Package Details: fish-git 2:2.3.0.r233.g7ef40f4-1

Git Clone URL: https://aur.archlinux.org/fish-git.git (read-only)
Package Base: fish-git
Description: User friendly shell intended mostly for interactive use.
Upstream URL: http://fishshell.com
Keywords: shell
Licenses: GPL, BSD, LGPL, MIT
Conflicts: fish, fish-shell
Provides: fish, fish-shell
Submitter: SanskritFritz
Maintainer: SanskritFritz
Last Packager: SanskritFritz
Votes: 51
Popularity: 0.691059
First Submitted: 2010-11-17 22:27
Last Updated: 2016-06-07 23:13

Dependencies (7)

Required by (11)

Sources (1)

Latest Comments

SanskritFritz commented on 2016-04-20 07:23

etu just recompile the package and you got it.

etu commented on 2016-04-20 06:59

2.3b1 is out :)

SanskritFritz commented on 2016-04-17 18:06

Guys, thanks, done.

RaphaelCosta commented on 2016-04-17 00:38

Emergency:
Updated today and all logins were broken complaining about missing libpcre2-32.so.0. Tested with no .config/fish. Using community/fish works without pcre2.

Users afflicted by this issue would have to chroot. I apologize but I'm flagging this out of date. There must be a problem somewhere but for now you need to add community/pcre2 to list of dependencies until it's fixed.

alfa commented on 2015-10-13 22:03

This currently builds with the bundled pcre2. If you want to use system-pcre2 (which presumably arch proper would want to do once 2.2.0+1 is released), you need to build with "--without-included-pcre2" and add pcre2 to the deps.

SanskritFritz commented on 2015-09-18 21:18

Important: after ncurses upgrade, fish-shell must be rebuilt.

SanskritFritz commented on 2015-02-06 06:21

Thanks for the heads-up beatgammit.

beatgammit commented on 2015-02-06 03:55

Could you add `pkgfile` as an optdepends? I added a command-not-found handler for Arch Linux and it depends on it: https://github.com/fish-shell/fish-shell/pull/1925

SanskritFritz commented on 2014-10-04 17:03

Ok, for the time being I remove docker.fish. Future fish versions will provide a vendor_completions folder, we will have to pursuade the docker packagers to change the destination of docker.fish.

SanskritFritz commented on 2014-09-29 18:47

beatgammit thanks for the modified PKGBUILD.
Also see https://github.com/fish-shell/fish-shell/issues/1726 and comment if necessary.

beatgammit commented on 2014-09-29 17:53

@SanskritFritz

I understand that, but the official docker package in the official repositories packages that file and the official fish package doesn't have those completions (it is slated for 2.2 AFAICT). So, to get the PKGBUILD to work today, I've made the change.

I understand if you're not interested, but I'll leave this PKGBUILD here for others in case they're too lazy to remove it for themselves:

http://pastebin.com/1LMtJymN

Once fish 2.2 is released, a bug should be filed to decide where this file belongs.

beatgammit commented on 2014-09-29 17:52

@SanskritFritz

I understand that, but the official docker package in the official repositories packages that file and the official fish package doesn't have those completions (it is slated for 2.2 AFAICT). So, to get the PKGBUILD to work today, I've made the change.

I understand if you're not interested, but I'll leave this PKGBUILD here for others in case they're too remove it for themselves:

http://pastebin.com/1LMtJymN

Once fish 2.2 is released, a bug should be filed to decide where this file belongs.

beatgammit commented on 2014-09-29 17:48

To get this to build with docker installed, I manually rm'd docker.fish in the PKGBUILD.

Is this something that should be upstreamed, or should this PKGBUILD just ignore that file?

For the lazy, here's my PKGBUILD: http://pastebin.com/1LMtJymN

SanskritFritz commented on 2014-09-29 17:47

beatgammit could you read the previous comments please.

beatgammit commented on 2014-09-29 17:44

To get this to build with docker installed, I manually rm'd docker.fish in the PKGBUILD.

Is this something that should be upstreamed, or should this PKGBUILD just ignore that file?

beatgammit commented on 2014-09-29 17:43

To get this to build with docker installed, I manually rm'd docker.fish. Is this something that should be upstreamed, or should this PKGBUILD just ignore that file?

