I am struggling with running go in the fish shell.
A) I followed the installation procedure and installed go as specified in the documentation
However, when I run go version, I am getting a fatal error:
fatal: not a git repository (or any parent up to mount point /)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
When I initialize the git by git init, I receive another error:
error: pathspec 'version' did not match any file(s) known to git
I have the paths configured in my fish config as follows:
set -gx GOPATH /home/<user>/go
set -gx PATH $PATH /usr/local/bin/go $GOPATH/bin
where /usr/local/bin/go is an output of which go in both bash and fish
Running it like bash -c 'go version', however, works without any problems (even without the git directory).
B) Also tried gofish and dnf to install go, the same results
Any help is greatly appreciated!
Additional info:
Fish version: fish, version 3.0.2
Operating system: Linux t460s 5.0.13-300.fc30.x86_64 x86_64 GNU/Linux
Could this be some kind of alias that is used instead? The error message is an error from the git binary.
Take a look in your file ~/.config/fish/config.fish and see if you find an alias there. If you do not find anything there look at the files in ~/.config/fish/ and subfolders.
Related
I'm transitioning from berkshelf to the Policyfile workflow, and everything was running fine until I attempted to resolve all the dependencies. Two of my dependencies are from git, so when I run chef install in the cookbook directory, I get the following error:
Error: Failed to generate Policyfile.lock
Reason: (CookbookOmnifetch::GitCommandError) Git error: command `git reset --hard 7e5e4bccf57718cacc246f1b4d501ff7de8ac13d` failed. If this error persists,
try removing the cache directory at 'C:/Users/pcummuskey/AppData/Local/chefdk/cache/.cache/git/780b7dbee933c3a9efbfd3898d54cdf72784fa7e'.Output from the command:
error: unable to create file test/kitchen/.kitchen/tmp/librarian/cache/source/chef/site/877777683730772c36b1e6a3fc3aa2c3/windows/version-uri/a3432e93fb3c21cb7111fde71e435661/package/files/default/handlers/windows_reboot_handler.rb: Filename too long
Checking out files: 100% (389/389), done.
fatal: Could not reset index file to revision '7e5e4bccf57718cacc246f1b4d501ff7de8ac13d'.
Predictably, removing the cache folder had zero effect on the problem. I'm in the middle of checking whether I can get it working in the ubuntu subsystem and/or docker atm, but it'd be nice to be able to run it natively on Windows.
I don't get this problem using "git bash" (MINGW64) packaged in Git for windows installer.
As an example:
$ pwd
/c/Users/foobar/AppData/Local/chefdk/cache/.cache/git/13036489648c6f8811262c6a50ff6b03da2395b7/4416ca3886d025b3ead519f1ee53af03ecb630f64416ca3886d025b3ead519f1ee53af03ecb630f64416ca3886d025b3ead519f1ee53af03ecb630f64416ca3886d025b3ead519f1ee53af03ecb630f6/4416ca3886d025b3ead519f1ee53af03ecb630f64416ca3886d025b3ead519f1ee53af03ecb630f64416ca3886d025b3ead519f1ee53af03ecb630f61
Managed to work around the issue by making a local cookbooks folder and checking out the dependencies to there and referencing them by path. It's not pretty, but it does the trick.
If I had a compiled Golang program that I wanted to install such that I could run it with a bash command from anywhere on my computer, how would I do that? For example, in nodejs
npm install -g express
Installs express such that I can run the command
express myapp
and express will generate a file directory for a node application called "myapp" in whatever my current directory is. Is there an equivalent command for go? I believe now with the "go install" command you have to be in the directory that contains the executable in order to run it
Thanks in advance!
Update: If you're using Go 1.16, this answer still works, but go install has changed and is now the recommended method for installing executable packages. See Karim's answer for an explanation: https://stackoverflow.com/a/68559728/10490740
Using Go >= 1.11, if your current directory is within a module-based project, or you've set GO111MODULE=on in your environment, go get will not install packages "globally". It will add them to your project's go.mod file instead.
As of Go 1.11.1, setting GO111MODULE=off works to circumvent this behavior:
GO111MODULE=off go get github.com/usr/repo
Basically, by disabling the module feature for this single command, it will install to GOPATH as expected.
