Updated to nikic/PHP-Parser 4.x
#410
Conversation
kukulich
force-pushed
the
php-parser4
branch
4 times, most recently
from
29b9d86 to
8e1b674
Compare
|
Build is failing because PHPStan doesn't support PHP-Parser 4. @ondrejmirtes PHP-Parser 4 has been released 11 hours ago. What have you been doing the whole night? 😄 |
|
@kukulich @ondrejmirtes Have already create a issue there to track this migration 😂 |
|
Looks like my plans for Thursday have been sorted out for me 😂 |
|
@kukulich You could use the shim https://github.com/phpstan/phpstan-shim to focus on package upgrade at a time |
|
@TomasVotruba I tried, it doesnt work. |
|
That's how Rector can run on php-parser 4 and use phpstan with php-parser 3 at the same time:
What exactly did you try and what happened? |
|
PHP-Parser is not prefixed in the shim (because of extensions support) and
the two versions are incompatible.
On Sat, 3 Mar 2018 at 21:20, Tomáš Votruba ***@***.***> wrote:
It works for me:
- https://github.com/rectorphp/rector/blob/master/composer.json#L14
- https://github.com/rectorphp/rector/blob/master/composer.json#L26
What exactly did you tried and what happened?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#410 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAGZuJo_zkcf7pcogM-k6nJXx7HqKMPBks5tavr_gaJpZM4SXfS9>
.
--
Ondřej Mirtes
|
|
Then |
|
You can also go with |
|
I don’t think that any of this would help - PHPStan uses PHP-Parser 3.x
which it has loaded into memory and when the analysis stumbles upon code
that uses 4.x, it will always crash.
On Sun, 4 Mar 2018 at 13:19, Théo FIDRY ***@***.***> wrote:
You can also go with barmani/composer-bin-plugin. Until a shim is
officially provided or a PHAR is used it's a good solution.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#410 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAGZuO2LXC_FKVO28K_wvVx4EkywRIBZks5ta9vLgaJpZM4SXfS9>
.
--
Ondřej Mirtes
|
|
@kukulich what's the status of the PR? |
|
@theofidry The same. PHPStan doesn’t support PHP-Parser 4. The only solution is to disable phpstan check temporarily. |
|
Support in PHPStan is hopefully coming in June: i was relying on a
contributor to do the migration but he left the effort so I have to do it
myself.
On Tue, 8 May 2018 at 07:49, Jaroslav Hanslík ***@***.***> wrote:
@theofidry <https://github.com/theofidry> The same. PHPStan doesn’t
support PHP-Parser 4. The only solution is to disable phpstan check
temporarily.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#410 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AAGZuC3V7SRdSI4rTVXJjC4772RDUFh9ks5twTH_gaJpZM4SXfS9>
.
--
Ondřej Mirtes
|
|
Can't we disable PHPStan temporary then? It's a very nice to have tool but it doesn't feel like it should be a blocker neither and according to @ondrejmirtes it's gonna take another 1-2 months |
|
I'll prepare PR without PHPStan (already on my localhost) but I need to fix #422 first. |
|
Ready to review. I would like to have it in master so I can work again on #373 without future conflicts. |
| # env: DEPENDENCIES="" | ||
| # before_script: | ||
| # - travis_retry composer require --dev --prefer-dist --prefer-stable phpstan/phpstan:^0.9.2 | ||
| # script: vendor/bin/phpstan analyse -l 5 -c phpstan.neon src |
There was a problem hiding this comment.
I suppose it is impossible to analyse this codebase due to autoloading reflection?
| if ($node instanceof Node\Scalar\MagicConst\Class_) { | ||
| return $this->compileClassConstant($context); | ||
| } | ||
| $constExprEvaluator = new ConstExprEvaluator(function (Node\Expr $node) use ($context) { |
There was a problem hiding this comment.
Can the evaluator instance be preserved?
There was a problem hiding this comment.
No, because it needs the fallback evaluator with the right context :( I tried to implemented it without the fallback evaluator and call other evaluations in catch but it doesn't work. The test CompileNodeToValueTest::testClassConstantResolutionWithAnotherClassConstant was failing because the test is combination of multiplication (supported by PHP-Parser itself) and class constants (supported by BR).
|
Thanks @kukulich! |
|
When can we expect a tag? It would help other packages to bump on php-parser 4.0 |
|
I was going to create a new tag, but I'll most likely first merge in #423 first |
|
Oh, that's great 👍 Sure! |
No description provided.