Hook soup

Wednesday, 17 June, 2009.

Of late, I seem to keep stumbling upon Drupal hooks that I've never heard of before. For example, I was just reading a blog post about what you can't modify in a _preprocess() function, when I saw mention of hook_theme_registry_alter(). What a mouthful. I ain't seen that one 'til now. Is it just me, or are new hooks popping up every second day in Drupal land? This got me wondering: exactly how many hooks are there in Drupal core right now? And by how much has this number changed over the past few Drupal versions? Since this information is conveniently available in the function lists on api.drupal.org, I decided to find out for myself. I counted the number of documented hook_foo() functions for Drupal core versions 4.7, 5, 6 and 7 (HEAD), and this is what I came up with (in pretty graph form):

And those numbers again (in plain text form):

Drupal 4.7: 41
Drupal 5: 53
Drupal 6: 72
Drupal 7: 183

Aaaagggghhhh!!! Talk about an explosion — what we've got on our hands is nothing less than hook soup. The rate of growth of Drupal hooks is out of control. And that's not counting themable functions (and templates) and template preprocessor functions, which are the other "magically called" functions whose mechanics developers need to understand. And as for hooks defined by contrib modules — even were we only counting the "big players", such as Views — well, let's not even go there; it's really too massive to contemplate.

In fairness, there are a number of good reasons why the amount of hooks has gone up so dramatically in Drupal 7:

Splitting various "combo" hooks into a separate hook for each old $op parameter, the biggest of these being the death of hook_nodeapi()
The rise and rise of the _alter() hooks
Birth of fields in core
Birth of file API in core

Nevertheless, despite all these good reasons, the number of core hooks in HEAD right now is surely cause for concern. More hooks means a higher learning curve for people new to Drupal, and a lot of time wasted in looking up API references even for experienced developers. More hooks also means a bigger core codebase, which goes against our philosophy of striving to keep core lean, mean and super-small.

In order to get a better understanding of why D7 core has so many hooks, I decided to do a breakdown of the hooks based on their type. I came up with the "types" more-or-less arbitrarily, based on the naming conventions of the hooks, and also based on the purpose and the input/output format of each hook. The full list of hooks and types can be found further down. Here's the summary (in pretty graph form):

And those numbers again (in plain text form):

Type	No. of hooks
misc action	44
info	30
alter	27
delete	20
insert	13
load	12
update	10
validate	6
form	4
misc combo	4
prepare	4
view	4
presave	3
check	2

As you can see, most of the hooks in core are "misc action" hooks, i.e. they allow modules to execute arbitrary (or not-so-arbitrary) code in response to some sort of action, and that action isn't covered by the other hook types that I used for classification. For the most part, the misc action hooks all serve an important purpose; however, we should be taking a good look at them, and seeing if we really need a hook for that many different events. DX is a balancing act between flexibility-slash-extensibility, and flexibility-slash-extensibility overload. Drupal has a tendency to lean towards the latter, if left unchecked. Also prominent in core are the "info" and "alter" hooks which, whether they end in the respective _info or _alter suffixes or not, return (for info) or modify (for alter) a more-or-less non-dynamic structured array of definitions. The DX balancing act applies to these hooks just as strongly: do we really need to allow developers to define and to change that many structured arrays, or are some of those hooks never likely to be implemented outside of core?

I leave further discussion on this topic to the rest of the community. This article is really just to present the numbers. If you haven't seen enough numbers or lists yet, you can find some more of them below. Otherwise, glad I could inform you.