Projects not using modules can still go get normally to install binaries to $GOPATH/bin.
There's a lengthy conversation and multiple issues logged about this change in behavior branching from here: golang/go - cmd/go: go get should not add a dependency to go.mod #27643.
Starting with Go >= 1.16 the recommended way to install an executable is to use
go install package#version
For example, go install github.com/fatih/gomodifytags#latest.
Executables (main packages) are installed to the directory named by the GOBIN environment variable, which defaults to $GOPATH/bin or $HOME/go/bin if the GOPATH environment variable is not set. You need to add this directory to your PATH variable to run executables globally. In my case, I've added this line to my ~/.zshrc file. (if you are using bash, add it to the ~/.bash_profile file):
export PATH="$HOME/go/bin:$PATH"
Go team published a blog post about this change, here's the explanation quote:
We used to recommend go get -u program to install an executable, but this use caused too much confusion with the meaning of go get for adding or changing module version requirements in go.mod.
Refer to go install documentation for more details
As far as I know, there is no direct equivalent to npm install -g. The closest equivalent would not be go install, but go get. From the help page (go help get):
usage: go get [-d] [-f] [-fix] [-insecure] [-t] [-u] [build flags] [packages]
Get downloads and installs the packages named by the import paths,
along with their dependencies.
By default, go get installs binaries to $GOPATH/bin, so the easiest way to make those binaries callable from everywhere is to add that directory to your $PATH.
For this, put the following line into your .bashrc (or .zshrc, depending on which shell you're using):
export PATH="$PATH:$GOPATH/bin"
Alternatively, you could also copy or link the executables to /usr/local/bin:
ln -s $GOPATH/bin/some-binary /usr/local/bin/some-binary
Short solution for Linux users:
Use the go get command as usual
Add the following lines to .bashrc:
# This is the default GOPATH, you should confirm with the 'go env' command
export GOPATH=$HOME/go
export PATH=$PATH:$GOROOT/bin:$GOPATH/bin
Restart terminal or source it. Installed binaries will be available globally.
For Go v1.8+
go install package_name#latest
Caveat: this answer is outdated following the 2020 deprecation of go get. The solution presented here won't work with newer Go runtime installs.
The closest analogue of this in Go would be go get. By default, it will fetch a Go package from a supplied repository URL, and requires a $GOPATH variable to be set in your shell so that Go knows where to store the packages (and subsequently where to find them when compiling code depending on go get-ted packages).
Example syntax:
$ go get github.com/user/repo
The behaviour supplied by npm's -g flag is default, and packages installed using go get are normally available globally.
See go get --help for more information about the command.
As mentioned by #helmbert, adding your $GOPATH to your $PATH is useful if you're installing standalone packages.
if you are using zsh :
first: install your package using :
go install package#version
then , you edit your .zshrc file
nano ~/.zshrc
Add this line to the end of .zshrc file :
export PATH="$HOME/go/bin:$PATH"
last but not least :
source ~/.zshrc
then open a new terminal and execute your command :)
TL;DR at the bottom. I'm going to walk you through how I came to this conclusion, and why the more obvious solutions don't work.
Upon seeing this question, I thought "If I could set root's GOPATH=/usr, it would install things in /usr/bin/ and /usr/src!"
So I tried the obvious thing:
Add GOPATH=/usr to root's .bashrc.
And it worked!
Sort of.
Not really.
Turns out, sudo doesn't execute root's .bashrc. For "security" or something like that.
Do env_set or something in /etc/sudoers
Turns out, /etc/sudoers can only remove environment variables. There's no env_set directive.
(As far as I can find)
Dig through man sudoers.
Where does sudo get it's default set of environment variables from?
Well, the first one in the list is /etc/environment, so that's the one I used.
sudo echo "GOPATH=/usr" >> /etc/environment
sudo go get <repo>
Binaries will be put in /usr/bin, and sources will be put in /usr/src.
Running go as non-root will use GOPATH the "normal" way.
If you don't have go installed, you may use the gobinaries. it builds an on-demand binary of the project from github repo.
