The Platform.sh CLI is the official command-line interface for Platform.sh

... to easily look up mounts and rsync files up or down. I've never worked with any of the used components before so I'm pretty sure some stuff (like verbose error reporting, etc.) can be improved but I think it's a pretty solid base for you guys to work with. push and pull duplicate quite some code too and I think you might want to refactor part of this into some service or similar.

With the latest release, v3.39.0, our team's drush aliases can't find the document root after regenerating. We have to manually add the /app directory before our document root to our alias files for the aliases to function properly. Our sites are currently running Drupal 8.6.7 and Drush 9.4.

Hello I am trying to set up my local development environment and i am following this link

But whenever i download a project usning platform get [project-id] [folder-name] command. I got following sh: 1: drush: not found error, while i have downloaded drush by composer global require 'drush/drush:6.*' command.

linux@ubuntu:~$ platform get uoml5m5oa72ya my-project Created project directory: my-project Enter a number to choose which environment to check out: [0] Master [1] sprint1 [2] Test Arif

Enter passphrase for key '/home/linux/.ssh/id_rsa': Cloning into 'my-project/repository'... Enter passphrase for key '/home/linux/.ssh/id_rsa': The project Genero was successfully downloaded to: my-project Building application php using the toolstack php:drupal sh: 1: drush: not found Changed current directory to /home/linux/.composer ./composer.json has been updated Loading composer repositories with package information Updating dependencies (including require-dev) Nothing to install or update Generating autoload files sh: 1: drush: not found The build failed with an error Drush must be installed linux@ubuntu:~$

Can anyone please tell me how can i solve this issue?

Thanks.

When I try the command platform environment:activate --project=project_id --parent=develop branch -vvv, I get this issue:

Waiting for the activity xxx (Nek activated environment feature/assets-packages):

[-->-------------------------] 3 secs (in progress)
Warning: stream_select(): Cannot cast a filtered stream on this system in phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php on line 203

Fatal error: Uncaught ValueError: No stream arrays were passed in phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php:203
Stack trace:
#0 phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php(203): stream_select(Array, Array, Array, 0, 200000)
#1 phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php(124): Platformsh\Cli\Service\ActivityMonitor->canRead(Resource id #810, 200000)
#2 phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php(252): Platformsh\Cli\Service\ActivityMonitor->waitAndLog(Object(Platformsh\Client\Model\Activity))
#3 phar:///Users/nek/.platformsh/bin/platform/src/Command/Environment/EnvironmentActivateCommand.php(118): Platformsh\Cli\Service\ActivityMonitor->waitMultiple(Array, Object(Platformsh\Client\Model\Project))
#4 phar:///Users/nek/.platformsh/bin/platform/src/Command/Environment/EnvironmentActivateCommand.php(43): Platformsh\Cli\Command\Environment\EnvironmentActivateCommand->activateMultiple(Array, Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\StreamOutput))
#5 phar:///Users/nek/.platformsh/bin/platform/vendor/symfony/console/Command/Command.php(255): Platformsh\Cli\Command\Environment\EnvironmentActivateCommand->execute(Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\ConsoleOutput))
#6 phar:///Users/nek/.platformsh/bin/platform/vendor/symfony/console/Application.php(1010): Symfony\Component\Console\Command\Command->run(Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\ConsoleOutput))
#7 phar:///Users/nek/.platformsh/bin/platform/src/Application.php(297): Symfony\Component\Console\Application->doRunCommand(Object(Platformsh\Cli\Command\Environment\EnvironmentActivateCommand), Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\ConsoleOutput))
#8 phar:///Users/nek/.platformsh/bin/platform/vendor/symfony/console/Application.php(255): Platformsh\Cli\Application->doRunCommand(Object(Platformsh\Cli\Command\Environment\EnvironmentActivateCommand), Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\ConsoleOutput))
#9 phar:///Users/nek/.platformsh/bin/platform/vendor/symfony/console/Application.php(148): Symfony\Component\Console\Application->doRun(Object(Symfony\Component\Console\Input\ArgvInput), Object(Symfony\Component\Console\Output\ConsoleOutput))
#10 phar:///Users/nek/.platformsh/bin/platform/bin/platform(32): Symfony\Component\Console\Application->run()
#11 /Users/nek/.platformsh/bin/platform(10): require('phar:///Users/m...')
#12 {main}
  thrown in phar:///Users/nek/.platformsh/bin/platform/src/Service/ActivityMonitor.php on line 203

Another issue not-really-related to the cli is that when you try to create an environment while you cannot because you're out of the authorized number of environments deployed, it will create the environment anyway but it's broken. (it would be nice if nothing happen)

I installed platformsh-cli whitout problems, but when I try connect to Plataform.sh I get this error:

[~]$ platform
Welcome to Platform.sh!