SanskritFritz commented on 2014-09-24 16:13

I'd say fish should pull it, probably the docker people do a better job maintaining it. I certainly would choose their version, unless... :)

beatgammit commented on 2014-09-24 15:44

@SanskritFritz - which upstream should this be reported to? Docker's policy seems to be to ship completions for all shells (at least they include them in the Ubuntu PPA [1]).

Since the "official" community package includes the completions, perhaps they should not be installed by this package? I'm don't really know what the best policy is here...

[1] - https://github.com/docker/docker/pull/4853

SanskritFritz commented on 2014-09-24 06:59

Well, this is the policy of fish, it installs completions itself. On the other hand, I can understand that docker upstream thinks thy can do it better, so they provide one. This should be resolved upstream.

euclio commented on 2014-09-24 03:18

This package appears to conflict with docker in community. Both packages attempt to install /usr/share/fish/completions/docker.fish

SanskritFritz commented on 2014-08-09 20:14

Honestly, I don't know what you're saying, it builds and works here flawlessly.

lowfatcomputing commented on 2014-08-09 15:35

In order for this to build FISH-BUILD-VERSION-FILE needed to be generated among other things. Here is an updated PKGBUILD:

http://pastebin.com/hcTKEJhq

SanskritFritz commented on 2013-10-29 12:49

Kamek press ctrl-z in vim and see tha error messages. Then report it upstream please if you feel there is a problem in fish. I don't have any problems running vim during my daily use.

Kamek commented on 2013-10-29 02:01

Appears to have some problems running Vim. Anyone else getting this? It said fish doesn't use standard shell operators or some such as that it goes by too fast to read.

SanskritFritz commented on 2013-06-12 11:50

bananaoomarang please read this:
https://wiki.archlinux.org/index.php/Arch_CVS_%26_SVN_PKGBUILD_guidelines#VCS_sources

bananaoomarang commented on 2013-06-12 11:45

Doesn't seem to compile properly as it doesn't use normal conventions for cloning a git repo, makepkg just downloads the .git file, not what we want! I made a new package build which seems to work, if anyone else has the same problem.

http://pastebin.com/JAHJjSfZ

SanskritFritz commented on 2013-04-06 20:43

pkgver() is used, it provides last commit date.

SanskritFritz commented on 2012-12-27 13:49

@spicey Thanks for the heads-up, updated.

spicey commented on 2012-12-27 11:56

make_completions.py is also recently removed, so that sed call needs to go as well.

SanskritFritz commented on 2012-12-12 21:55

@mikezackles: Indeed, thanks for the heads-up.

mikezackles commented on 2012-12-12 18:37

internalize_scripts.py doesn't appear to exist in the latest checkout, so I had to remove the sed call to swap python for python2 on line 38.

fosskers commented on 2012-07-21 15:38

Thanks for going out of your way to ask about it.

SanskritFritz commented on 2012-07-21 14:44

Ok, arm added, according to this: http://archlinux.2023198.n4.nabble.com/AUR-and-unsuported-architectures-td4655207i20.html#a4657627

fosskers commented on 2012-07-21 06:53

Forgive my ignorance about base-devel.

Arch Linux Arm users _do_ install from _this_ AUR, but this is usually done by passing flags to yaourt that ignore architecture.
If 'arm' is included for packages that have proven to build on arm, then this architecture ignoring would not be necessary.

Either way, no harm could come by adding 'arm' to the PKGBUILD. It wouldn't even require a dedicated push by SanskritFritz. He could just include it the next time he decides to make a push from the git repo.

Note that there are many other packages here that _do_ include 'arm' in their architecture field.

fosskers commented on 2012-07-21 06:52

Forgive my ignorance about base-devel.

Arch Linux Arm users _do_ install from _this_ AUR, but this is usually done by passing flags to yaourt that ignore architecture.
If 'arm' is included for packages that have proven to build on arm, then this architecture ignoring would not be necessary.

Either way, no harm could come by adding 'arm' to the PKGBUILD. It wouldn't even require a dedicated push by SanskritFritz. He could just include it the next time he decides to make a push from the git repo.

spirtbrat commented on 2012-07-21 05:55

All packages from base-devel group are assumed to be installed. Check the wiki about that.
Also check archlinuxarm.org if you need packages that build on various ARM devices.

fosskers commented on 2012-07-21 03:52

Hey again.