The command to install the go package would be:
curl -sf https://gobinaries.com/rakyll/hey | sh
Using Cygwin on Windows 7 64 bit.
Trying to rebase msys-2.0.dll: for Git with the below command.
However I am getting the message saying its skipped because its wrong machine type.
cd /cygdrive/c/Program Files (x86)/Git/bin
$ rebase.exe -b 0x50000000 /cygdrive/c/"Program Files (x86)"/Git/bin/msys-2.0.dll
/cygdrive/c/Program Files (x86)/Git/bin/msys-2.0.dll: skipped because wrong machine type.
What does this mean? Trying to look up on Google but didn't find much information anywhere - so hoping someone in this community can help.
As mentioned here, it could be a Cygwin issue depending on your current version of Cygwin.
But you can also try the alternative without cygwin, using the latest git-for-windows:
unzip PortableGit-2.6.3-64-bit.7z.exe anywhere you want
add to your %PATH% the folders C:\path\to\PortableGit-2.6.3-64-bit\bin and C:\path\to\PortableGit-2.6.3-64-bit\usr\bin (that last path comes with 200+ Linux gnu commands: ls, xargs, awk,... all accessible from your CMD.exe session!)
Then try again your git commands.
I know that this should probably be very easy, and I have looked into the read me files and other files in cgdb like INSTALL that talk about isntalling cgdb. I was trying to install cgdb and the instructions said:
$ ./configure --prefix=/usr/local
$ make
$ sudo make install
However, there is no such file inside of the file I got from running the git clone command. i.e. from running:
$ git clone git://github.com/cgdb/cgdb.git
I went inside that file to see the installation instructions and it says the following:
Basic Installation
==================
These are generic installation instructions.
The `configure' shell script attempts to guess correct values for
various system-dependent variables used during compilation. It uses
those values to create a `Makefile' in each directory of the package.
It may also create one or more `.h' files containing system-dependent
definitions. Finally, it creates a shell script `config.status' that
you can run in the future to recreate the current configuration, a file
`config.cache' that saves the results of its tests to speed up
reconfiguring, and a file `config.log' containing compiler output
(useful mainly for debugging `configure').
If you need to do unusual things to compile the package, please try
to figure out how `configure' could check whether to do them, and mail
diffs or instructions to the address given in the `README' so they can
be considered for the next release. If at some point `config.cache'
contains results you don't want to keep, you may remove or edit it.
The file `configure.in' is used to create `configure' by a program
called `autoconf'. You only need `configure.in' if you want to change
it or regenerate `configure' using a newer version of `autoconf'.
The simplest way to compile this package is:
1. `cd' to the directory containing the package's source code and type
`./configure' to configure the package for your system. If you're
using `csh' on an old version of System V, you might need to type
`sh ./configure' instead to prevent `csh' from trying to execute
`configure' itself.
Running `configure' takes awhile. While running, it prints some
messages telling which features it is checking for.
2. Type `make' to compile the package.
3. Optionally, type `make check' to run any self-tests that come with
the package.
4. Type `make install' to install the programs and any data files and
documentation.
5. You can remove the program binaries and object files from the
source code directory by typing `make clean'. To also remove the
files that `configure' created (so you can compile the package for
a different kind of computer), type `make distclean'. There is
also a `make maintainer-clean' target, but that is intended mainly
for the package's developers. If you use it, you may have to get
all sorts of other programs in order to regenerate files that came
with the distribution.
and there is more I didn't paste.
The part that confuses me is the following paragraph:
The file `configure.in' is used to create `configure' by a program
called `autoconf'. You only need `configure.in' if you want to change
it or regenerate `configure' using a newer version of `autoconf'.
does it mean I should run ./configure.in? I am honeslty a little scared of running it because I need to run it in as sudo plus, that file doesn't even exit in the code I got from git clone. This is what I do have in that directory:
➜ cgdb git:(master) ls
AUTHORS Makefile.am autogen.sh doc roadmap.txt
COPYING NEWS autorelease.sh indent.sh test
ChangeLog README cgdb lib
FAQ README.md config packages
INSTALL TODO configure.init release-todo.txt
➜ cgdb git:(master)
I did grep for config or config.in in that directory and it didn't yield anything useful.
