adrian

Tracy Debugger

Recommended Posts

Introducing a new module which brings the Tracy Debugger (https://tracy.nette.org/) to ProcessWire.
 
A big thanks to @tpr for introducing me to Tracy and for the idea for this module and for significant feedback, testing, and feature suggestions.

The most comprehensive set of instructions and examples is now available on the PW blog: https://processwire.com/blog/posts/introducing-tracy-debugger/

Modules Directory: http://modules.processwire.com/modules/tracy-debugger/
Github: https://github.com/adrianbj/TracyDebugger

  • quickly detect and correct errors with an expandable call stack tree
  • log errors (and optionally receive emails when an error occurs in production mode)
  • dump variables
  • measure execution time of scripts/queries
  • see memory consumption between breakpoints

 
Debug Bar
 
I have added a custom ProcessWire panel which shows all the content of the current page object, along with all the info from the admin debug tools mode.
 
It also includes a PW Debug Mode badge to warn you that debug mode is on - great reminder for live sites.
 
A Method Info tab details the dumping/logging calls that you can make from your templates etc.
 
Screen Shot 2016-02-15 at 1.19.29 PM.png
 
 

Error Reporting
 
Expandable call stack of errors and other useful information:
 
joxi_screenshot_1455322556737.png
 
 
Manual Dumping, Logging, and Timing

Content can be dumped to the page via TD::dump() or to the debug bar via TD::barDump(), or logged via TD::log() from PW template files. eg.

 

 

 

TD::barDump($page->body, 'Body Field');
TD::barDump($fields->body, 'Body Settings');

Screen-Shot-2016-02-12-at-4.07.05-PM.png
 

TD::dump($page);

Screen Shot 2016-02-12 at 4.11.34 PM.png
 
 
Logging to Console (FireLog)
 
This works with Chrome and Firefox

TD::fireLog()
fireLog()
fl()

To make this work you need to install these browser extensions:
 
Chrome:
https://github.com/MattSkala/chrome-firelogger
 
Firefox:
http://www.getfirebug.com/
http://firelogger.binaryage.com/
 
Screen Shot 2016-02-15 at 1.21.36 PM.png 
 
Logging to File

TD::log('Log Message');

By default, manually logged content is sent to: /site/assets/logs/tracy/info.log, but you can specify an optional second parameter to one of the following: 'debug', 'info', 'warning', 'error', 'exception', 'critical' files.
eg. which will put the message in the debug.info file. 

 

 

 

TD::log('Log Message', 'debug'); 

 
Available Methods

TD::dump()
TD::barDump()
TD::log()
TD::fireLog()
TD::addBreakpoint()
TD::timer()

 
You can also use these shortcuts/aliases:
 
TD::dump($var)

dump()
d()

 
TD::barDump($var, $title)

barDump()
bd()

 
TD::log($message)

l()

Just in case you are wondering, log() is already defined in PW, which is why it is not included here.
 
 
TD::fireLog($message)

fireLog()
fl()

TD:addBreakpoint($name)

addBreakpoint()
bp()

 
TD:timer($name)

timer()
t()

 
These aliases/shortcuts can be disabled in the module config setting, either altogether, or individually, just in case you have functions with the same names defined them elsewhere on your site and are getting a "previously declared" error.
 
  
 
Config Settings 
 
There are several config settings to control the output of errors, dumps, logging, and emails of errors. If you find any additional Tracy settings that you would also like to have access to, please let me know and I'll be happy to add them.

  • Like 37

Share this post


Link to post
Share on other sites

Thanks to a bug report from horst I just committed a new version that fixes a Windows directory path issue.

He also pointed me to another problem which is turns out means that this module currently only works with the 3.x branch of PW. I will look into supporting 2.x, but for the moment if you want to test, you will need 3.x or you can probably make it work on 2.7 if you uncheck "Show ProcessWire Panel" in the config settings. You'll miss out on that functionality, but hopefully everything else should work.

