Administrators can configure the look and feel of Bugzilla without having to edit Perl files or face the nightmare of massive merge conflicts when they upgrade to a newer version in the future. Templatization also makes localized versions of Bugzilla possible, for the first time. It's possible to have Bugzilla's UI language determined by the user's browser. More information is available in .

The template directory structure is that there's a top level directory, template, which contains a directory for each installed localization. The default English templates are therefore in en. Underneath that, there is the default directory and optionally the custom directory. The default directory contains all the templates shipped with Bugzilla, whereas the custom directory does not exist at first and must be created if you want to use it. There are two different ways of editing Bugzilla's templates, and which you use depends mainly on the method you plan to use to upgrade Bugzilla. The first method of making customizations is to directly edit the templates in template/en/default. This is probably the best method for small changes if you are going to use the CVS method of upgrading, because if you then execute a cvs update, any template fixes will get automagically merged into your modified versions. If you use this method, your installation will break if CVS conflicts occur. The other method is to copy the templates to be modified into a mirrored directory structure under template/en/custom. The templates in this directory automatically override those in default. This is the technique you need to use if you use the overwriting method of upgrade, because otherwise your changes will be lost. This method is also better if you are using the CVS method of upgrading and are going to make major changes, because it is guaranteed that the contents of this directory will not be touched during an upgrade, and you can then decide whether to continue using your own templates, or make the effort to merge your changes into the new versions by hand. If you use this method, your installation may break if incompatible changes are made to the template interface. If such changes are made they will be documented in the release notes, provided you are using a stable release of Bugzilla. If you use using unstable code, you will need to deal with this one yourself, although if possible the changes will be mentioned before they occur in the deprecations section of the previous stable release's release notes. Don't directly edit the compiled templates in data/template/* - your changes will be lost when Template Toolkit recompiles them. It is recommended that you run ./checksetup.pl after any template edits, especially if you've created a new file in the custom directory.

If you are making template changes that you intend on submitting back for inclusion in standard Bugzilla, you should read the relevant sections of the Developers' Guide. The syntax of the Template Toolkit language is beyond the scope of this guide. It's reasonably easy to pick up by looking at the current templates; or, you can read the manual, available on the Template Toolkit home page. One thing you should take particular care about is the need to properly HTML filter data that has been passed into the template. This means that if the data can possibly contain special HTML characters such as <, and the data was not intended to be HTML, they need to be converted to entity form, ie &lt;. You use the 'html' filter in the Template Toolkit to do this. If you forget, you may open up your installation to cross-site scripting attacks. Also note that Bugzilla adds a few filters of its own, that are not in standard Template Toolkit. In particular, the 'url_quote' filter can convert characters that are illegal or have special meaning in URLs, such as &, to the encoded form, ie %26. This actually encodes most characters (but not the common ones such as letters and numbers and so on), including the HTML-special characters, so there's never a need to HTML filter afterwards. Editing templates is a good way of doing a "poor man's custom fields". For example, if you don't use the Status Whiteboard, but want to have a free-form text entry box for "Build Identifier", then you can just edit the templates to change the field labels. It's still be called status_whiteboard internally, but your users don't need to know that.

Some CGI's have the ability to use more than one template. For example, buglist.cgi can output itself as RDF, or as two formats of HTML (complex and simple). The mechanism that provides this feature is extensible. Bugzilla can support different types of output, which again can have multiple formats. In order to request a certain type, you can append the &ctype=<contenttype> (such as rdf or html) to the <cginame>.cgi URL. If you would like to retrieve a certain format, you can use the &format=<format> (such as simple or complex) in the URL. To see if a CGI supports multiple output formats and types, grep the CGI for "GetFormat". If it's not present, adding multiple format/type support isn't too hard - see how it's done in other CGIs, e.g. config.cgi. To make a new format template for a CGI which supports this, open a current template for that CGI and take note of the INTERFACE comment (if present.) This comment defines what variables are passed into this template. If there isn't one, I'm afraid you'll have to read the template and the code to find out what information you get. Write your template in whatever markup or text style is appropriate. You now need to decide what content type you want your template served as. The content types are defined in the Bugzilla/Constants.pm file in the $contenttypes variable. If your content type is not there, add it. Remember the three- or four-letter tag assigned to you content type. This tag will be part of the template filename. After adding or changing a content type, it's suitable to edit Bugzilla/Constants.pm in order to reflect the changes. Also, the file should be kept up to date after an upgrade if content types have been customized in the past. Save the template as <stubname>-<formatname>.<contenttypetag>.tmpl. Try out the template by calling the CGI as <cginame>.cgi?format=<formatname>&ctype=<type> .

