I'm trying to setup a makefile to create a virtual environment using Make in Windows.
I want to create an empty requirement.txt file from within the Makefile. This file shall be filled later with pip-package names.
The Code to create the empty file in windows looks like this:
build/requirements.txt:
echo >> //C:\Users\xxx\environments\venv\homepage\requirement.txt
When I run the Makefile no file is created. But when I use the command in Powershell directly, it creates the requirements.txt file.
I suppose it has something to do with Escaping the command. but I couldnt find any hint how to do this.
Has anyone an Idea, why the file is not created from the Makefile?
Merci A
I found out, that the syntax must be like that:
build/requirements.txt:
type nul >> environments/venv/homepage/requirements.txt
The reason is, that Make writes to the $USER folder as standard. So that is the starting point for any PATH used here.
Related
I have a the following command in a file called $stat_val_result_command.
I want to add -Xms1g parameter at the end of the file so that is should look like this:
<my command in the file> -Xms1g
However, I want to run this command after append. I am running this in a workflow system called "nextflow". I tied many things, including following, but it does not working. check the script section which runs in Bash by default:
process statisticalValidation {
input:
file stat_val_result_command from validation_results_command.flatten()
output:
file "*_${params.ticket}_statistical_validation.txt" into validation_results
script:
"""
echo " -Xms1g" >> $stat_val_result_command && ```cat $stat_val_result_command```
"""
}
Best to avoid appending to or manipulating input files localized in the workdir as these can be, and are by default, symbolic links to the original files.
In your case, consider instead exporting the JAVA_TOOL_OPTIONS environment variable. This might or might not work for you, but might give you some ideas if you have control over how the scripts are being generated:
export JAVA_TOOL_OPTIONS="-Xms1g"
bash "${stat_val_result_command}"
Also, it's generally better to avoid localizing and running scripts like this. It might be unavoidable, but usually there are better options. For example, third-party scripts, like your Bash script could be handled more simply:
Grant the execute permission to these files and copy them into a
folder named bin/ in the root directory of your project repository.
Nextflow will automatically add this folder to the PATH environment
variable, and the scripts will automatically be accessible in your
pipeline without the need to specify an absolute path to invoke them.
This of course assumes you can control and parameterize the process that creates your Bash scripts.
I have a /.cust_dev_cmds/ directory on my MBP machine that is part of a parent sysadmin-work directory that uses a git repo here. I would like to be able to:
Not have to use a for loop in my .bash_profile to source all the *.sh files.
Add the directory to PATH with and export line in the .bash_profile instead.
# from my .bash_profile
export PATH="/Users/<my-usr-name>/Public/sharable-dev-scripts:$PATH"
This does show up with a echo $PATH but when I try to invoke a function from within the scripts I have created that worked with sourcing directly within the .bash_profile in a loop (like with point #1 above) like this
# create a directory with a builtin command
mkdir test-dir
# use one of my custom ones to create a simple readme.md within that directory:
mkr !!:1
# I am getting >>> mkr: command not found
Use whatever type of link to not have a duplicated directory structure on the machine.
Any good explanations to read up on here without using $10 words would be great.
Define a means to test the link works and works through PATH. It would also be nice that something like declare -F would be available to see that the scripts within the directory are in fact becoming part of callable functions in the shell.
is there a command anyone knows to do this?
Step this up a notch for a shared network directory. I have created a shared directory through apple > System Preferences > Sharing, and turned on the ability to share this directory in the Public folder.
Is there a tutorial that can outline this with something like VirtualBox and an Ubuntu guest that is accessing the commands from the MBP shared directory?
I have realized point #1, so really the question begins with #2 so no one would suggest the first one. I have read a bit on links but the way most of the articles I come across describing them are difficult to wrap my head around- especially when wishing to add this functionality to PATH. I believe the answer may revolve around how links are followed, but it may be better to back up and punt- dig back into linking first- then export my directory appropriately without a link, and eventually get the proper resolution to this situation.
The last thought on links before I try a few hacks on my own is do I need to only add a link to the Public directory and somehow place a flag to look at all the directories within the /Public, or is it better to drill all the way down to the /Public/shared-directory/.cust_dev_cmds? Any direction would be greatly appreciated. My goal is to be able to have a few custom command directories for various tasks, and eventually have them across networks/instances.