Hooks in Drupal 4.7 core

Hooks in Drupal 5 core

Hooks in Drupal 6 core

Hooks in Drupal 7 core

Hook	Type
hook_access	check
hook_actions_delete	delete
hook_action_info	info
hook_action_info_alter	alter
hook_aggregator_fetch	misc action
hook_aggregator_fetch_info	info
hook_aggregator_parse	misc action
hook_aggregator_parse_info	info
hook_aggregator_process	misc action
hook_aggregator_process_info	info
hook_aggregator_remove	delete
hook_block_configure	form
hook_block_list	info
hook_block_list_alter	alter
hook_block_save	misc action
hook_block_view	view
hook_boot	misc action
hook_comment_delete	delete
hook_comment_insert	insert
hook_comment_publish	misc action
hook_comment_unpublish	misc action
hook_comment_update	misc action
hook_comment_validate	validate
hook_comment_view	view
hook_cron	misc action
hook_db_rewrite_sql	alter
hook_delete	delete
hook_disable	misc action
hook_elements	info
hook_enable	misc action
hook_exit	misc action
hook_fieldable_info	info
hook_field_access	check
hook_field_attach_create_bundle	insert
hook_field_attach_delete	delete
hook_field_attach_delete_bundle	delete
hook_field_attach_delete_revision	delete
hook_field_attach_form	form
hook_field_attach_load	load
hook_field_attach_presave	presave
hook_field_attach_pre_insert	insert
hook_field_attach_pre_load	load
hook_field_attach_pre_query	misc action
hook_field_attach_pre_update	update
hook_field_attach_submit	misc action
hook_field_attach_validate	validate
hook_field_attach_view_alter	alter
hook_field_build_modes	info
hook_field_create_field	insert
hook_field_create_instance	insert
hook_field_delete	delete
hook_field_delete_field	delete
hook_field_delete_instance	delete
hook_field_delete_revision	delete
hook_field_info	info
hook_field_insert	insert
hook_field_load	load
hook_field_prepare_translation	prepare
hook_field_presave	presave
hook_field_read_field	load
hook_field_read_instance	load
hook_field_rename_bundle	update
hook_field_sanitize	info
hook_field_schema	info
hook_field_storage_create_bundle	insert
hook_field_storage_create_field	insert
hook_field_storage_delete	delete
hook_field_storage_delete_field	delete
hook_field_storage_delete_instance	delete
hook_field_storage_delete_revision	delete
hook_field_storage_load	load
hook_field_storage_query	misc action
hook_field_storage_rename_bundle	update
hook_field_storage_write	misc action
hook_field_update	update
hook_field_update_instance	update
hook_field_validate	validate
hook_field_widget	form
hook_field_widget_error	misc action
hook_field_widget_info	info
hook_file_copy	misc action
hook_file_delete	delete
hook_file_download	misc action
hook_file_insert	insert
hook_file_load	load
hook_file_move	misc action
hook_file_references	misc action
hook_file_update	update
hook_file_validate	validate
hook_filter	misc combo
hook_filter_tips	info
hook_flush_caches	misc action
hook_footer	misc action
hook_form	form
hook_forms	info
hook_form_alter	alter
hook_form_FORM_ID_alter	alter
hook_help	info
hook_hook_info	info
hook_image_toolkits	info
hook_init	misc action
hook_insert	insert
hook_install	misc action
hook_js_alter	alter
hook_link	info
hook_link_alter	alter
hook_load	load
hook_locale	info
hook_mail	misc action
hook_mail_alter	alter
hook_menu	info
hook_menu_alter	alter
hook_menu_link_alter	alter
hook_modules_disabled	misc action
hook_modules_enabled	misc action
hook_modules_installed	misc action
hook_modules_uninstalled	misc action
hook_node_access_records	misc action
hook_node_access_records_alter	alter
hook_node_build_alter	alter
hook_node_delete	delete
hook_node_delete_revision	delete
hook_node_grants	info
hook_node_grants_alter	alter
hook_node_info	info
hook_node_insert	insert
hook_node_load	load
hook_node_operations	info
hook_node_prepare	prepare
hook_node_prepare_translation	prepare
hook_node_presave	presave
hook_node_search_result	misc action
hook_node_type	info
hook_node_update	update
hook_node_update_index	misc action
hook_node_validate	validate
hook_node_view	view
hook_openid	misc action
hook_page_alter	alter
hook_perm	info
hook_prepare	prepare
hook_profile_alter	alter
hook_query_alter	alter
hook_query_TAG_alter	alter
hook_registry_files_alter	alter
hook_requirements	misc combo
hook_schema	info
hook_schema_alter	alter
hook_search	misc combo
hook_search_preprocess	misc action
hook_system_info_alter	alter
hook_taxonomy_term_delete	delete
hook_taxonomy_term_insert	insert
hook_taxonomy_term_load	load
hook_taxonomy_term_update	update
hook_taxonomy_vocabulary_delete	delete
hook_taxonomy_vocabulary_insert	insert
hook_taxonomy_vocabulary_load	load
hook_taxonomy_vocabulary_update	update
hook_test_finished	misc action
hook_test_group_finished	misc action
hook_test_group_started	misc action
hook_theme	info
hook_theme_registry_alter	alter
hook_translated_menu_link_alter	alter
hook_translation_link_alter	alter
hook_uninstall	misc action
hook_update	update
hook_update_index	misc action
hook_update_last_removed	misc action
hook_update_N	misc action
hook_update_projects_alter	alter
hook_update_status_alter	alter
hook_user	misc combo
hook_user_cancel	misc action
hook_user_cancel_methods_alter	alter
hook_user_categories	info
hook_user_load	load
hook_user_operations	info
hook_validate	validate
hook_view	view
hook_watchdog	misc action
hook_xmlrpc	info

(D7 list accurate as of 17 Jun 2009; type breakdown for D7 list added arbitrarily by yours truly)

Filed in: Computing • Drupal

Wow

Submitted by ksenzee@drupal.org on 17 June 2009 - 9:32pm.

That was a lot of work! Thanks for putting it together, and bringing up the (very valid) question.

My instinct is that lots of hooks don't hurt anything except performance. In my perfect world, there's a hook for just about everything. I don't need to know all of them -- I can just assume they're there. And they're all named so well they're no trouble to find in the API anyway. For that matter, after all the work you've done here, I'll just bookmark this page for my hook documentation needs. :)

» login or register to post comments

More hooks != bigger core or bad DX

Submitted by catch (not verified) on 18 June 2009 - 1:13am.