  • Like 1

Share this post


Link to post
Share on other sites

That is something nearly every developer can make use of. Thanks adrian and tpr!

Could you please suggest a proper way to debug process modules stuff with it? On the tracy page I have read about outputting stuff to fiebug...

 
 
 

Share this post


Link to post
Share on other sites

That is something nearly every developer can make use of. Thanks adrian and tpr!

Could you please suggest a proper way to debug process modules stuff with it? On the tracy page I have read about outputting stuff to fiebug...

 

Thanks Ivan, there are several options - Firebug is actually something I haven't implement yet in this module. It requires an additional package which I haven't included. I figure that most of us are on Chrome, so didn't bother with it yet. But if there is interest, I will definitely take a look.

I should have pointed to the best of the Tracy docs: https://doc.nette.org/en/2.3/debugging - for some reason the ones on the main page don't include the barDump() option yet.

You can use: TD::dump(), TD::barDump(), TD::log()

I also implemented some aliases - d(), bd(), l() as very short options for lazy typists, but not sure about keeping these due to possible name conflicts.

Anyway, let me know if that links helps and if you have any specific questions.

Edited by adrian
Corrected info about the Firebug option.

Share this post


Link to post
Share on other sites

Thanks everyone for the great interest in this module!

I have just added 2 new sections to the PW panel that shows details of all fields on the current page.

I would love to hear from you if you have any ideas for other sections/info that you would find useful to have in the PW panel.

List View

Shows basic details with the name linked to the edit page for that field:

post-985-0-77931100-1455395881_thumb.png

Object View

Shows the fields object - for details of each field, expand the "data" section:

post-985-0-51904500-1455395880_thumb.png

  • Like 4

Share this post


Link to post
Share on other sites

Sorry, this is just too much fun :)

Just added a Current Page Info section, with links to edit the page (from the name), edit the template, and open the page tree with this page expanded to show the children (from the number of children).

post-985-0-19806300-1455399024_thumb.png

  • Like 3

Share this post


Link to post
Share on other sites

Looks like you're really in love with Tracy :)

Actually I am, and have been for 18 years now, but that is a very different kind of love ;)

  • Like 4

Share this post


Link to post
Share on other sites

That was high time then to make a module from it :)

Page Info: it's useful, but it's a bit inconvenient to show. How about adding it to a separate tab? That way it would only require to hover. 

I don't know how much work is it, just asking.

Share this post


Link to post
Share on other sites

That was high time then to make a module from it :)

Page Info: it's useful, but it's a bit inconvenient to show. How about adding it to a separate tab? That way it would only require to hover. 

I don't know how much work is it, just asking.

New tab/panel isn't hard to do at all, but I wonder whether just having that section always expanded would be ok/better?

There are so many options and ways to present info that I am not totally sure the best option yet. I'd really love some thoughts from others about what they think they would use/access most often. Perhaps I could add several PW tabs/panels with different elements on each. Anyone have any ideas on what else they would like to have accessible and what should be grouped with what?

EDIT: I have just made the Page Info section open by default so a hover over the PW tab is enough to see that info. Still amenable to making separate tabs/panels if that is the consensus though.

  • Like 1

Share this post


Link to post
Share on other sites

Just been playing around with shortcuts to the various dump/log methods.

TD::dump()
TD::barDump()
TD::log()
TD::fireLog()
TD::timer()

 
You can now also use these shortcuts/aliases:
 
TD::dump($var)

dump()
d()

 
TD::barDump($var, $title)

barDump()
bd()

 
TD::log($message)

l()

Just in case you are wondering, log() is already defined in PW, which is why it is not included here.
 

TD::fireLog($message)

fireLog()
fl()

TD:timer($name)

timer()
t()

These aliases/shortcuts can be disabled in the module config setting, either altogether, or individually, just in case you have functions with the same names defined them elsewhere on your site and are getting a "previously declared" error.
 

I'd really appreciate any feedback on these shortcuts/aliases - I think they make dumping/logging substantially quicker, but I am definitely open to suggestion/criticisms to this approach.