I have also tried brew installing it but I ran into problems. When I cgdb the file and hit run it frozen at a print statment instead of moving on to the next command prompt for gdb as in (gdb). Anyone knows whats up? Is it because I am using go source file?
What have people done to have cgdb working on a mac?
Неу, yeah, the git source for cgdb doesn't come with the configure script pre-built. (The official release distributions do, however.)
If you have the right tools installed, you can fix this by running:
./autogen.sh
This will generate the configure script, and then you can install it as you would any other autoconf-based source package.
Forget about all that aggro with configuration and dependency libraries; install MacPorts and do:
$ sudo port install cgdb
They have 0.6.7 at the time of writing.
What worked for me was doing:
brew install cgdb
However, for me that was not enough to get it going. I needed to sudo it for it to run:
sudo cgdb
I would have never guessed I needed to sudo it...I figured it out because it was throwing me the error talked about in the following question:
gdb fails with "Unable to find Mach task port for process-id" error
I am getting below error when taking checkout from svn using a script.
**[Test] $ /bin/sh -xe /tmp/hudson8576425899836211909.sh
+ sh /cvsrx/rxapp/build_dir/Jenkins_Scripts/test.sh
Could not load program svn:
Could not load module /opt/freeware/lib/libssl.so.
Dependent module /usr/lib/libcrypto.a(libcrypto.so.1.0.1) could not be loaded.
Member libcrypto.so.1.0.1 is not found in archive
Could not load module svn.
Dependent module /opt/freeware/lib/libssl.so could not be loaded.
Could not load module .
Build step 'Execute shell' marked build as failure
Finished: FAILURE**
In test.sh I have written just one line svn co /path to svn branch/
I am in middle of some test so please don't ask why am not using jenkins in build svn plugin.
here,I am able to take checkout on command prompt using svn co /path to svn branch/
But not if I write this command line in script and run in execute shell of jenkins.
Any help please ?
Am using jenkins on AIX 7 platform.
I had softlinks from /usr/bin/svn to /opt/freeware/bin/svn.SVN installed at /opt/freeware/bin/svn ..... By default when i do which svn its showing /usr/bin/svnBut When i deleted those softlinks and exported path,Jenkins didnt recognize SVN at all.And which svn command doesnot show any svn installed . PFB logs of jenkins :
`
/bin/sh -xe /tmp/hudson5607872610124977868.sh
+ export PATH=/opt/freeware/bin/svn/:/opt/freeware/bin/svnversion:/opt/freeware/bin/svn:/opt/freeware/bin/svnversion/:/usr/java5/lib:/opt/freeware/bin/svnversion/bin:/usr/local/bin:/usr/bin:/usr/X11R7/bin:/bin:/usr/games:/opt/gnome/bin:/opt/kde3/bin:/usr/lib/mit/bin:/sbin:/
+ echo /opt/freeware/bin/svn/:/opt/freeware/bin/svnversion:/opt/freeware/bin/svn:/opt/freeware/bin/svnversion/:/usr/java5/lib:/opt/freeware/bin/svnversion/bin:/usr/local/bin:/usr/bin:/usr/X11R7/bin:/bin:/usr/games:/opt/gnome/bin:/opt/kde3/bin:/usr/lib/mit/bin:/sbin:/opt/freeware/bin/svn/:/opt/freeware/bin/svnversion:/opt/freeware/bin/svn:/opt/freeware/bin/svnversion/:/usr/java5/lib:/opt/freeware/bin/svnversion/bin:/usr/local/bin:/usr/bin:/usr/X11R7/bin:/bin:/usr/games:/opt/gnome/bin:/opt/kde3/bin:/usr/lib/mit/bin:/sbin
+ cd /usr/local/apps/Jenkins_new/scripts
+ ./test.sh
Could not load program /opt/freeware/bin/svn:
Could not load module /opt/freeware/lib/libssl.so.
Dependent module /usr/lib/libcrypto.a(libcrypto.so.1.0.1) could not be loaded.
Member libcrypto.so.1.0.1 is not found in archive
Could not load module svn.