There are a few templates you may be particularly interested in customizing for your installation. index.html.tmpl: This is the Bugzilla front page. global/header.html.tmpl: This defines the header that goes on all Bugzilla pages. The header includes the banner, which is what appears to users and is probably what you want to edit instead. However the header also includes the HTML HEAD section, so you could for example add a stylesheet or META tag by editing the header. global/banner.html.tmpl: This contains the "banner", the part of the header that appears at the top of all Bugzilla pages. The default banner is reasonably barren, so you'll probably want to customize this to give your installation a distinctive look and feel. It is recommended you preserve the Bugzilla version number in some form so the version you are running can be determined, and users know what docs to read. global/footer.html.tmpl: This defines the footer that goes on all Bugzilla pages. Editing this is another way to quickly get a distinctive look and feel for your Bugzilla installation. global/variables.none.tmpl: This defines a list of terms that may be changed in order to "brand" the Bugzilla instance. In this way, terms like "bugs" can be replaced with "issues" across the whole Bugzilla installation. The name "Bugzilla" and other words can be customized as well. bug/create/user-message.html.tmpl: This is a message that appears near the top of the bug reporting page. By modifying this, you can tell your users how they should report bugs. bug/create/create.html.tmpl and bug/create/comment.txt.tmpl: You may not wish to go to the effort of creating custom fields in Bugzilla, yet you want to make sure that each bug report contains a number of pieces of important information for which there is not a special field. The bug entry system has been designed in an extensible fashion to enable you to add arbitrary HTML widgets, such as drop-down lists or textboxes, to the bug entry page and have their values appear formatted in the initial comment. An example of this is the mozilla.org guided bug submission form. The code for this comes with the Bugzilla distribution as an example for you to copy. It can be found in the files create-guided.html.tmpl and comment-guided.html.tmpl. So to use this feature, create a custom template for enter_bug.cgi. The default template, on which you could base it, is custom/bug/create/create.html.tmpl. Call it create-<formatname>.html.tmpl, and in it, add widgets for each piece of information you'd like collected - such as a build number, or set of steps to reproduce. Then, create a template like custom/bug/create/comment.txt.tmpl, and call it comment-<formatname>.txt.tmpl. This template should reference the form fields you have created using the syntax [% form.<fieldname> %]. When a bug report is submitted, the initial comment attached to the bug report will be formatted according to the layout of this template. For example, if your custom enter_bug template had a field <input type="text" name="buildid" size="30"> and then your comment.txt.tmpl had BuildID: [% form.buildid %] then something like BuildID: 20020303 would appear in the initial comment.

Bugzilla honours the user's Accept: HTTP header. You can install templates in other languages, and Bugzilla will pick the most appropriate according to a priority order defined by you. Many language templates can be obtained from . Instructions for submitting new languages are also available from that location. After untarring the localizations (or creating your own) in the BUGZILLA_ROOT/template directory, you must update the languages parameter to contain any localizations you'd like to permit. You may also wish to set the defaultlanguage parameter to something other than en if you don't want Engish to be the default language.