Share this post


Link to post
Share on other sites

In followup to that last post, I decided to add an option to load the Tracy library for non-superusers.

This will NOT show errors or the debug bar for non-superusers. It just loads the Tracy library so that you can call Tracy methods directly, eg. Tracy\Debugger::dump() without non-superusers getting a "Fatal error: Class \'Tracy\Debugger\' not found" fatal error.
 
The downside to this setting is that the library is loaded for everyone - while it may not be a significant increase in resources, it is still something, so in general I recommend leaving this unchecked.

Share this post


Link to post
Share on other sites

How about making a dummy Tracy class with barDump/dump/log methods to load for non-superusers? (as a workaround)

This way less resources would be needed.

  • Like 1

Share this post


Link to post
Share on other sites

How about making a dummy Tracy class with barDump/dump/log methods to load for non-superusers? (as a workaround)

This way less resources would be needed.

I think that sounds like a decent approach, so I just implemented it.

Now there is no need to worry if for some reason you want to make direct calls via (Tracy\Debugger::method()) on a live site as non-superusers will just get the dummy methods which return nothing.

I also just implemented the TD::timer() method, along with timer() and t() shortcut aliases.

With timer() you can do:

t();
// insert resource intensive code here
bd(t());

to dump the results of the time to the debug bar.

  • Like 1

Share this post


Link to post
Share on other sites

t() returns number like 3.0994415283203e-6 Could that be "better" formated?

Sure, it is now rounded to 4 decimal places. This matches the precision of PW's built in Debug::timer() which of course you could also use and dump or bardump to Tracy.

Is that what you were getting at?

I have also added the name of the timer to the output - if you specify it like:

t('first');
sleep(1);
bd(t('first'));

t('second');
sleep(2);
bd(t('second'));

This way you can have multiple timers and output them at once, and know which value responds to which timer.

Please let me know if that is better for you, or if you have any more specific suggestions.

Share this post


Link to post
Share on other sites

There's a typo in the name of "ShortcutMethods": the first "S" is lowercase in the filename which causes an error (perhaps not on every server).

Share this post


Link to post
Share on other sites

There's a typo in the name of "ShortcutMethods": the first "S" is lowercase in the filename which causes an error (perhaps not on every server).

That's weird - it was actually a git issue - I had already changed the filename locally to match the require statement, but git didn't update it when syncing. Looks to be fixed now.

  • Like 1

Share this post


Link to post
Share on other sites

4 decimal places is ok (from my point of view), but now it show 0.0 instead of 0.0000

Weird - I am not seeing that at all - could you please post exactly how you are calling this? Also, how many times have you reloaded the page. As you can image if the result was 3.100000000 it would round to be 3.1, rather than 3.1000, although it seems unlikely that the result would only have 2 significant figures.

I am wondering what might be different on your setup. I don't see why PHP version would matter and can't think of anything else.

Anyone else out there seeing only one decimal place?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


  • Recently Browsing   0 members

    No registered users viewing this page.

