aboutsummaryrefslogtreecommitdiff
path: root/bash
Commit message (Collapse)AuthorAgeFilesLines
* Move .bashrc interactivity test to topTom Ryder2016-08-221-6/+6
|
* Remove double-up import of completionsTom Ryder2016-08-211-17/+0
|
* Quote a bash->sh variableTom Ryder2016-08-211-1/+1
|
* Many stylistic fixes/optimisationsTom Ryder2016-08-216-90/+34
| | | | In particular, limit the verbose >=2.05 test to just one file: .bashrc
* Force type for Git ahead/behind countsTom Ryder2016-08-211-4/+6
|
* Use dynamic completion if availableTom Ryder2016-08-213-41/+75
|
* Anticipate Bath versions > 4Tom Ryder2016-08-211-1/+1
|
* Use clearer quoting for status flagsTom Ryder2016-08-211-6/+6
|
* Target terminfo first, then termcapTom Ryder2016-08-211-9/+9
|
* Break up long printf linesTom Ryder2016-08-211-2/+4
|
* Slightly quicker git/svn detectionTom Ryder2016-08-211-15/+6
|
* Add some upstream checking to Git promptTom Ryder2016-08-211-3/+19
| | | | Also add some comments to be a bit less cryptic
* Use terser prompt status flag checksTom Ryder2016-08-211-8/+4
|
* Resolve an issue with incorrect prompt infoTom Ryder2016-08-211-0/+3
| | | | | | | | | | | | | From `man 1 git-config`: > diff.autoRefreshIndex > When using git diff to compare with work tree files, do not > consider stat-only change as changed. Instead, silently run git > update-index --refresh to update the cached stat information for > paths whose contents in the work tree match the contents in the > index. This option defaults to true. Note that this affects only > git diff Porcelain, and not lower level diff commands such as git > diff-files.
* Update path() completionTom Ryder2016-08-201-3/+3
|
* Port path() to POSIX shTom Ryder2016-08-201-180/+0
| | | | That was a lot easier than I thought
* Port fnl() to POSIX sh script fnl(1)Tom Ryder2016-08-201-41/+0
| | | | | No real compelling reason to make it a shell function in the first place.
* Remove readv() and readz()Tom Ryder2016-08-202-29/+0
| | | | Neither are terribly useful
* Port vr(1) to POSIX shTom Ryder2016-08-201-59/+0
|
* Port lhn() to POSIX shTom Ryder2016-08-201-7/+0
|
* Port sd() to POSIX shTom Ryder2016-08-201-109/+0
|
* Port pd() to POSIX shTom Ryder2016-08-201-53/+0
|
* Port ud() to POSIX shTom Ryder2016-08-201-50/+0
|
* Port bd() to POSIX shTom Ryder2016-08-201-78/+0
|
* Port all bash_profile.d scripts to POSIX shTom Ryder2016-08-204-71/+0
| | | | | Also require flag files in ~/.welcome for displaying or not displaying login stuff
* Move OLDPWD setting to POSIX sh dirTom Ryder2016-08-201-3/+0
| | | | With an attempt at correct trapping; may still require tweaking
* Change cd() into POSIX shell scriptTom Ryder2016-08-191-32/+0
|
* Add lhn()Tom Ryder2016-08-181-0/+7
|
* Drastically limit amount of history keptTom Ryder2016-08-181-4/+4
| | | | | | Reading really large files seems to be slow in Bash in general, but it's particularly bad in 4.4rc1. I keep encrypted snapshots of my HISTFILE on my home machine, so it's just a little extra step to search them.
* Don't do Git prompt work if not in working treeTom Ryder2016-08-181-0/+7
|
* Use explicit -n test within [[ ]]Tom Ryder2016-08-182-3/+3
| | | | | Bash doesn't require it, but there's no real advantage to it and it's a better habit for complying with e.g. pdksh, which does
* Use consistent syntax for sh source loopsTom Ryder2016-08-182-9/+5
|
* Port grep() and ls() to POSIX shTom Ryder2016-08-182-55/+0
| | | | | | Check capabilities of wrapped programs at runtime, not declaration time. Also do away with the silly GREP_COLORS and GREP_OPTS variables. Considering doing the same with LS_COLORS.
* Update bashrc/zshrc with POSIX funcs sourceTom Ryder2016-08-171-1/+1
|
* Simplify end-of-file loop in bashrcTom Ryder2016-08-171-24/+8
|
* Use terser loop syntax in ~/.bash_profileTom Ryder2016-08-171-3/+2
|
* Source ~/.shrc.d in Bash if not login shellTom Ryder2016-08-171-0/+12
|
* Move interactive tests back into *rc filesTom Ryder2016-08-172-4/+10
| | | | | | I don't fully understand why I have to do it this way yet, but if I don't, calling SSH with a command raises "stdin: not a terminal" because ~/.bashrc was called.
* Move -i test in Bash to end of ~/.bash_profileTom Ryder2016-08-172-9/+4
|
* Move simple Bash/pdksh functions into POSIX shTom Ryder2016-08-1716-171/+0
| | | | | | | Have only translated the scripts that translate readily into POSIX sh for now. More complex stuff like that bd/pd/sd/ud navigation for Bash doesn't port as easily, mostly because there isn't an analogue for the "local" keyword in POSIX.
* Keep grep(1) option metadata in cacheTom Ryder2016-08-151-9/+11
| | | | | | | | | It's stupid to run `grep --help` once per shell (twice for login shells!) when it's so unlikely to change, and way faster to check for the presence or absence of hint files rather than pattern-match the output with the shell. ls(1) will get the same treatment in a minute.
* Don't use git: prompt prefix if PROMPT_VCS unsetTom Ryder2016-08-151-2/+3
| | | | | All of my personal stuff is in Git, so this is only really applicable at work
* Remove square brackets around promptTom Ryder2016-08-151-1/+1
| | | | Just noise
* Quote a PS1 character to appease shellcheckTom Ryder2016-08-141-1/+1
|
* Two extra commentsTom Ryder2016-08-141-0/+2
|
* Rearrange prompt variable assignments a bitTom Ryder2016-08-141-10/+12
|
* Use --quiet flag for prompt git-rev-parse(1) callTom Ryder2016-08-141-1/+1
|
* Use porcelain commands for Git promptTom Ryder2016-08-131-27/+9
| | | | | | | | | | | | Checking the output of `git status -z` works, but to be correctly handled requires using a null delimiter for read. Because I want to port this to OpenBSD pdksh (which doesn't have the -d option for read), this seems to be a workable alternative. It also enables me to remove the process substitution (another thing pdksh doesn't support), and the array of flags. I haven't yet tested this for speed.
* Fix comment typoTom Ryder2016-08-111-1/+1
|
* Test bc(1) for non-POSIX -qTom Ryder2016-08-111-0/+3
|