u), adds a real advantage for interactive use (such as colored output for ls), or is widely used for convenience and does not add much overhead (such as grep -r). I am not much interested in extensions that change the behavior of tools for use in scripts, though. Such extensions lead to subtly incompatible shell scripts, and these should be avoided by all means. It is much better to make a fresh start by creating a new utility with a different name for such cases. Such utilities are then beyond the scope of this collection. In general, chances are rather low that I will accept your extension for my version. But that doesn't mean that you should not do it for yourself. Reporting bugs ============== Before you report behavior that you suspect to be a bug, check at least the following to be sure not to report a feature: - Does it leave a core dump? If yes, it's certainly a bug (unless you sent a signal, of course). - Read the manual page and the source for the utility carefully. Did you use the utility properly? - If the utility is used within a shell script or with another user's environment, try how it behaves if used standalone. - If the behavior disturbs you just because your vendor's utilities do something different, that alone is no reason to report it. The Heirloom Toolchest is oriented on System V behavior. Verify that real System V derivatives behave differently before reporting. - If you are using one of the POSIX personalities and the specification says something different, it is likely to be a bug. But be careful to read the respective specification itself; some incompatibilities to the wording of the specification may even exist in certified systems, so you cannot rely on a vendor's implementation or documentation here. If you went through these statements and you're still sure that it is a bug, report it. Gunnar Ritter Freiburg i. Br. Germany <http://heirloom.sourceforge.net> <[email protected]> 2/24/07
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.