EZP-31930: Improved output of composer ezplatform-install command

[alongosz]

JIRA: EZP-31930
Target version: 2.5.x, 3.1.x, 3.2

This is yet another attempt like ezsystems/ezpublish-kernel#2593 to fix the output of ezplatform:install Symfony command piped through Composer (composer ezplatform-install).
The issue is visible e.g. here or locally.

Seems it looks better when bin/console has --ansi option. TBD if this solves the issue for all platforms.

QA

See TODO list.

TODO

• Create JIRA ticket if this works
• Platforms: Linux hosted on Travis: slightly better
• Platforms: Linux hosted locally: works very well locally on Ubuntu 20.04 (@alongosz).
• Platforms: Linux hosted VBox (TBD by QA)
• Platforms: MacOS: TBD
• Platforms: Windows: TBD

[bogusez]

@alongosz here you can see how the script behave on ezplatform2.5 (macOS) after script is finish.

[bogusez]

3.1 commerce
When --ansi parameter is provided then an exception occur while script is processed.

[alongosz]

@bogusez
OSS/EE 2.5: Reindexing command output is buffered through install command, which causes the mentioned effect. The same can be observed when used directly via php bin/console --ansi ezplatform:install clean. When reindexing is called directly (php bin/console ezplatform:reindex) this does not occur. I'm not able to find a fix for this case, I'm getting that output from Symfony Process.

EC 3.1: Cannot reproduce.

[bogusez]

@alongosz thank you for information. By the way, the problem on EC 3.1 is already solved (output formatting works properly without any errors).

[bogusez]

@alongosz I have tested formatting output on all systems. I have found several discrepancies especially on windows OS (all ezplatfrom versions: 2.5, 3.1 and 3.2) but also on macOS (ezplatform 2.5 version). The noticed by me problem might be combined with symfony problem that you mentioned before. Here are the screen shots:

Windows / ezp 2.5

Windows / ezp 3.1

Windows / ezp 3.2

MacOS / ezp 2.5

[alongosz]

Hmm, not sure if anything can be done with those. If I add some extra lines, for the cases where it's correct now it's gonna produce 2 blank lines. The question is if we're willing to accept the current state, given it improves output significantly, or drop it because it's not complete. I mean, I'd go with it, but I'm biased ;-)

POV ping @ezsystems/php-dev-team

[lserwatka]

Hmm, not sure if anything can be done with those. If I add some extra lines, for the cases where it's correct now it's gonna produce 2 blank lines. The question is if we're willing to accept the current state, given it improves output significantly, or drop it because it's not complete. I mean, I'd go with it, but I'm biased ;-)

POV ping @ezsystems/php-dev-team

Hmm, not sure if anything can be done with those. If I add some extra lines, for the cases where it's correct now it's gonna produce 2 blank lines. The question is if we're willing to accept the current state, given it improves output significantly, or drop it because it's not complete. I mean, I'd go with it, but I'm biased ;-)

POV ping @ezsystems/php-dev-team

I would add this extra line. I don't have problem with extra space. This looks broken to me.

[alongosz]

@bogusez further adjustments for the commands causing this are available via ezsystems/ezpublish-kernel#3070
Let me know if that works for you.