rm -Rf node_modules
rm package-lock.json
npm install --package-lock-only

Now its hanging

(⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂⠂) ⠇ idealTree:[redacted-current-folder]: sill idealTree buildDeps

still hanging... still hanging... still hanging... still hanging...

give up...

ctrl+c disable ipv6...

npm install --package-lock-only
up to date, audited 667 packages in 44s
72 packages are looking for funding
npm install --no-optional
added 665 packages, and audited 666 packages in 5s

I have used mousetrap in other projects successfully. They all have their annoying bugs and like you say, most of them are wildly out of date and unsupported nowadays with tonnes of bugs reported and unaddressed.

There really is space for an amazing new project for keyboard binding. Im mainly a PHP dev, but I dabble in JS, but as always lack of time is the problem :)

'cmd+a' works perfectly.

'cmd+a+d' doesn't work and the first press of 'cmd+a' in safari just selects the whole page.

According to https://github.com/jaywcjlove/hotkeys/issues/55 this should be possible?

Im going to close this. Having spent an hour looking through hotkeys issue list in their gitrepo it looks like most of these are upstream bugs, and bugs interacting with safari :-(

Hotkeys promises so much... it's just a little lacking at the moment.

set a hotkey for and item for cmd+n

attempt to use cmd+n

It opens a new browser window, when it should trigger the action/url from the item

The hotkeys docs https://github.com/jaywcjlove/hotkeys#defining-shortcuts state you should be able to return false to override browser default behaviour, I tried that in your source code and it did not work either, you are already having event.preventDefault() in your code.. Im no JS expert sorry, Im mainly PHP :)

// Returning false stops the event and prevents default browser events
// Mac OS system defines `command + r` as a refresh shortcut
hotkeys('ctrl+r, command+r', function() {
  alert('stopped reload!');
  return false;
});

Im going to close this. Having spent an hour looking through hotkeys issue list in their gitrepo it looks like most of these are upstream bugs, and bugs interacting with safari :-(

Hotkeys promises so much... it's just a little lacking at the moment.

This is repeatable in the demo here where you have homepage set as cmd+h but when pressing that, the browser window "hides" as thats the browser keystroke - and for some reason hotkeys is not overruling it.

set a hotkey for and item for cmd+n

attempt to use cmd+n

It opens a new browser window, when it should trigger the action/url from the item

The hotkeys docs state you should be able to return false to override browser default behaviour, I tried that in your source code and it did not work either, you are already having event.preventDefault() in your code.. Im no JS expert sorry, Im mainly PHP :)

// Returning false stops the event and prevents default browser events
// Mac OS system defines `command + r` as a refresh shortcut
hotkeys('ctrl+r, command+r', function() {
  alert('stopped reload!');
  return false;
});

'cmd+a' works perfectly.

'cmd+a+d' doesn't work and the first press of 'cmd+a' in safari just selects the whole page.

According to https://github.com/jaywcjlove/hotkeys/issues/55 this should be possible?

Assign the letter A to a menu item (Like "Your Account")

invoke cmd+k

note that Your Account has a [A] correctly - note you can arrow down and select it and it runs...

BUT note that if you type a in the search box the hotkey is invoked INCORRECTLY and navigates you away to the url of the action.

This is wrong, this should not happen.

I should be able to have hotkeys, but they should not invoke if I am in a search box, text box etc... (They dont invoke, correctly, when Im in a normal input box on my page - they do invoke when the dialog search box has focus - this is incorrect and needs a fix)

Without this fixed, it makes hotkeys unusable if they are single letters like a

are you taking feature requests?

1. Ability to color the text (for example using a css class to show a red text for dangerous commands
2. The ability to use FontAwesome (or other) icon classes as image instead of having to provide the svg source.

Loving this so far...

A quick bug report

1. invoke cmd+k
2. close the popup dialog
3. smash the keyboard sdfkguys
4. invoke cmd+K
5. Note that the smashed keys are in the search box, but not searched for, and overlaid the placeholder

ok I'll chime in...

Dual stack ip4 and IPv6 on latest MacOS Mac Studio with 128Mb RAM on 1Gbps fibre connection to the desk. both ip4 and ipv6 are valid and work perfectly all the time during the day with anything non-npm like

Dual stack ip4 and IPv6. = npm hangs 100% of the time.

Disable IPv6 and use ip4 only = npm works 100% of the time.

As soon as I disable ipv6 I can 100% of the time get npm upgrade to complete.

As soon as I enable ipv6 (dual stack with ip4 as is the norm) then 100% of the time npm upgrade stalls/hangs...

ok I'll chime in...

Dual stack ip4 and IPv6 on latest MacOS on 1Gbps figure to the desk. both ip4 and ipv6 are valid and work perfectly all the time during the day with anything non-npm like

Dual stack ip4 and IPv6. = npm hangs 100% of the time.

Disable IPv6 and use ip4 only = npm works 100% of the time.

As soon as I disable ipv6 I can 100% of the time get npm upgrade to complete.

As soon as I enable ipv6 (dual stack with ip4 as is the norm) then 100% of the time npm upgrade stalls/hangs...

I would vote just close it, there are workarounds which are more sensible than changing the current behaviour - once you understand that the email is NOT rendered when you send it to an async message queue, and only rendered when its "actually sent" by the message handler then the lightbulb goes on over your head and you can code for that scenario.

The docs have been updated to include this statement https://symfony.com/doc/current/mailer.html#sending-messages-async

Note that the "rendering" of the email (computed headers, body rendering, ...) is also deferred and will only happen just before the email is sent by the Messenger handler.

which covers this issue and makes it an educational issue instead of a bug to fix.

