Inicio Explorar Axuda
Iniciar sesión
domonkos
/
ed
Seguir 1
Destacar 0
Fork 0
Ficheiros Incidencias 0 Pull Requests 0 Wiki

ed(1) is the standard text editor.

Rama: master
Ramas Etiquetas
ed
ZIP TAR.GZ
sean 2ab97f145f Shuffling bits of code around, and making mkstemp work for %!s(int64=2) %!d(string=hai) anos
doc 2ab97f145f Shuffling bits of code around, and making mkstemp work for %!s(int64=2) %!d(string=hai) anos
.gitignore 98ef09326e gitignore: igonore .Z tarballs %!s(int64=4) %!d(string=hai) anos
Makefile 335051c9b1 Large patchset. %!s(int64=2) %!d(string=hai) anos
README.md 335051c9b1 Large patchset. %!s(int64=2) %!d(string=hai) anos
ed.c 2ab97f145f Shuffling bits of code around, and making mkstemp work for %!s(int64=2) %!d(string=hai) anos
utf.c 335051c9b1 Large patchset. %!s(int64=2) %!d(string=hai) anos
utf.h 2ab97f145f Shuffling bits of code around, and making mkstemp work for %!s(int64=2) %!d(string=hai) anos
utfio.c 98d675472e all: many small fixes to keep -Wall quiet. %!s(int64=4) %!d(string=hai) anos
utfio.h 98d675472e all: many small fixes to keep -Wall quiet. %!s(int64=4) %!d(string=hai) anos

README.md

ED, the standard editor, now UTF-8/Unicode-aware.

This is a port of the Research Unix (v10) ED, with UTF-8/Unicode support. In particular, regexes and subs are fully Unicode-aware.

The sources that this port began with were scavenged from the Research Unix sources available from The Unix Heritage Society.

The original manpage is in the doc subdirectory. Where the behaviour of this ed differs significantly from that described in the manpage is noted below.

The Makefile provides a simple way to build and install an ed binary. make clean && make will get you a standalone binary ed that can be installed anywhere. make install will install ED as ed and e, with corresponding manpages. Edit the Makefile to change where all this goes.

Differences

This ed differs from the ED described in the manpage as follows:

  • This ed consumes and outputs UTF-8 Unicode. Internally, all "characters" are stored and manipulated as Unicode code-points represented as int-s. This means that even emojis can appear in regexes and character classes.

  • In list mode, this ed displays non-printing codepoints in the ASCII range, and all Unicode codepoints above the printing ASCII range as a sequence \ddd of octal triplets, one for each byte in the UTF-8 encoding of the codepoint. Octal is the traditional ED list mode format, and it is a surprisingly good way of reading raw UTF-8, as all multi-byte sequences begin with byte \3.., and all subsequent sequence bytes begin \2... (Non-printing) ASCII bytes are all of the form \0.., or \1...

  • The original manpage states that ED discards all text in a file that appears between the last newline and the end-of-file. The original ed.c source code does not do that. Instead it supplies a newline at the end of the file, and notifies the user with a: '\n' appended message. This behaviour has been retained.