Template hooks are a way for extensions to Bugzilla to insert code into the standard Bugzilla templates without modifying the template files themselves. The hooks mechanism defines a consistent API for extending the standard templates in a way that cleanly separates standard code from extension code. Hooks reduce merge conflicts and make it easier to write extensions that work across multiple versions of Bugzilla, making upgrading a Bugzilla installation with installed extensions easier. A template hook is just a named place in a standard template file where extension template files for that hook get processed. Each hook has a corresponding directory in the Bugzilla directory tree. Hooking an extension template to a hook is as simple as putting the extension file into the hook's directory. When Bugzilla processes the standard template and reaches the hook, it will process all extension templates in the hook's directory. The hooks themselves can be added into any standard template upon request by extension authors. To use hooks to extend a Bugzilla template, first make sure there is a hook at the appropriate place within the template you want to extend. Hooks appear in the standard Bugzilla templates as a single directive in the format [% Hook.process("name") %], where name is the unique (within that template) name of the hook. If you aren't sure which template you want to extend or just want to browse the available hooks, either use your favorite multi-file search tool (e.g. grep) to search the standard templates for occurrences of Hook.process or browse the directory tree in BUGZILLA_ROOT/template/en/extension/hook/, which contains a directory for each hook in the following location: BUGZILLA_ROOT/template/en/extension/hook/PATH_TO_STANDARD_TEMPLATE/STANDARD_TEMPLATE_NAME/HOOK_NAME/ If there is no hook at the appropriate place within the Bugzilla template you want to extend, file a bug requesting one, specifying: the template for which you are requesting a hook; where in the template you would like the hook to be placed (line number/position for latest version of template in CVS or description of location); the purpose of the hook; a link to information about your extension, if any. The Bugzilla reviewers will promptly review each hook request, name the hook, add it to the template, check the new version of the template into CVS, and create the corresponding directory in BUGZILLA_ROOT/template/en/extension/hook/. You may optionally attach a patch to the bug which implements the hook and check it in yourself after receiving approval from a Bugzilla reviewer. The developers may suggest changes to the location of the hook based on their analysis of your needs or so the hook can satisfy the needs of multiple extensions, but the process of getting hooks approved and checked in is not as stringent as the process for general changes to Bugzilla, and any extension, whether released or still in development, can have hooks added to meet their needs. After making sure the hook you need exists (or getting it added if not), add your extension template to the directory within the Bugzilla directory tree corresponding to the hook. That's it! Now, when the standard template containing the hook is processed, your extension template will be processed at the point where the hook appears. For example, let's say you have an extension named Projman that adds project management capabilities to Bugzilla. Projman has an administration interface edit-projects.cgi, and you want to add a link to it into the navigation bar at the bottom of every Bugzilla page for those users who are authorized to administer projects. The navigation bar is generated by the template file useful-links.html.tmpl, which is located in the global/ subdirectory on the standard Bugzilla template path BUGZILLA_ROOT/template/en/default/. Looking in useful-links.html.tmpl, you find the following hook at the end of the list of standard Bugzilla administration links: keywords' IF user.groups.editkeywords %] [% Hook.process("edit") %] ...]]> The corresponding directory for this hook is BUGZILLA_ROOT/template/en/extension/hook/global/useful-links.html.tmpl/edit/. You put a template named projman-edit-projects.html.tmpl into that directory with the following content: projects' IF user.groups.projman_admins %]]]> Voila! The link now appears after the other administration links in the navigation bar for users in the projman_admins group. Notes: You may want to prefix your extension template names with the name of your extension, e.g. projman-foo.html.tmpl, so they do not conflict with the names of templates installed by other extensions. If your extension includes entirely new templates in addition to extensions of standard templates, it should install those new templates into an extension-specific subdirectory of the BUGZILLA_ROOT/template/en/extension/ directory. The extension/ directory, like the default/ and custom/ directories, is part of the template search path, so putting templates there enables them to be found by the template processor. The template processor looks for templates first in the custom/ directory (i.e. templates added by the specific installation), then in the extension/ directory (i.e. templates added by extensions), and finally in the default/ directory (i.e. the standard Bugzilla templates). Thus extension templates can override standard templates, but installation-specific templates override both. Note that overriding standard templates with extension templates gives you great power but also makes upgrading an installation harder. As with custom templates, we recommend using this functionality sparingly and only when absolutely necessary. Installation customizers can also take advantage of hooks when adding code to a Bugzilla template. To do so, create directories in BUGZILLA_ROOT/template/en/custom/hook/ equivalent to the directories in BUGZILLA_ROOT/template/en/extension/hook/ for the hooks you want to use, then place your customization templates into those directories. Obviously this method of customizing Bugzilla only lets you add code to the standard templates; you cannot change the existing code. Nevertheless, for those customizations that only add code, this method can reduce conflicts when merging changes, making upgrading your customized Bugzilla installation easier.