Dependent module /opt/freeware/lib/libssl.so could not be loaded.
Could not load module .
./test.sh[3]: svn: not found
Build step 'Execute shell' marked build as failure
Finished: FAILURE`
ReInstalling Jenkins solved my problem !! It was due to corrupt plugin which happened due to server restart.
Thanks Everyone
I have not seen the exact command of svn you have used in test.sh, but would advice you to give full path of SVN in your calling line, eg /usr/bin/svn co .....
And if also possible make an entry in your test.sh for export PATH and export LD_LIBRARY_PATH setting them to values which are paths of the mentioned .so files
I see you're using Hudson/Jenkins. Hudson and Jenkins use SVNKit internally when checking stuff in and out of Subversion. This means that the command line svn client may not be installed on your system, or that it may have other problems.
It is also possible that there are multiple svn clients on your system. For example, you may have one under /usr/bin/svn and one under /usr/local/bin/svn. If Subversion is working from the command line, but not in the script, you may have a different $PATH setup when you're executing from the command line vs. the script from Hudson/Jenkins. You can add to your script (if it's BASH) the line type svn to see where your executing svn from. It may be different from what you are using from the command line. It may also be nice to print out $PATH as part of your script.
It would also be helpful to see the svn command that your script is executing, and tell us what you're trying to do. You can also add to your script the following lines:
PS4="\$LINE: "
set -xv
These lines will turn on shell script debugging, and help you locate where your script is having problems.
This will give you some clues as to what is going wrong in your script.
Reply
Thanks David. Here only 1 svn path exists (usr/bin/svn)but created as a softlink see -->cd /usr/bin/svn lrwxrwxrwx 1 root system 26 Jul 1 14:34 svn -> ../../opt/freeware/bin/svn. Also soflink of libs.so is created see --> lrwxrwxrwx 1 root system 15 Sep 13 18:15 libssl.so -> libssl.so.1.0.1 .......... Is it possible that softlink is creating these problems ?? Before requesting to remove these softlinks ,i need confirmation that these softlinks are creating problems.
Softlinking isn't unusual for Unix. For example, I have Ant, Grails, Maven, Subversion, and many other packages installed under /opt on my Mac. In order not to have to include each and every one of these in my path, I soft link all of the binaries for those programs under /usr/local/bin. About 80% of the programs under /usr/local/bin are merely soft links elsewhere.
Library soft links are also very common. This usually has to do with version numbering. When a program requests a library, it may or may not include the library's version number. So, you have libfoo-2.0.3.so on your drive. This is the actual version of foo. However, few programs will request that particular version. Instead, they may simply request they need Version #2 of foo or just say they need to link to foo.
To handle this, you will have libfoo2.0.3.so soft linked to libfoo-2.so for programs that specify they need version 2 of foo. Then, libfoo-2.so will be soft linked to libfoo.so. This way, libfoo will be called no matter what. If I install, libfoo2.0.4.so, I can change the link to libfoo-2.so to point to version 2.0.4 instead of version 2.0.3, and anything that depends upon Foo will be picking up the correct version.
Instead, let's look at the error message:
Could not load module /opt/freeware/lib/libssl.so.
Dependent module /usr/lib/libcrypto.a(libcrypto.so.1.0.1) could not be loaded.
Member libcrypto.so.1.0.1 is not found in archive
For some reason, it couldn't access the file /usr/lib/libcryto.a. Is this file on you machine? Is it in /usr/lib? If not, where is it located?
So, where did you get this version of Subversion from? Why is the link to the /opt/freeware/bin/ directory? Was this part of your system?
It could be that Subversion on your system is not complete and never did work. In Jenkins, the Subversion repository is accessed by the SVNKit Jarfile which is embedded inside Jenkins/Hudson itself, so it wouldn't really be a surprise to find that the Subversion binary didn't work.
Are you able to do anything with Subversion from the command line? If not, you may have to install a new version of Subversion from Perzl which is where CollabNet points to for an AIX version of Subversion. (It's at least up to date at version 1.8.4).
You may even want to change the soft link at /usr/bin/svn to point to the newer, working version of Subversion.