  • Similar Content

    • By Macrura
      Field Descriptions Extended
      This module enables you to extend field descriptions by dividing short descriptions with a longer text that is revealed in a toggle.
      Github: https://github.com/outflux3/FieldDescriptionsExtended
      Extending your field descriptions using the standard field's description field.
      Once this module is installed, it will automatically search your description field for the presence of 5 dashes (-----).
      Any content above the 5 dashes will be visible and the content below the dashes will be hidden. A 'More...' link will appear at the end of the short description which when clicked will reveal the rest of the description.
      Using Simple Markdown Editor with the description field
      If you have Simple Markdown Editor (InputfieldSimpleMDE) installed, you can enable the field description to have that editor.
      *When using Simple MDE, you can use the button (Insert Horizontal Line) instead of typing 5 dashes. More about SimpleMDE.
      Extending your field descriptions using content from a ProcessWire Page for the field description.
      You may use the content from a ProcessWire page as a field description. This would allow you to easily insert images, links, and use hanna codes.
      To use page content for your field descriptions, please follow these instructions:
      Install Select Fields module (FieldtypeFields) http://modules.processwire.com/modules/fieldtype-fields/ Create a new field using this field type, e.g. field_select. Add the field to any template you will be using for your field descriptions. Setup your help pages (for example under a settings branch) where you will store the field description content,using the template containing the Field Select. Add content to a page and select the field where that content should show. To show a short text before the link to the longer content, separate them with 5 dashes Be sure to update your settings on this page, first enable page content descriptions,then specify the name of the Select Fields field, template to search, and content field. If you create a field description using this method, please note that the description field must be blank for contexts where you want the page content to appear.
      You can freely use template context for field descriptions, but the Page Content method is not context sensitive and will display under all contexts where the description is blank.
      ----
      original post:
      This is a new module, hope to release soon, which allows extended field descriptions, in currently 2 ways.
      The main feature of the module is that you can have a short description and then a 'more...' link which drops down a longer block of text.
      This is achieved by separating the intro/visible text and the rest with 5 dashes.

      Example setup:

      the 2nd way is if you are using AdminThemeUiKit, you can show extended field instructions in a panel. The content of the panel is edited on a regular PW page. This use case would probably not be that common, but if you had a field that required some extended instructions for how to use, this could be useful; Also, since this allows you to target information and instructions down at the field level, it could reduce the amount of documentation needed on a global level, since it is a lot more context targeted.

    • By netcarver
      Part 1 of a 2 part Module & Service Reveal.
      I'm currently working on a new module: ModuleReleaseNotes that was inspired by the work I originally did on making Ryan's ProcessWireUpgrades module "release" aware. In the end, I decided to ditch the approach I was originally taking and instead work on a module that hooked in to the UpgradeConfirmation dialog and the module edit page.
      Aims
      My aims for this module are as follows...
      Make discovery of a module's changes prior to an upgrade a trivial task. Make breaking changes very obvious. Make reading of a module's support documentation post-install a trivial task. Make module authors start to think about how they can improve the change discovery process for their modules. Make sure the display of information from the module support files/commit messages doesn't introduce a vulnerability. Looking at these in turn...
      Making discovery of a module's changes prior to upgrade a trivial task.
      This is done by adding a "What's changed section" to the upgrade confirmation dialog.  This section takes a best-effort approach to showing what's changed between the installed version and the updated version that's available via the module repository.
      At present, it is only able to talk to github-hosted repositories in order to ask them for the release notes, the changelog file (if present) and a list of commits between the git tag that matches the installed version and the tag matching the latest version.
      It will display the Release Notes (if the author is using the feature), else it will display the commits between the tags (if tagging is used by the module author) else it will show the changelog file (if present) else it will show the latest N commits on the master branch (N, of course, being configurable to your liking.)
      An example of the Github Release Notes pulled in for you, taken from Mike Rockett's TextformatterTypographer Module...

      An example of a tag-to-tag commit list from the same module...

      An example of a changelog - formatted to show just the changes (formatting styles will change)...

      Finally, an example of a fallback list of commits - sorry Adrian ...

       
      Making breaking changes obvious.
      This is currently done by searching for a set of configurable search strings. Later versions may be able to support breaking change detection via use of Semantic Versioning - but this may require some way of signalling the use of this versioning standard on a module-by-module basis.
      For now, then, you can customise the default set of change markers. Here I have added my own alias to the list of breaking change markers and the changes section of the changelog is styled accordingly (these will be improved)...

       
      Make reading of a module's support documentation, post-install, a trivial task.
      This is done by making some of the support files (like the README, CHANGELOG and LICENSE files) readable from the module's information/settings screen. There is an option to control the initial open/closed state of this section...

      Here is Tracy's README file from within the module settings page... 