This feature should be considered experimental; the Bugzilla code you will be changing is not stable, and could change or move between versions. Be aware that if you make modifications as outlined here, you may have to re-make them or port them if Bugzilla changes internally between versions, and you upgrade. Companies often have rules about which employees, or classes of employees, are allowed to change certain things in the bug system. For example, only the bug's designated QA Contact may be allowed to VERIFY the bug. Bugzilla has been designed to make it easy for you to write your own custom rules to define who is allowed to make what sorts of value transition. For maximum flexibility, customizing this means editing Bugzilla's Perl code. This gives the administrator complete control over exactly who is allowed to do what. The relevant function is called CheckCanChangeField(), and is found in process_bug.cgi in your Bugzilla directory. If you open that file and search for "sub CheckCanChangeField", you'll find it. This function has been carefully commented to allow you to see exactly how it works, and give you an idea of how to make changes to it. Certain marked sections should not be changed - these are the "plumbing" which makes the rest of the function work. In between those sections, you'll find snippets of code like: # Allow the owner to change anything. if ($ownerid eq $whoid) { return 1; } It's fairly obvious what this piece of code does. So, how does one go about changing this function? Well, simple changes can be made just be removing pieces - for example, if you wanted to prevent any user adding a comment to a bug, just remove the lines marked "Allow anyone to change comments." And if you want the reporter to have no special rights on bugs they have filed, just remove the entire section which refers to him. More complex customizations are not much harder. Basically, you add a check in the right place in the function, i.e. after all the variables you are using have been set up. So, don't look at $ownerid before $ownerid has been obtained from the database. You can either add a positive check, which returns 1 (allow) if certain conditions are true, or a negative check, which returns 0 (deny.) E.g.: if ($field eq "qacontact") { if (Bugzilla->user->groups("quality_assurance")) { return 1; } else { return 0; } } This says that only users in the group "quality_assurance" can change the QA Contact field of a bug. Getting more weird: user->email =~ /.*\@example\.com$/)) { if ($oldvalue eq "P1") { return 1; } else { return 0; } }]]> This says that if the user is trying to change the priority field, and their email address is @example.com, they can only do so if the old value of the field was "P1". Not very useful, but illustrative. For a list of possible field names, look in data/versioncache for the list called @::log_columns. If you need help writing custom rules for your organization, ask in the newsgroup.

Bugzilla optimizes database lookups by storing all relatively static information in the versioncache file, located in the data/ subdirectory under your installation directory. If you make a change to the structural data in your database (the versions table for example), or to the constants encoded in defparams.pl, you will need to remove the cached content from the data directory (by doing a rm data/versioncache ), or your changes won't show up. versioncache gets automatically regenerated whenever it's more than an hour old, so Bugzilla will eventually notice your changes by itself, but generally you want it to notice right away, so that you can test things.