This builds on arm devices, so could you add 'arm' to the architecture field?

Also, could you add `autoconf` and `make` as makedeps?

SanskritFritz commented on 2012-07-12 11:42

Yeah, that's why I bumped the version number.

fosskers commented on 2012-07-12 10:27

Builds no problem now. Thanks.

SanskritFritz commented on 2012-07-11 12:35

https://github.com/fish-shell/fish-shell/issues/223

fosskers commented on 2012-07-11 12:30

Building explodes with the following error (omitting before and after stuff):
.
.
.
checking for /usr/local/lib library directory... yes
checking for /usr/local/bin command directory... yes
checking for g++... g++
checking whether the C++ compiler works... no
configure: error: in `/home/colin/fish-shell-git-25756/fish-shell-git/src/fish-shell-build':
configure: error: C++ compiler cannot create executables
See `config.log' for more details
==> ERROR: A failure occurred in build().
Aborting...
.
.
.

SanskritFritz commented on 2012-06-16 20:05

koral please use the fish-shell-git package

koral commented on 2012-06-16 17:43

New development upstream is now: https://github.com/fish-shell/fish-shell .
Could you please update the package accordingly, or disown it ?

SanskritFritz commented on 2012-06-15 14:04

zwastik, this is a development version, git head, beta testing phase. What did you expect?

zwastik commented on 2012-06-15 13:50

I installed this package and when running
fish -v
I get fish, version 1.23.1 =S, that's supposed to be the 2009 stable version!

SanskritFritz commented on 2012-06-12 23:54

http://sourceforge.net/mailarchive/message.php?msg_id=29387090
Repository was moved to github, under the name fish-shell. I created this new package with the new name. The old one was merged with this one.

SanskritFritz commented on 2012-06-12 23:53

thejcs please delete your comment, packages were merged.

SanskritFritz commented on 2012-06-12 23:52

thejcs please delete your comments, packages were merged.

Anonymous comment on 2012-06-11 20:54

Relevant discussion, regarding the missing 'help_doc/man/man1/history.1' file:

https://aur.archlinux.org/packages.php?ID=59641

SanskritFritz commented on 2012-06-11 18:28

http://sourceforge.net/mailarchive/message.php?msg_id=29387090
Repository moved to github, under the name fish-shell. I created a new package with the new name. This one will be deleted or merged with the new one. Please use the new one from now.

dpaneda commented on 2012-06-10 20:55

It can be reproduced with a clean checkout of the fishfish. I just follow the fishfish compilation steps and it fails:
http://ridiculousfish.com/shell/beta.html

This is therebefore an error of the fishfish code and not a PKGBUILD one.
There is already an issue (and some-like-a patch) for this problem on github:
https://github.com/fish-shell/fish-shell/issues/32


However, for some reason it only fails on the first compilation, so an easy workaround is to change "make" to "make || make" on PKGBUILD.

SanskritFritz commented on 2012-06-10 20:31

That is strange, I don't even have such a file.

spirtbrat commented on 2012-06-10 18:46

I can confirm the install error. Also 64-bit system here.

SanskritFritz commented on 2012-06-10 08:26

Sorry guys, cannot reproduce the error (64 bit archlinux here).
--without-xsel is added because xsel is provided by the xsel package and it should not be bundled with fish.

silenc3r commented on 2012-06-09 21:54

I'm getting the same eorror as delluminatus and RedLotus.
Is there any particular reason that you add --without-xsel flag?

Anonymous comment on 2012-06-09 15:22

You should add autoconf to make dependencies.

Anonymous comment on 2012-06-09 04:16

Same problem as delluminatus
I'll try delluminatus's solution.

SanskritFritz commented on 2012-06-08 23:28

delluminatus: I don't get this history.1 error, sorry. On the other hand I modified the PKGBUILD in such a way that it can be compiled more times.

Anonymous comment on 2012-06-08 17:53

I get the following error on installation:

/bin/sh: line 4: help_doc/man/man1/history.1: No such file or directory

It can be fixed by manually creating the file "./src/fishfish/help_doc/man/man1/history.1". However, if you try to run makepkg again after manually changing the src tree, the build process substitutes "python2" for "python" a second time resulting in files referencing "python22."