This is a valid issue

I am already sourcing the .zshrc and accepting the 4 second delay (my causing, my zshrc does a lot :-))

The problem though is that your code:

if [[ -n "{query}" ]]; then
	QUERY="{query}"
	BIN=${jb_bin}
	PROJECT_NAME=${jb_project_name}

	source ~/.zshrc

	$(${BIN} "${QUERY}")

    echo ${PROJECT_NAME}
fi

runs AFTER the pstorm script filter which runs src/index.js

I have tried modifying the PATH also by changing the pstorm script filter to be

export PATH="/Applications/PhpStorm.app/Contents/MacOS:$PATH"
./node_modules/.bin/run-node src/index.js "$1" "PhpStorm"

and that did not work either

Im still unable to get this plugin to find the phpstorm binary

No idea where else to specifically state where the binary is

I even tried symlinking the binary into /usr/local/bin without success.

    0 lrwxr-xr-x@  1 phil        61 Mar 13 12:50 phpstorm -> /Applications/PhpStorm 2023.1 EAP.app/Contents/MacOS/phpstorm
    0 lrwxr-xr-x@  1 phil        61 Mar 13 12:51 pstorm -> /Applications/PhpStorm 2023.1 EAP.app/Contents/MacOS/phpstorm

Back in 2013 https://github.com/twigphp/Twig/issues/962 - the solution proposed was to " recommend you to fix your code to avoid any issues in the future"

Well here we are in 2022... and using the Symfony Messenger component to run message:consume workers...

Now given that fetching services (auto injection) gets the SAME service every time

And given that a single message:consume process can run many messages, thus reusing the same container

And given that each message is going to render a Twig template, with a DIFFERENT set of sandboxed security policies in my case, I have today stumbled upon this error message on the SECOND message consumed by Message Handlers:

Unable to register extension "Twig\Extension\SandboxExtension" as extensions have already been initialized

My exact extract of code is not important, but here it is, basically the Security policy will change depending on the user that owns the template

The FIRST message consumed works perfect. The SECOND message errors with the above error.

        $policy     = new SecurityPolicy($tags, $filters, $methods, $properties, $functions);
        $sandbox    = new SandboxExtension($policy, true);
        $this->twig->setExtensions([$sandbox]);

Now reading https://symfony.com/doc/current/service_container/shared.html it says that if we dont share the service in services.php that it will create a new instance each time - this doesnt seem to be the case.

in services.php I tried the following with no luck.

    $services->set(Environment::class)
        ->args([service('twig.loader.native_filesystem')])
        ->share(false);

Maybe related? https://github.com/symfony/symfony/issues/40642

I dont think this is a support request, but rather reporting a limitation now of twig environment, which only allows a single set of extensions to be loaded, and thus in long running processes like messenger:consume (or other runtimes like RoadRunner) this can be a problem

Or maybe I have missed something and Im being an idiot again?

thoughts?

Sorry, so much time has elapsed since I opened this, or worked on this area in my code.

I have just checked, and my workaround was not to use sandboxing when rendering the twig template in the messenger fed long-running process, but to use sandboxing while the users were creating their custom twig layouts (Which are email bodies for reports and reporting) and to validate the user input when they designed their twig layouts before I saved their custom twig layouts to the db.

This then means when I pull the custom twig layouts from the users account, when im running in messenger fed long running processes, they are already "clean" and only contain the tags/methods/functions/filters that the sandbox allowed them, so its safe to run without a sandbox in the long running process Twig service environment, when rendering the actual template (for sending by email)

Hope that helps someone else.

I still think there is a valid issue to be fixed somewhere here, but as Im the only one with this specific issue, and I have worked around it a long time ago, I'll close this issue today as not to be just old noise.

https://gpgtools.org/releases/gpgsuite/release-notes.html

The issue with SMIME and drafts is fixed in the Feb 28th release of Official GPGMail.

Fixes a problem where saving a draft or sending a reply to an encrypted or signed message was not possible when replying the successful activation dialog would be shown on every Mail launch in case automatic support plan activation was configured [#1135]

The other issue with saving in drafts is fixed in the Feb 28th release of Official GPGMail.

Draft saving and sending error on macOS 13.2
Under specific circumstances, saving a draft or sending a reply to an encrypted or sign message failed with an error that attachments are too large or that the message could not bet sent

Neither are yet fixed in Free-GPGMail as the source code has not yet been released.

FAQ

• [X] Yes, my issue is not listed in troubleshooting guide.
• [X] Yes, my issue is not listed in other issues.

Describe the bug

in phpStorm 2023.1 EAP (without toolbox) it looks like the command line launcher pstorm is no longer a thing

To Reproduce

When you go to Tools -> Create Command Line Launcher you get:

The link for online help goes to https://www.jetbrains.com/help/phpstorm/working-with-the-ide-features-from-command-line.html

I have added the following in my .zshrc (but including the whole .zshrc slows things down anyway https://github.com/bchatard/alfred-jetbrains/issues/271)

export PATH="/Applications/PhpStorm 2023.1 EAP.app/Contents/MacOS:$PATH"

Expected behavior

Alfred can no longer find phpStorm to launch

Alfred Version

5.06[2110]

Node Version

19.7.0

NPM Version

9.5.1

Jetbrains Product & Version

phpStorm 2023.1 EAP

Installation type

Classic via .dmg

OS

macOS 13.3 Beta 2

To be clear, the current version (28th Feb) from GPGTools.org, and the nightly (27th Feb), still do not work on the very latest Beta of MacOS 13.3 Beta 2

maybe you just don't care that this is about to break all PHP 7 compatibility for Joomla 4.2.... ????

No worries - thanks for the quick reply, as long as I was not already missing some nice flag I could set, I'll just have to live with this for now :)

I get your point though if the array had more items in it, it might make sense to wrap it,

There are many crazy things happening with attributes and I dont think there are solid conventions yet for attributes.