On my Mac, tar defaults to BSD tar.
$ tar --version
bsdtar 2.8.3 - libarchive 2.8.3
There is a GNU tar that I can invoke with gnutar.
$ gnutar --version
tar (GNU tar) 1.17
Copyright (C) 2007 Free Software Foundation, Inc.
...
With Ant Tar task, how do I configure it to use the GNU tar?
As #Raghuram wrote, modern versions of Ant do not use an external tar program; instead, they have their own pure-Java implementation. However, you can produce behavior similar to GNU tar by using the longfile="gnu" attribute. From the manual:
If the loss of path or file information is not acceptable, and it rarely is, longfile may be set to the value gnu. The tar task will then produce a GNU tar file which can have arbitrary length paths. Note however, that the resulting archive will only be able to be untarred with GNU tar.
ant tar task does not use the system tar command. Thus the question of configuring an alternate tar implementation does not arise.
Related
To decrease library size, used cmake -E tar "zcvf" "lib.tar.gz" out.lib in cmake to automatically create .tar.gz then lib.tar.gz is moved to folder f, ie. f/lib.tar.gz.
However, to use this library in another project, it's needed to extract to specific folder. The command without cmake is clear: tar -xzf f/lib.tar.gz -C target-folder.
The problem is that, in cmake, how to combine cmake -E tar "xzf" f/lib.tar.gz with -C option? Commands such as cmake -E tar "xzf" f/lib.tar.gz -C f/ will result error.
Ideal script would be sth. like
add_custom_target(PrepareLib
COMMAND ${CMAKE_COMMAND} -E tar "xzf" command....
WORKING_DIRECTORY ${CMAKE_SOURCE_DIR} ## note, this line must not be modified
)
Tools version: Visual Studio 2017, Windows 10 and CMake 3.17.1.
I need to install GNU grep on my Mac but I'm finding some difficulties.
I tried doing this:
brew install grep --with-default-names
But this is no longer an option since Homebrew removed --with-default-names.
Can anyone provide a solution for this?
Yes, --with-default-names was removed.
But some formulas, like grep, provided a workaround for this:
$ brew info grep
...
==> Caveats
All commands have been installed with the prefix "g".
If you need to use these commands with their normal names, you
can add a "gnubin" directory to your PATH from your bashrc like:
PATH="/usr/local/opt/grep/libexec/gnubin:$PATH"
...
First, to install, just do install without --with-default-names.
$ brew install grep
...
==> Summary
🍺 /usr/local/Cellar/grep/3.3: 21 files, 880.7KB
You should also see that same "Caveats" info I mentioned at the start. Now, by default, the Homebrew grep would be prefixed by a "g", so it's accessible as ggrep.
$ ggrep -V
ggrep (GNU grep) 3.3
Packaged by Homebrew
Copyright (C) 2018 Free Software Foundation, Inc.
...
This prevents it from shadowing the built-in grep that comes with Mac.
$ grep -V
grep (BSD grep) 2.5.1-FreeBSD
If you really need to use grep and not ggrep, just follow the instructions and put /usr/local/opt/grep/libexec/gnubin at the start of your PATH. You have to do this in your .bashrc or .bash_profile (whichever one you use).
$ echo 'export PATH="/usr/local/opt/grep/libexec/gnubin:$PATH"' >> ~/.bash_profile
$ source ~/.bash_profile
$ grep -V
grep: warning: GREP_OPTIONS is deprecated; please use an alias or script
grep (GNU grep) 3.3
Packaged by Homebrew
...
I installed pigz via homebrew on my Macbook Air (OS X 10.10.5) to get better performance for compress/decompress.
To compress, I use tar --use-compress-program=pigz -cf test.tgz test and it's ok.
But the command to uncompress, tar --use-compress-program=pigz -xf test.tgz output error:
tar: Unrecognized archive format
tar: Error exit delayed from previous errors.
Or sometimes it output:
tar: Unrecognized archive format
pigz: abort: write error on <stdout> (Broken pipe)
tar: Child process exited with status 32
tar: Error exit delayed from previous errors.
I read the manual of tar, and have no clue why it doesn't work.
I noticed that even tar --use-compress-program=gzip -xf test.tgz generate the same error. So is this a bug of OSX's tar implementation?
Note: I know pipe style pigz -d test.tgz | tar -xf works and in this case I could also just use tar -xf test.tgz which call built-in gzip. But I just want to confirm whether it is a bug.
The program works as designed: there is no provision in its command-line to pass along the options needed to use gzip for decompressing. Instead of "gzip" for decompressing, you should use the wrapped gzcat, e.g.,
tar --use-compress-program gzip -cf foo.compressed foo
tar --use-compress-program gzcat -tf foo.compressed
A quick check shows that this does not work:
tar --use-compress-program 'gzip -d' -tf foo.compress
although that could change some time (it is doable, but not done).