This information comes straight from my life. I was forced to learn how Bugzilla organizes database because of nitpicky requests from users for tiny changes in wording, rather than having people re-educate themselves or figure out how to work our procedures around the tool. It sucks, but it can and will happen to you, so learn how the schema works and deal with it when it comes. So, here you are with your brand-new installation of Bugzilla. You've got MySQL set up, Apache working right, Perl DBI and DBD talking to the database flawlessly. Maybe you've even entered a few test bugs to make sure email's working; people seem to be notified of new bugs and changes, and you can enter and edit bugs to your heart's content. Perhaps you've gone through the trouble of setting up a gateway for people to submit bugs to your database via email, have had a few people test it, and received rave reviews from your beta testers. What's the next thing you do? Outline a training strategy for your development team, of course, and bring them up to speed on the new tool you've labored over for hours. Your first training session starts off very well! You have a captive audience which seems enraptured by the efficiency embodied in this thing called "Bugzilla". You are caught up describing the nifty features, how people can save favorite queries in the database, set them up as headers and footers on their pages, customize their layouts, generate reports, track status with greater efficiency than ever before, leap tall buildings with a single bound and rescue Jane from the clutches of Certain Death! But Certain Death speaks up -- a tiny voice, from the dark corners of the conference room. "I have a concern," the voice hisses from the darkness, "about the use of the word 'verified'." The room, previously filled with happy chatter, lapses into reverential silence as Certain Death (better known as the Vice President of Software Engineering) continues. "You see, for two years we've used the word 'verified' to indicate that a developer or quality assurance engineer has confirmed that, in fact, a bug is valid. I don't want to lose two years of training to a new software product. You need to change the bug status of 'verified' to 'approved' as soon as possible. To avoid confusion, of course." Oh no! Terror strikes your heart, as you find yourself mumbling "yes, yes, I don't think that would be a problem," You review the changes with Certain Death, and continue to jabber on, "no, it's not too big a change. I mean, we have the source code, right? You know, 'Use the Source, Luke' and all that... no problem," All the while you quiver inside like a beached jellyfish bubbling, burbling, and boiling on a hot Jamaican sand dune... Thus begins your adventure into the heart of Bugzilla. You've been forced to learn about non-portable enum() fields, varchar columns, and tinyint definitions. The Adventure Awaits You!

If you were like me, at this point you're totally clueless about the internals of MySQL, and if it weren't for this executive order from the Vice President you couldn't care less about the difference between a bigint and a tinyint entry in MySQL. I recommend you refer to the MySQL documentation . Below are the basics you need to know about the Bugzilla database. Check the chart above for more details. To connect to your database: bash# mysql -u root If this works without asking you for a password, shame on you ! You should have locked your security down like the installation instructions told you to. You can find details on locking down your database in the Bugzilla FAQ in this directory (under "Security"), or more robust security generalities in the MySQL searchable documentation. You should now be at a prompt that looks like this: mysql> At the prompt, if bugs is the name you chose in the localconfig file for your Bugzilla database, type: mysql use bugs;