Nice work on the statistics, really interesting to read and a fun post.

Have to disagree on it being a bad thing though. In terms of DX, splitting the $ops out of hooks makes things a lot easier. Very few people need to worry about hook_node_search_index(), whereas you couldn't avoid reading about it if looking at the docs for hook_nodeapi(). Hooks don't really affect performance at all either - if a hook isn't implemented, the module_invoke_all() only needs to check if a module implements the hook, and when it finds it doesn't, job done. Modules can do nasty things in hooks of course, but they can do that anywhere else too.

Similarly adding hooks in different places also allows for performance improvements - letting modules interact earlier or later in processes to swap in their own implementation or make small alterations can reduce the need to use bigger sledgehammers (you can use hook_menu_alter() to completely replace a page callback with your own version, but why do that when you can use hook_form_FORM_ID_alter() just to set a checkbox to required).

» login or register to post comments

Very interesting insight

Submitted by Kieran Lal (not verified) on 18 June 2009 - 1:26am.

Hi, thanks for that summary. Very interesting insight. Small APIs are certainly more immediately attractive to developers. Hopefully, as catch indicates, these APIs are simpler to learn.

It would certainly be interesting to compare to other Web Application Frameworks.

Kieran

» login or register to post comments

More, cleaner hooks are better: D7 = Good

Submitted by Benjamin Melançon (not verified) on 18 June 2009 - 2:12am.

I add my thanks for the breakdown and analysis, but have to say very strongly that this apparent increase in hooks is very much a good thing. While you mentioned "de-opping" as a source, I think it is the dominant source, and having seven clearly defined hooks rather than one giant switch-statement style hook (nodeapi) is a major developer usability win.

On the broader philosophy, I can just say I have never experienced a hook I wished weren't in core, and frequently when developing run into situations where I did want a hook.

^{benjamin, Agaric Design Collective}

» login or register to post comments

Code registry - no performance loss

Submitted by moshe weitzman (not verified) on 18 June 2009 - 3:10am.

You missed a key difference for D7 - there is a code registry which tracks which modules implement which hooks so adding a hook has zero performance impact. Freed from performance concerns, we have added hooks that let developers freely do their thing. If you don't know about a hook or don't use it, you are no worse off than you were before the hook was added.

The upgrade docs are very carefully maintained these days. I assure you that all new hooks are documented there shortly after their birth. Folks who skim the upgrade docs or skip them are naturally unaware of new hooks. Thats fine, they'll stumble on them when they need them.

Adding a hook is typically a single line of code, and a bunch of lines of documentation. I am all for #smallcore, but hooks do not cause code bloat.

There is nothing "unchecked" about core drupal development. Every line gets strip searched before commit. Every hook gets interrogated about its life purpose.

If you are going to say "surely this is a concern", you ought to name 5 hooks which should not exist.

» login or register to post comments

DX, not performance

Submitted by Jaza on 18 June 2009 - 7:04pm.

Perfectly valid point, Moshe (and the other folks who commented similarly). Hooks do indeed have zero performance impact in D7. And that's great.

However, my main concern is not with the performance impact of hooks, but with the DX impact of them. I'm aware of the high standards and the up-to-date-ness of the upgrade docs and the API docs. But good documentation doesn't mitigate the fact that there are simply too many hooks for the average developer to easily get acquainted with. At the end of the day, anyone can see that the number of hooks in core has more than doubled between D6 and D7, and that fact is going to scare people no matter how good our docs are (and they ARE good!).

My biggest gripe would have to be with the alter hooks. If you want me to name 5 hooks which should not exist, I could pick almost any 5 of the alter hooks in D7 core that have been added more recently. Here are 5 such ones:

hook_action_info_alter
hook_block_list_alter
hook_system_info_alter
hook_theme_registry_alter
hook_update_projects_alter

(No doubt people will reply with valid reasons why each of these SHOULD be, and IS in core — go ahead, I'm not stopping you).

The original one was hook_form_alter, and then we added a few more sorely-needed ones, such as hook_link_alter, hook_mail_alter and hook_menu_(link_)alter. Now, the trend seems to be that any info-style hook should have an accompanying alter-style hook, and I find that ridiculous. There simply isn't a valid use case for adding that many alter hooks. It caters for some very small edge cases, at the cost of bloating both the code and the API docs. I'm also inclined to agree with Walkah's argument that he presented in "Why I Hate Drupal" at the DC conference — that the entire philosophy of alter hooks is bad, and that they're a poor substitute for more mainstream alternatives such as object-based inheritance.

The new field API hooks also make up over a quarter of the hooks in HEAD right now (49 / 183). The field API alone, therefore, has almost as many hooks as all of D5 core (which has 53). That makes me strongly suspect that the present field API is over-engineered. Hopefully, we'll be able to remove some of these hooks as the field API matures.

» login or register to post comments

GreenAsh

Hook soup