According to pigz's manual page, it has unpigz, which is what you can use for that program.
This is an issue between BSD tar and GNU tar.
I fixed this by installing gnu-tar from homebrew and placing that on the path.
See this similar question on superuser: https://superuser.com/questions/318809/linux-os-x-tar-incompatibility-tarballs-created-on-os-x-give-errors-when-unt
on linux I renamed gzip to gzip.sav and made a soft link to pigz so tar, dolphin, yum, and any other program that calls gzip actually calls pigz
On Linux, I have a --parents option available for the cp command so I can do
cp --parents test/withintest/go.rb test2
http://www.gnu.org/software/coreutils/manual/html_node/cp-invocation.html
On Mac, I do not have this option available. Is there a way to do this on Mac? Why is this option not available?
PS. The purpose of --parents is the following:
‘--parents’ Form the name of each destination file by appending to the
target directory a slash and the specified name of the source file.
The last argument given to cp must be the name of an existing
directory.
For example, the command:
cp --parents a/b/c existing_dir
copies the file a/b/c to existing_dir/a/b/c, creating any missing intermediate directories.
This bothered me quite a lot as well.
A workaround for this could be to use rsync.
rsync -R test/withintest/go.rb test2
has the same effect as cp --parents and OS X comes standard with rsync.
You can use the ditto command on Mac OS X:
The basic form
ditto <src-path> <dst-path>
does what you want. There's a lot more options too - check out the man page.
You can install the GNU version of cp using MacPorts.
After MacPorts is installed you can install the coreutils packages:
sudo port install coreutils
Then you will be able to use the GNU version cp and other core utilitites (ls, date, cat, etc.) by prefixing the command with a g:
gcp --parents test/withintest/go.rb test2
If you want these GNU versions to be used by default you can add the GNU bin update your path. Add the following to your ~/.bash_profile:
export PATH="/opt/local/libexec/gnubin:$PATH"
The Homebrew way:
Install coreutils
brew install coreutils
Use the GNU g- prefixed command
gcp --parents test/withintest/go.rb test2
I used rsync and what I did was:
rsync -R dir/**/file.json destination
Try
mkdir -p `dirname "$file_path"` && cp "$old_dir/$file_path" "$file_path"
This first creates the directory with all itermediates in the relative file path. Then it copies the file to the newly created directory.
I would not replace mac cp with GNU cp. I would also not used ditto because it is not cross-platform. Instead use cross-platform tools, such as rsync:
rsync <srcDir/srcFile> <dst>
Result: dst/srcDir/srcFile
I'm writing some small bash scripts for copiyng certain files/directories in GNU/Linux and Solaris. Everything is OK in Linux, but cp command hasn't the same options in Linux and Solaris.
Copy command is something like this:
cp -ruv $source $dest
Unfortunately I don't know how to achieve copy verbose and copy update in Solaris. Any idea?
Thanks
Unfortunately, cp under Solaris doesn't have that option. man solaris should reveal that.
Are you comfortable making your script depend on rsync?
Or, if possible, you can install the coreutils package and use GNU's cp.
I ran into a similar issue myself and found that gcp takes care of it too. I've made installing coreutils part of my standard system setup.
I run these on a new Solaris install:
pkgadd -d http://get.opencsw.org/now
pkgutil -U
pkgutil -i -y coreutils
pkgutil -a vim
pkgutil -i -y vim
pkgutil -i -y findutils
Remember to add the path - and the documentation path - to your profile, and possibly to the system profile at /etc/profile:
# Set the program path
PATH=$PATH:/usr/sfw/bin:/usr/sfw/sbin:/usr/openwin/bin:/opt/csw/bin:/usr/ccs/bin:/usr/local/bin:/usr/local
export PATH
# Set the documentation path
MANPATH="$MANPATH:/usr/share/man:/opt/sfw/man:/opt/csw/man"
export MANPATH
It sounds like you might be new to Solaris - as I am relatively new. I also do these, which shouldn't affect anything.
I set VIM as the default editor instead of VI - it's compatible, but has more features, including ANSI color, and some terminal emulators will pass your mouse clicks and scrolling through for even more flexibility:
# Set the default editor
EDITOR=vim
export EDITOR
Then if you are still using the default prompt that doesn't say anything, you might want to add some information - this version requires a Bash shell:
# Set the command prompt, which includes the username, host name, and the current path.
PS1='\u#\h:\w>'
export PS1
To recreate verbose mode, you can tee the output to the controlling terminal (/dev/tty) while the stdoout output of tee itself is passed to cp via xargs.
find /some/source/directory -type f | \
tee /dev/tty | xargs -I {} cp {} /copy/to/this-directory/
Replace the find with whatever you like, so long as it passes the paths to the files to be copied through the pipe to tee.
Tested on a standard Solaris 10 system without extra GNU utils.