When you want all functions that you wrote in files in /.cust_dev_cmds/, the normal way would be sourcing all the files.
When you want to avoid a loop, you can use
utildir="$HOME/.cust_dev_cmds/" # I think the path is relative to your home).
source <(cat ${utildir}/*)
When you want the functions found with the PATH, you should make a file for each function.
Before:
# cat talk
ask() { echo "How are you?"; }
answer() { echo "Fine, thank you"; }
After:
# cat ask
echo "How are you?"
# cat answer
echo "Fine, thank you"
When you want all users to use the same set of functions, consider a master script that sources all scripts (the masterfile can use user=dependent settings like HOME or VERSION):
# cat /Public/shared-directory/setup_functions
utildir="$HOME/.cust_dev_cmds/" # I think the path is relative to your home).
source <(cat ${utildir}/*)
source some_other_file
Now each user only needs to source one file.
I have several text docs that are created each day from templates. This process I've achieved successfully albeit probably in a Cro-Magnon way. I want these newly created text files to be filed within a newly created dated folder.
The script creates the file docs from the templates successfully and also creates the newly dated directory. I don't really want to create these text files somewhere else and then move them to the newly created directory. Rather that they be created directly within it. All my research tends to involve directories that already exist rather than one created from a var.
I've included just one file creation example below.
Hope you can help. TIA
today=`date '+%y%m%d'`;
today_Folder=~/Desktop/test/"${today}"
if [[ ! -d $today_Folder ]]
then
mkdir "${today_Folder} `(date '+%A')`"
fi
cat ~/Desktop/test/template.txt >> ~/Desktop/test/dest.txt
P.S. I've tried to make the cat command regarding the text files clearer - it simply creates files. I'm NOT trying to create a tree of directories. Simply ONE newly created directory that could be in test along with the text files.
Your question is how to dynamically create a file, also creating all the path to contain that file? That's not possible in any intuitive/portable way, and it's not typically programs always have to create the directory before the file. What you can do is pass the -p flag to mkdir. On Linux systems (this may also not be portable), this flag means "create all the directories necessary for this path". Zero directories is okay, so you don't need to check whether the directory already exists. So change the whole if block to just this:
mkdir -p "${today_Folder} `(date '+%A')`"
Also, it's kind of smelly the way you want a string (the path) and you're using three operations to create it. Could it be simpler? You want more statements when they add clarity, but in this case the steps are so simple that the only thing accomplished is to make your colleagues go up and read what you wrote more than once. It might suit to change it to:
dir_path=...
mkdir -p "${dir_path}"
To accomplish this, keep in mind that instead of backticks, you can add command substitution with $(). It helps since backticks can't be nested--it makes the line more readable, since you clearly see the command's start/end.
I have two environment variables defined as:
test1=C:\something\dir1
test2=C:\something\dir2
And I'm trying to run the following command:
copy dir1\filename.txt dir2\filename.txt
I know that if I write the copy command with the environment variables it will work, like below:
copy %test1%\filename.txt %test2%\filename.txt
But isn't there a better way to do this? If Windows doesn't find the "dir1" directory in its current directory, won't it try to find it with the system variables it has?
EDIT: Im trying to use the copy command without typing the enviroment variable's name in the command.
Something like "copy dir1\filename.txt dir2\filename.txt", where, if Windows cant find the dir1 directory in its current directory, it would automatically search this directory with the enviroment variables. Is this possible?
This will copy the fully qualified path and filename, and cater for spaces etc.
copy "%test1%\filename.txt" "%test2%\"
If it doesn't work for you then edit your question and give more details about the task.
I open up a .txt file. Make some modifications to it. Then I'd like to outsheet it with the same name. To do this, I presume either I'd have to delete the file right before I outsheet it or replace the file in the outsheeting process. As a backup, I can manually delete the .txt execute the relevant code.
I'd like to do this via Stata, however. Is it possible to either:
1. Delete a .txt file within a Stata .doe--say with a shell command ?
2. Overwrite a .txt file in the outsheeting process ?
I am outsheeting as in:
outsheet v1 v2 using "file.txt", nolabel delim(",")
I run Stata in Windows on an XP machine.
Add the replace option anywhere after the comma.
You can can also shell out to the OS. Type "h shell" to see how.
Or just use rm/erase directly.