Symfony2: dev to prod, win to unix, clear cache - windows

I have written a small Symfony2-Webapp. This worked on win in dev and prod. Than I uploaded all files to the provider and run into some issues:
The log-file tells me, that the twig-engine can't find a view. This helped, because a saw, that I had to correct one capital (index instead of Index) - it's always the same developing under windows and having production under unix.
I corrected the file and uploaded it. Than I deleted the cache (no console).
Still, the error resists!
What can I do?
New informations:
The problem exists, when not warming up the cache (dev and prod). If using console cache:warmup, than the cache is generated right.
As I found out, the in one generated file, the function 'getTemplateName()' is wrong. All other references in that file are correct.
Sadly, I have no console on my server.
EDIT: Answers:
1) Error is: request.CRITICAL: Uncaught PHP Exception InvalidArgumentException: "Unable to find template "sisicalBundle:Index:index.html.twig"." at /var/www/web1261/html/sisical/vendor/symfony/symfony/src/Symfony/Bridge/Twig/TwigEngine.php line 133 {"exception":"[object]
I searched the hole src-folder in eclipse: none of my code uses 'Index' (with capital) any more (since I corrected this). Does symphony create sources in 'vendors...' when installing them via composer?
2) I think I don't use assets and they are not the problem.
3) As logs and cache are written, permissions seem to be OK.
4) Good idea, but I can't find an use. I use templating and for that no 'use'-imports for templates. Seaching for 'Index' I found the wrong template-name generated in the cache. Clearing and warming cache does not help.
Is it possible, that twig generates wrong when using Index as a package/folder name?

I debugged a little bit:
When warming the cache, the FilesystemLoader parses the file system... > works well. But when running with no cache, the template-engine uses the TemplateGuesser.php which calculates the foldername from the controllername. This one is capital, because it's from a class-name....
I now saw, that I use upper cases in my other project. This time I thought 'always lower cases is the easiest...'.

Related

Prestashop 1.7.6.5 not taking into account php module modifications

I've been modifying a module and for some reason prestashop does not take into consideration the code modifications in module.php.
I've made modifications a while back and I need to modify it again due to an update that raises an error, the problem is that I get the error even if I remove said line.
I get this type of error: PHP message: PHP Notice: Undefined index: ordersorderFilter!active in /modules/module/module.php on line 104
I've commented first than removed that line but I still get the error.
It seems like there is a cache somewhere that I can't find, I've deactivated all possible cache in prestashop, activated debug mode, deleted the dev and prod folders in /var/cache/ but I still get the error.
I've used grep on the server to try to find that line, and when I delete it from the file I can not find it anywhere else on the server (but can find it if I add it).
Anybody has any idea?
Thanks a lot!
Prestashop does not perform any kind of caching on the modules PHP files so the problem lies elsewhere or at "lower level" (webserver / proxy ie.).
So check your environment for server-side caching, proxies or PHP accelerators such as opCache that may not serve changes from the source server.
Also try renaming or deleting the offending file altogether and see if the error changes, if not you might simply be looking at the wrong file :)

Sulu, strange hash related error when trying to save page / post?

Until recently everything was working well. Now, when I try to save (create or update) any page or post I get error message at top of the form "Error - There was an error when trying to save the form".
In error log I see this error:
“Uncaught PHP Exception Sulu\Component\Rest\Exception\InvalidHashException: “The given hash for the entity of type “Sulu\Bundle\ArticleBundle\Document\ArticleDocument” with the id “9e0720a7-5565-4a6f-a735-8a186b8fef9b” does not match the current hash. The entity has probably been edited in the mean time.” at /var/www/html/vendor/sulu/sulu/src/Sulu/Component/Hash/RequestHashChecker.php line 53"
Tried clearing symfony cache, website cache from admin, restarting docker containers.
I'm totally unaware that I did something to cause this error. Please help.
Update: strange thing I just noticed. When I try to save some article and I get that error and go back to overview page (where i.e. all articles of that type are listed) then I see unchanged article title. But when I click to edit it I see changed title?!? Like title on overview page and title on edit page are not used from the same place? How is that possible?
Update:
Now even I setup once more project from scratch saving articles causes that error. Some more info:
In stack trace last command executed is:
in vendor/elasticsearch/elasticsearch/src/Elasticsearch/ConnectionPool/StaticNoPingConnectionPool.php (line 64)
an it shoots out “No alive nodes found in your cluster”.
And while I'm setting up the project when executing:
php bin/console ongr:es:index:create
I get error:
{"error":{"root_cause":[{"type":"resource_already_exists_exception","reason
":"index [su_articles/sWs5F1uzSFO8bFiZqF1Egw] already exists","index_uuid":
"sWs5F1uzSFO8bFiZqF1Egw","index":"su_articles"}],"type":"resource_already_e
xists_exception","reason":"index [su_articles/sWs5F1uzSFO8bFiZqF1Egw] alrea
dy exists","index_uuid":"sWs5F1uzSFO8bFiZqF1Egw","index":"su_articles"},"st
atus":400}
And when I run:
php bin/console ongr:es:index:create --manager=live
I get similar:
In Connection.php line 675:
{"error":{"root_cause":[{"type":"resource_already_exists_exception","reason":"index [su_articles_live/Pissm9ycRj-o79K4wrrD
AA] already exists","index_uuid":"Pissm9ycRj-o79K4wrrDAA","index":"su_articles_live"}],"type":"resource_already_exists_exc
eption","reason":"index [su_articles_live/Pissm9ycRj-o79K4wrrDAA] already exists","index_uuid":"Pissm9ycRj-o79K4wrrDAA","i
ndex":"su_articles_live"},"status":400}
Also to mention that now saving pages works, but saving articles doesn't.
This solved the issue on ElasticSearch index creation for me:
php bin/console ongr:es:index:drop --force
The error can happen in the following cases.
Expected Case somebody else did edit the same article like you and did save it
Unexpected Case your phpcr cache is not in sync
Unexpected Case you have a multi server setup but your cache.app is not configured to use a central cache
So if its one of the unexpected cases first you should clear your cache.pools with:
bin/console cache:pool:prune
If you have a multi server setup make sure you configure a central cache. Most use in this case a redis-server which you configure in your cache.yaml e.g.:
# config/packages/prod/cache.yaml
framework:
cache:
default_redis_provider: "%env(resolve:REDIS_DSN)%"
app: cache.adapter.redis
Also make sure that you use the latest version and maybe update your phpcr cache configuration based on the sulu/skeleton: https://github.com/sulu/skeleton/blob/2.x/config/packages/prod/sulu_document_manager.yaml, there you could when performance doesn't matter in your case disable the phpcr cache, I would not recommend that.