Basically I solved this problem by:

echo "history.1 contents" > src/fishfish/help_doc/man/man1/history.1

then perform a re-substitution:

find -wholename "./src/*.py" | xargs sed "s/python2/python/" -i

Alternatively, you could alter the PKGBUILD by changing the three substitutions in the build() function to the following:

sed -i "s/python[^2]/python2/" "internalize_scripts.py"
sed -i "s/python[^2]/python2/" "make_completions.py"
sed -i "s/python[^2]/python2/" "create_wajig_completions.py"

SanskritFritz commented on 2012-06-08 12:06

Added python2 as makedepend, thanks miffe.

SanskritFritz commented on 2012-06-06 12:06

Ahah, thanks, will be fixed soon.

miffe commented on 2012-06-06 11:58

I get this when building:
/usr/bin/env: python2: No such file or directory
make: *** [builtin_scripts.h] Error 127

Adding python2 to makedepends solves it.

SanskritFritz commented on 2011-09-13 12:26

157ca12 Merge commit 'refs/merge-requests/12' of git://gitorious.org/fish-shell/fish-shell into merge-requests/12
e9d1a54 fix: Fish may hang in Linux virtual console
c8981c0 Merge commit 'refs/merge-requests/10' of git://gitorious.org/fish-shell/fish-shell into merge-requests/10
1ae44c6 still didn't work correctly, there were delimiters missing in the conditional as well
b288747 last commits fix now also applied in the 'else' clause
8e2db29 fixed an issue when trying to complete something like 'cd /mnt/windows7/Program\ Files\ \('
d38de73 Change git-symbolic-ref to git symbolic-ref.
cde6067 minor changes on user doc

SanskritFritz commented on 2011-01-11 22:04

2583638: Merge commit 'refs/merge-requests/8' of git://gitorious.org/fish-shell/f...
11360b0: mimedb: Search all the defaults.list files rather than just the first.
bdfc621: fish_indent: indent "case" with the same level as "swith&...
74f06d5: fish_indent: don't indent commands after a pipe
46986ad: Use eix instead of emerge when possible
6a5d647: mimedb: Search for exact mime type match in search_ini.
cf0850e: mimedb: Add seperator between directory and filename in file_exists.
fd4ac87: funcsave: Fixed error and help handling.
164144f: default_key_bindings: Added mappings for ctrl-arrow keys.
f2ad4a2: Makefile: test for xsel directory before trying to clean it.
df23913: fish_indent: only output fd number when have to
4ccf26f: builtin: count should not accept options, not even help
081e277: builtin: read's --shell long option should require an argument

esodax commented on 2010-12-23 21:11

Haven't yet stumbled upon any unexpected login-issues using fish. The biggest
problem is that some commands expect you to be able to source standard shell
scripts or output. e.g from 'dircolors', 'gpg-agent' and others... luckily
most of them is just to "export <somevar>=<someval>" and can quite easy be
converted to fish syntax.

You can get to a jump start by browsing through my fish config files
repo here: https://github.com/esodax/fishystuff

Not much in there really, but feel free to extract what you want, or just to
get an overall idea on how to solve 'source this-and-that' problems you
might encounter yourself.
Also, my way might not be the best way, but it works for me at least :)

For example, "include/colors.fish" and "functions/run_gpg-agent.fish" resolves
the problem I mentioned above with dircolors and the gpg-agent.

And, again an extra thanks to SanskritFritz.

Merry Christmas fish lovers!

(sorry, couldn't access the ml from here)

Anonymous comment on 2010-12-23 10:11

@esodax: I'm trying to pick fish again while using (very little) of zsh. Can you write something more about issues when using fish as login shell? (you can reply to fish ml)


Sincerely,
Gour

SanskritFritz commented on 2010-11-29 15:02

It is good to know that fish development picked up some spin again, see here: https://wiki.archlinux.org/index.php/Fish#Current_state_of_fish_development

esodax commented on 2010-11-29 14:19

I've actually switched from using zsh for years as my main shell to fish, I'm even using it as my login shell now.
I never managed to fully get a grip of all the stuff in zsh. However, fish I was able to mostly understand in just one day, and still it gives me the same power as zsh did.
Well, this is just about to let others know how great fish really is. :)

SanskritFritz commented on 2010-11-28 20:37

Good idea, package updated. Fish FTW!

esodax commented on 2010-11-28 00:00

Maybe python2 should be added to a optdepends array, because of the make_completions script?
Thanks for this package.