Please log in using your Platform.sh account

Your email address: [email protected]
Your password:

[GuzzleHttp\Exception\ConnectException]                                      
  Error creating resource: [message] fopen(https://marketplace.commerceguys.c  
  om/oauth2/token): failed to open stream: Connection refused                  
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 405                                                                   
  [message] Undefined variable: http_response_header                           
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 406

[GuzzleHttp\Exception\ConnectException]                                      
  Error creating resource: [message] fopen(https://marketplace.commerceguys.c  
  om/oauth2/token): failed to open stream: Connection refused                  
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 405                                                                   
  [message] Undefined variable: http_response_header                           
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 406

[GuzzleHttp\Ring\Exception\RingException]                                    
  Error creating resource: [message] fopen(https://marketplace.commerceguys.c  
  om/oauth2/token): failed to open stream: Connection refused                  
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 405                                                                   
  [message] Undefined variable: http_response_header                           
  [file] /home/rpayanm/.composer/vendor/guzzlehttp/ringphp/src/Client/StreamH  
  andler.php                                                                   
  [line] 406 

I am behind a proxy and I config the terminal for work with my proxy:

[~]$ echo $http_proxy
http://192.168.1.1:3128
[~]$ echo $https_proxy
http://192.168.1.1:3128
[~]$ echo $ftp_proxy
http://192.168.1.1:3128

How I can fix that?

This replaces fix-aliases with a drush-aliases command, usage:

platform project:drush-aliases [-r|--recreate] [-g|--group[="..."]] [-p|--pipe]

This allows you to specify a new group for all the project's Drush aliases, e.g. @mysite instead of @s0mel0ngPl4tf0rmpr0ject1d.

Or, if you want to keep the obscure group name, you can use the pipe feature to run commands like drush @platform drush-aliases -p._local status inside any project.

This PR also sets a 'uri' key on Drush aliases, useful for commands like drush @mysite.master uli.

Unfortunately the Platform CLI doesn't distinguish between Drupal and non-Drupal applications - so it creates Drush aliases for any project - which is a problem for another day.

I'm seeing a SEGV on db:dump (Debian Buster, on ddev). It's worked fine for a few months with various versions of ddev and various versions of platformsh-cli, but now I just see the Segmentation Fault no matter what version of platformsh-cli I use. I've tried 3.66.3 (which is where I started getting it) but I manually installed 3.65.4 and saw the same thing with that. 3.66.1 has the same behavior, same SEGV, but it has definitely worked and passed tests before. Using same PHP version as previously, 7.4.18, same build.

platform user:list also does a SEGV

platform db:dump -vvv --yes --gzip --file=/var/www/html/.ddev/.downloads/db.sql.gz --project=lago3j23xu2w6 --environment=master
DEBUG Selected project: lago3j23xu2w6
Segmentation fault

Is this a result of something on the server end perhaps? The project it's acting on is running fine.

Suggestions how to debug a SEGV here?

Edit: List of some commands that SEGV:

• user:list
• db:dump
• route:list
• environment:list
• commit:list
• domain:list

By now, when the "drush make" command is called, the --contrib-destination option is harcoded to the current directory. That might be good for production reasons but during the dev it would be better to be able to build all the contributed projects in the sites/all folder. It would avoid polluting the repository with things that don't need to be there.

I don't know what's the best way to add this to the command line but I think it would only take a few minutes, so I hope you won't wait too long to code it. Alternatively, you can just advice me on how you would see this option and I'll make it.

This is a duplicate of https://github.com/platformsh/platformsh-cli/issues/1079 but with new insights.

TL;DR: please enable non-ECDHE ciphers on api.platform.sh

While helping ppl migrate from SymfonyCloud to Platform.sh, I saw MANY ppl having this issue, "certificate has expired".

This happens on both Windows and Mac, and the reason is an outdated openssl lib. Ppl using SymfonyCloud don't have this issue because the symfony binary embeds it's own SSL routines. But ppl that use the platform binary use the system's openssl, and :boom:

While we might dream of telling ppl to upgrade, it's out of reach from a DX perspective. The current situation gives a terrible experience to those ppl.

I dig a bit more to understand the root cause and here it is:

First, I reproduced the issue on a wsl on Windows 10. Then, I realized that composer doesn't have the issue, which means that php is able to connect to both packagist.org and github.com.

Here is why: api.platform.sh supports only ECDHE ciphers, while packagist and github both also allow non-ECDHE ciphers. See below for details.

The solution is then simple: the TLS endpoint serving api.platform.sh should be reconfigured to allow the same ciphers as github.

