PIO shortcoming in Inclusion / Exclusion of libraries

I an missing an option to exclude a specific library. Here my trouble example. I use:

#include <Arduino.h>
#include <ESPAsyncWebServer.h>

and in my pio.ini have:

    board = esp32dev
    framework = arduino
    platform = 
        https://github.com/platformio/platform-espressif32.git#feature/idf-master
    platform_packages = 
        framework-arduinoespressif32 @ https://github.com/espressif/arduino-esp32#2.0.0-rc1
    lib_deps = 
        https://github.com/me-no-dev/ESPAsyncWebServer.git#master
        https://github.com/me-no-dev/AsyncTCP.git#master
        powerbroker2/SafeString@^4.1.5
    lib_ignore = 
        LittleFS_esp32

So, I want to use a specific webserver AND for it a specific AsyncTCP. However, as long as the default AsyncTCP still exists in the project, both seem to be included as seen in the Dependency graph:
Auswahl_061

Setting the old lib to “Ignoring” is not possible, because within the triangular brackets the old and new one have the identical name AsyncTCP - I can only ignore either both or none.

I am not sure about what happens with the two libs are seemingly both included, but some big troubles in my project went away only after I uninstalled the old lib!

This shortcoming could be overcome if the lib_ignore allowed the same type of specification for version like e.g. @^1.1.1 as is used for lib_deps.

Indeed, that was also noted in Libraries showing twice in .pio/libdeps - #8 by maxgerhardt. Referencing a library via git will not change the fact that the library itself references a dependent library via a regular name and version in the library.json, so that will be pulled in from PlatformIO too, if the user additionally specifies a differing version of the library to be in the project, that too will be in there, thus you see it double.

Looks like I am not the first and only one to be stuck in that trap. And with all the chances for non-predictability of what will happen with which lib, doing a fork of something in development, which may change daily, to maybe perhaps, not even fully sure, correct the problem seems less than workable!

I did consider my post as a bug-report, but it is not perceived as such?

And my suggestion to offer a way to exclude a specific version of a lib is also not being considered?

I have problems with the two libs webserver and AsyncTCP. The point of specifying both is to test them and their interaction. It is thus pointless to test webserver with its old AsyncTCP; using it for the webserver is the only purpose to have it!