Error after login to Magento on core/Mage/Admin/Model/User.php on line 437

None of out programmers has changed anything to the code, but out of nothing now, when a user tries to login on the backend, we have this error:
Fatal error: Call to a member function children() on a non-object in
/home/r1tech/public_html/app/code/core/Mage/Admin/Model/User.php on
line 437
Any ideas why this is happening?
Not a real Magento problem.
Somehow the files got corrupted. I re-uploaded everything and now it works just fine.
Leaving this for the occasional googler with this issue:
We had this problem today, after we had tested different caching mechanisms for Magento. In the process, the file cache directory was somehow corrupted or otherwise made unusable. If you're not keen on reuploading the entire thing as Beto Frega suggests, you should be able to solve it by removing (or renaming) the var/cache (in the Magento webroot, not the server itself!!) directory and creating a new one, which you then obviously give write permissions to.
This solved the issue for us -- and incidentally also some performance issues in the backend, due to the sheer size of the old cache dir.

CodeIgniter 2.0.2 syntax error blank page

CodeIgniter 2.0.2 syntax error blank page
i have CodeIgniter 2.0.2 (CI) and its annoing, that when i write lots of code and then just get blank page. Hard to debug without error. i got error reporting E_ALL in php.ini and in config of CI its turned on for every case of enviroment (devel, product) just for sure...
Does anyone knows, where the bug can be? How to turn it on? I think CI is rewriting php.ini error reporting setting somewhere somehow...
Thanks.
The other thing to check in your php.ini are the 'display_errors' and 'display_startup_errors' variables, they should both be set to 1.
In my (little) experience with CI this issue has been the most frustating one. I can understand that no framework is perfect, and any web developer is used to deal with errors (no matter if the framework or the developer is to blame). But that ANY error results in a blank page, with ZERO information in logs, console, whatever, is unacceptable.
If I recall correctly, in my case this was caused by of some combination of CI (ab)using the "silent error" PHP operator (e.g. prepending some db statements with the # operator, a not very clever way of avoiding leaking potentially private information to the end user), together with some bad use of *error_reporting*. See my post here (and see how many answers I got!).
I reported another related CI's idiocyidiosyncrasy related with logging here (again, zero answers)
please check the index.php which locate in the SAME LEVEL of the application folder
inside the index.php the second line
line 2:
ini_set("display_errors", "1"); //ensure it is set to 1
line 23:
define('ENVIRONMENT', 'development'); // development | testing | production
grep your php.ini for the word "error", turn them to On etc...
Do you have output compression turned on in the config? I find that will cause white pages when you have errors.
$config['compress_output'] = FALSE; // Should be set to this when developing.
I had the same problem and eventually fixed it when i noticed I had no 'logs' directory in my application folder

Upload images problem: IO error. (Error #2038)

I'm using script which is uploading files to server via flash component. Sometimes, very rarely, when trying to upload images via Firefox I get following error: IO error #2038. Searching on the net I could find reason why is it really happening to me. But I found solution for my case:
I open IE6, do the same thing there (photos are always uploaded without problem) and the when I try again in Firefox problem disappears. If someone had similar problems maybe this could help or maybe this hint could help to someone discovering cause of the problem :)
I just had this error also using ASP.Net MVC and I discovered I just had my handler namespace wrong/missspellled in the web config section where the handlers go. HTH...
After far more time than I'd like to admit, here is the solution. Change the following variables in your php.ini to be large enough to handle your largest file:
; Maximum size of POST data that PHP will accept.
post_max_size = 32M
; Maximum allowed size for uploaded files.
upload_max_filesize = 32M
; Whether to allow HTTP file uploads. (this is on by default)
file_uploads = On
Since files are sent as POST data, both the first variables need to be set. Otherwise, there will be an error, but you won't see what the error is, except IO error #2038.
If you didn't add quote around the option names, try doing that. Sometimes in IE, quoting the option names fixes the issue. It would help if you posted some example code though.
Also make sure that the path to the server being used by the upload component is correct. If the path is a relative path, try to fully qualify it. IE can do some "magic" with relative paths that the other browsers cannot do, and sometimes trying the code from one browser to the other can tweak things.
Also, could be permissions...and having IE go there first could manipulate the permissions in a way that the other browser(s) then work "correctly".
I solved the problem.
In my office I use a proxy and I must change the settings to No Proxy.
Or if you do not use proxy, try to change in php.ini:
Hostname:Port localhost:0
to:
Hostname:Port localhost:80

Resources