Try to fix PHP/hhvm tests by working around an issue with SET NAMES #3220
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
I think we are running into this problem: facebook/hhvm#8111 We don't solve that problem, but we avoid it by specifying utf8 in a different fashion. I also took the opportunity to use the "TFB-database" hostname instead of the DBHOST environment variable because it's simpler that way. Here are the steps I took in a local TFB environment that make me think we're running into the same issue as that GitHub issue: - Run toolset/run-tests.py --mode verify --test hvm, see that it fails all tests with error messages that look like the application server simply isn't there - Enable logging in PHP/hhvm/deploy/config.hdf (basically copy the "Log" section over from config-debug.hdf), run the test again - Look at the log file it spit out, PHP/hhvm/error.log, notice messages like this: Core dumped: Segmentation fault Stack trace in /tmp/stacktrace.28653.log - /tmp is cleared by the TFB toolset each run, so I comment out that part of benchmarker.py that clears /tmp - Run the test again, look at the stacktrace file in /tmp, notice a stack trace referring to our implementation code like this: #0 PDO->__construct() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:14] TechEmpower#1 Benchmark->setup_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:31] TechEmpower#2 Benchmark->bench_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:10] TechEmpower#3 main() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:13] - Google a bit, come across a GitHub issue talking about a similar problem when they used "SET NAMES" - Comment the part of PHP/hhvm/once.php.inc that uses SET NAMES, run test again and it worked, except for fortunes which failed because of character encoding problems - Google for a different way to specify utf8 in the PDO constructor, try it again, all tests pass locally
zloster
pushed a commit
to zloster/FrameworkBenchmarks
that referenced
this pull request
Feb 7, 2018
…echEmpower#3220) I think we are running into this problem: facebook/hhvm#8111 We don't solve that problem, but we avoid it by specifying utf8 in a different fashion. I also took the opportunity to use the "TFB-database" hostname instead of the DBHOST environment variable because it's simpler that way. Here are the steps I took in a local TFB environment that make me think we're running into the same issue as that GitHub issue: - Run toolset/run-tests.py --mode verify --test hvm, see that it fails all tests with error messages that look like the application server simply isn't there - Enable logging in PHP/hhvm/deploy/config.hdf (basically copy the "Log" section over from config-debug.hdf), run the test again - Look at the log file it spit out, PHP/hhvm/error.log, notice messages like this: Core dumped: Segmentation fault Stack trace in /tmp/stacktrace.28653.log - /tmp is cleared by the TFB toolset each run, so I comment out that part of benchmarker.py that clears /tmp - Run the test again, look at the stacktrace file in /tmp, notice a stack trace referring to our implementation code like this: #0 PDO->__construct() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:14] #1 Benchmark->setup_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:31] #2 Benchmark->bench_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:10] #3 main() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:13] - Google a bit, come across a GitHub issue talking about a similar problem when they used "SET NAMES" - Comment the part of PHP/hhvm/once.php.inc that uses SET NAMES, run test again and it worked, except for fortunes which failed because of character encoding problems - Google for a different way to specify utf8 in the PDO constructor, try it again, all tests pass locally
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I think we are running into this problem:
facebook/hhvm#8111
We don't solve that problem, but we avoid it by specifying utf8 in a
different fashion. I also took the opportunity to use the "TFB-database"
hostname instead of the DBHOST environment variable because it's simpler
that way.
Here are the steps I took in a local TFB environment that make me think
we're running into the same issue as that GitHub issue:
Run toolset/run-tests.py --mode verify --test hvm, see that it fails
all tests with error messages that look like the application server
simply isn't there
Enable logging in PHP/hhvm/deploy/config.hdf (basically copy the
"Log" section over from config-debug.hdf), run the test again
Look at the log file it spit out, PHP/hhvm/error.log, notice messages
like this:
Core dumped: Segmentation fault
Stack trace in /tmp/stacktrace.28653.log
/tmp is cleared by the TFB toolset each run, so I comment out that
part of benchmarker.py that clears /tmp
Run the test again, look at the stacktrace file in /tmp, notice a
stack trace referring to our implementation code like this:
#0 PDO->__construct() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:14]
About PHP #1 Benchmark->setup_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/once.php.inc:31]
Benchmarking tool suggestion. #2 Benchmark->bench_db() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:10]
Update and improve Compojure #3 main() called at [/home/techempower/FrameworkBenchmarks/frameworks/PHP/hhvm/db.php:13]
Google a bit, come across a GitHub issue talking about a similar
problem when they used "SET NAMES"
Comment the part of PHP/hhvm/once.php.inc that uses SET NAMES, run
test again and it worked, except for fortunes which failed because of
character encoding problems
Google for a different way to specify utf8 in the PDO constructor,
try it again, all tests pass locally