Imagine your MySQL database as a series of spreadsheets, and you won't be too far off. If you use this command: mysql> show tables from bugs; you'll be able to see the names of all the spreadsheets (tables) in your database. From the command issued above, ou should have some output that looks like this: +-------------------+ | Tables in bugs | +-------------------+ | attachments | | bugs | | bugs_activity | | cc | | components | | dependencies | | fielddefs | | groups | | keyworddefs | | keywords | | logincookies | | longdescs | | milestones | | namedqueries | | products | | profiles | | profiles_activity | | tokens | | versions | | votes | | watch | +-------------------+ Here's an overview of what each table does. Most columns in each table have descriptive names that make it fairly trivial to figure out their jobs. attachments: This table stores all attachments to bugs. It tends to be your largest table, yet also generally has the fewest entries because file attachments are so (relatively) large. bugs: This is the core of your system. The bugs table stores most of the current information about a bug, with the exception of the info stored in the other tables. bugs_activity: This stores information regarding what changes are made to bugs when -- a history file. cc: This tiny table simply stores all the CC information for any bug which has any entries in the CC field of the bug. Note that, like most other tables in Bugzilla, it does not refer to users by their user names, but by their unique userid, stored as a primary key in the profiles table. components: This stores the programs and components (or products and components, in newer Bugzilla parlance) for Bugzilla. Curiously, the "program" (product) field is the full name of the product, rather than some other unique identifier, like bug_id and user_id are elsewhere in the database. dependencies: Stores data about those cool dependency trees. fielddefs: A nifty table that defines other tables. For instance, when you submit a form that changes the value of "AssignedTo" this table allows translation to the actual field name "assigned_to" for entry into MySQL. groups: defines bitmasks for groups. A bitmask is a number that can uniquely identify group memberships. For instance, say the group that is allowed to tweak parameters is assigned a value of "1", the group that is allowed to edit users is assigned a "2", and the group that is allowed to create new groups is assigned the bitmask of "4". By uniquely combining the group bitmasks (much like the chmod command in UNIX,) you can identify a user is allowed to tweak parameters and create groups, but not edit users, by giving him a bitmask of "5", or a user allowed to edit users and create groups, but not tweak parameters, by giving him a bitmask of "6" Simple, huh? If this makes no sense to you, try this at the mysql prompt: mysql> select * from groups; You'll see the list, it makes much more sense that way. keyworddefs: Definitions of keywords to be used keywords: Unlike what you'd think, this table holds which keywords are associated with which bug id's. logincookies: This stores every login cookie ever assigned to you for every machine you've ever logged into Bugzilla from. Curiously, it never does any housecleaning -- I see cookies in this file I've not used for months. However, since Bugzilla never expires your cookie (for convenience' sake), it makes sense. longdescs: The meat of bugzilla -- here is where all user comments are stored! You've only got 2^24 bytes per comment (it's a mediumtext field), so speak sparingly -- that's only the amount of space the Old Testament from the Bible would take (uncompressed, 16 megabytes). Each comment is keyed to the bug_id to which it's attached, so the order is necessarily chronological, for comments are played back in the order in which they are received. milestones: Interesting that milestones are associated with a specific product in this table, but Bugzilla does not yet support differing milestones by product through the standard configuration interfaces. namedqueries: This is where everybody stores their "custom queries". Very cool feature; it beats the tar out of having to bookmark each cool query you construct. products: What products you have, whether new bug entries are allowed for the product, what milestone you're working toward on that product, votes, etc. It will be nice when the components table supports these same features, so you could close a particular component for bug entry without having to close an entire product... profiles: Ahh, so you were wondering where your precious user information was stored? Here it is! With the passwords in plain text for all to see! (but sshh... don't tell your users!) profiles_activity: Need to know who did what when to who's profile? This'll tell you, it's a pretty complete history. versions: Version information for every product votes: Who voted for what when watch: Who (according to userid) is watching who's bugs (according to their userid). === THE DETAILS === Ahh, so you're wondering just what to do with the information above? At the mysql prompt, you can view any information about the columns in a table with this command (where "table" is the name of the table you wish to view): mysql> show columns from table; You can also view all the data in a table with this command: mysql> select * from table; -- note: this is a very bad idea to do on, for instance, the "bugs" table if you have 50,000 bugs. You'll be sitting there a while until you ctrl-c or 50,000 bugs play across your screen. You can limit the display from above a little with the command, where "column" is the name of the column for which you wish to restrict information: mysql> select * from table where (column = "some info"); -- or the reverse of this mysql> select * from table where (column != "some info"); Let's take our example from the introduction, and assume you need to change the word "verified" to "approved" in the resolution field. We know from the above information that the resolution is likely to be stored in the "bugs" table. Note we'll need to change a little perl code as well as this database change, but I won't plunge into that in this document. Let's verify the information is stored in the "bugs" table: mysql> show columns from bugs (exceedingly long output truncated here) | bug_status| enum('UNCONFIRMED','NEW','ASSIGNED','REOPENED','RESOLVED','VERIFIED','CLOSED')||MUL | UNCONFIRMED|| Sorry about that long line. We see from this that the "bug status" column is an "enum field", which is a MySQL peculiarity where a string type field can only have certain types of entries. While I think this is very cool, it's not standard SQL. Anyway, we need to add the possible enum field entry 'APPROVED' by altering the "bugs" table. mysql> ALTER table bugs CHANGE bug_status bug_status -> enum("UNCONFIRMED", "NEW", "ASSIGNED", "REOPENED", "RESOLVED", -> "VERIFIED", "APPROVED", "CLOSED") not null; (note we can take three lines or more -- whatever you put in before the semicolon is evaluated as a single expression) Now if you do this: mysql> show columns from bugs; you'll see that the bug_status field has an extra "APPROVED" enum that's available! Cool thing, too, is that this is reflected on your query page as well -- you can query by the new status. But how's it fit into the existing scheme of things? Looks like you need to go back and look for instances of the word "verified" in the perl code for Bugzilla -- wherever you find "verified", change it to "approved" and you're in business (make sure that's a case-insensitive search). Although you can query by the enum field, you can't give something a status of "APPROVED" until you make the perl changes. Note that this change I mentioned can also be done by editing checksetup.pl, which automates a lot of this. But you need to know this stuff anyway, right?

&integration;