       
      Make module authors start to think about how they can improve the change discovery process for their modules.
      There are notes in each of the sections displayed on the upgrade confirmation page that help authors use each of the features...

       
      Make sure display of external information doesn't introduce a vulnerability.
      This is an ongoing concern, and is the thing that is most likely to delay or prevent this module's release lead to this module's withdrawl should a vulnerability be found. Currently, output is formatted either via Markdown + HTML Purifier (if it was originally a Markdown file) or via htmlspecialchars() if it has come from a plaintext file.
      If you discover a vulnerability, please get in contact with me via the forum PM system.
       
      Ongoing...
      For now, I've concentrated on integration with GitHub, as most people use that platform to host their code. I know a few people are hosting their repositories with BitBucket (PWFoo comes to mind) and some with GitLab (Mike Rockett?) and I would eventually like to have adaptor implementations for these providers (and perhaps GitKraken) - but for now, GitHub rules and the other hosts are unsupported.
       
      Links
      Github: ModuleReleaseNotes
      PW Module Repository: Here (pending?)
    • By blynx
      Hej,
      A module which helps including Photoswipe and brings some modules for rendering gallery markup. Feedback highly appreciated

      Modules directory: http://modules.processwire.com/modules/markup-processwire-photoswipe
      .zip download: https://github.com/blynx/MarkupProcesswirePhotoswipe/archive/master.zip
      You can add a photoswipe enabled thumbnail gallery / lightbox to your site like this. Just pass an image field to the renderGallery method:
      <?php $pwpswp = $modules->get('Pwpswp'); echo $pwpswp->renderGallery($page->nicePictures); Options are provided like so:
      <?php $galleryOptions = [ 'imageResizerOptions' => [ 'size' => '500x500' 'quality' => 70, 'upscaling' => false, 'cropping' => false ], 'loresResizerOptions' => [ 'size' => '500x500' 'quality' => 20, 'upscaling' => false, 'cropping' => false ], 'pswpOptions' => (object) [ 'shareEl' => false, 'indexIndicatorSep' => ' von ', 'closeOnScroll' => false ] ]; echo $pswp->renderGallery($page->images, $galleryOptions); More info about all that is in the readme: https://github.com/blynx/MarkupProcesswirePhotoswipe
      What do you think? Any ideas, bugs, critique, requests?
      cheers
      Steffen
    • By mtwebit
      Tasker is a module to handle and execute long-running jobs in Processwire. It provides a simple API  to create tasks (stored as PW pages), to set and query their state (Active, Waiting, Suspended etc.), and to execute them via Cron, LazyCron or HTTP calls.
      Creating a task
      $task = wire('modules')->Tasker->createTask($class, $method, $page, 'Task title', $arguments); where $class and $method specify the function that performs the job, $page is the task's parent page and $arguments provide optional configuration for the task.
      Executing a task
      You need to activate a task first
      wire('modules')->Tasker->activateTask($task); then Tasker will automatically execute it using one of its schedulers: Unix cron, LazyCron or TaskerAdmin's REST API + JS client.
      Getting the job done
      Your method that performs the task looks like
      public function longTask($page, &$taskData, $params) { ... } where $taskData is a persistent storage and $params are run-time options for the task.
      Monitoring progress, management
      The TaskerAdmin module provides a Javascript-based front-end to list tasks, to change their state and to monitor their progress (using a JQuery progressbar and a debug log area). It also allows the on-line execution of tasks using periodic HTTP calls performed by Javascript.

       
      Monitoring task progress (and log messages if debug mode is active)

       
      Task data and log

      Detailed info (setup, task dependencies, time limits, REST API etc.) and examples can be found on GitHub.
      This is my first public PW module. I'm sure it needs improvement
       
    • By netcarver
      A very simple textformatter that was inspired by Diogo's RemoveHeight textformatter.
      This one strips the height from any images and either adds a custom class or adds a max-width:100% as an embedded style.
      Github: https://github.com/netcarver/TextformatterFluidImages
      PW Repo: To Be Confirmed.