A workaround in the platform client could be to use openssl_get_cipher_methods() to disable SSL verification if php doesn't support ECDHE ciphers. But it'd be a shame to go this way.

$> nmap --script ssl-enum-ciphers -p 443 api.platform.sh
[...]
]| ssl-enum-ciphers: 
|   TLSv1.2: 
|     ciphers: 
|       TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA (ecdh_x25519) - A
|     compressors: 
|       NULL
|     cipher preference: server
|_  least strength: A

$> nmap --script ssl-enum-ciphers -p 443 api.github.com
[...]
]| ssl-enum-ciphers: 
|   TLSv1.2: 
|     ciphers: 
|       TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_ECDSA_WITH_CHACHA20_POLY1305_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_CHACHA20_POLY1305_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_ECDSA_WITH_AES_128_CBC_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256 (ecdh_x25519) - A
|       TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA384 (ecdh_x25519) - A
|       TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA (ecdh_x25519) - A
|       TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA (ecdh_x25519) - A
|       TLS_RSA_WITH_AES_128_GCM_SHA256 (rsa 2048) - A
|       TLS_RSA_WITH_AES_256_GCM_SHA384 (rsa 2048) - A
|       TLS_RSA_WITH_AES_128_CBC_SHA256 (rsa 2048) - A
|       TLS_RSA_WITH_AES_256_CBC_SHA256 (rsa 2048) - A
|       TLS_RSA_WITH_AES_128_CBC_SHA (rsa 2048) - A
|       TLS_RSA_WITH_AES_256_CBC_SHA (rsa 2048) - A
|     compressors: 
|       NULL
|     cipher preference: server
|_  least strength: A

Hey guys, noticed this Exception popping up nowadays. I used to run platform ssh from random terminals most of the times and it would bring up a list from where i could choose which project and branch to connect to. I've used project:set-remote to map my projects to the CLI tool, but I still enjoyed opening up a quick random terminal, running platform ssh and off i went. I get the reasoning for this, but wouldn't it be easier to add a short info message to use platform:set-remote if it's supposed to be a project root, instead of an Exception? That way everyone can use the commands however and wherever they wish.

Thanks for this awesome tool! Kind regards, Alex.

The CLI installer fails when attempting to install behind a corporate firewall. The installer does not recognize HTTP_PROXY and HTTPS_PROXY env vars.

The issue seems to be with file_get_contents() -- which needs some special handling WRT proxies (See https://stackoverflow.com/questions/31634604/file-get-contents-with-https-requests-via-proxy )

The manual install method detailed at https://github.com/platformsh/platformsh-cli#installing-manually does work (but this is not in the docs for installing the CLI)

Can the installer be adjusted to recognize and use a proxy?

Seems a silly question but I am missing something.

When I run: platform auth:api-token-login --no-interaction I get:

Non-interactive use of this command is not supported.

To authenticate non-interactively, configure an API token using the PLATFORMSH_CLI_TOKEN environment variable.

Which I expect.

However if I set my environment variable first such as:

export PLATFORMSH_CLI_TOKEN="xxxx"

And then run platform auth:api-token-login --no-interaction I get:

An API token is already set via config

So when I run git fetch platform master for example I am greeted with:

Hello user (UUID: xxxxx), you successfully authenticated, but could not connect to service xxx (reason: access requires MFA)
[email protected]: Permission denied (publickey).
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

Worth adding, I have all my SSH keys correctly added to PlatformSH ssh-keys section.

Am I missing something here or is there a genuine bug.

Currently the only method of deleting an existing backup is to use the API.

https://api.platform.sh/docs/#tag/Environment-Backups/operation/delete-projects-environments-backups

Can a new backups command for backup:delete <backup_id> be implemented for self service management of backups?

Add quotes around the application executable before passing it to the autocomplete hook script. Quoting prevents issues with the path when containing special characters (WS and/or \).

On Windows 10, using Git Bash, after the regular setup, the autocompletion outputs the following message:

$ platform bash: C:UsersIEUser.platformshbinplatform: command not found

It seems that the \ character is stripped in the path. Having a look at the generated autocompletion.sh, I figured out that the path to the platform CLI is not quoted, and the \ are not doubled.

I was able to fix the issue by replacing:

RESULT=$(C:\Users\IEUser\.platformsh\bin\platform _completion --shell-type bash >/dev/null)

by:

RESULT=$("C:\Users\IEUser\.platformsh\bin\platform" _completion --shell-type bash >/dev/null)

After this change, and after sourcing ~/.bashrc, the autocompletion works. It ouputs another error, but probably related to the Windows Bash implementation:

__ltrim_colon_completions: command not found