We have so many, often useful, prefs, that aren't (and can't?) in the Pref UI. We could add an "Advanced Prefs" dialog, which is basically an editor for for the prefs.js file (on a lower level than the current Pref UI), that can insert any possible key/value pair, together with documentation. The documentation could be in XML or a database and contains all possible keys with description, value type, default value and category. The latter inserts it into a hierarchy, so that "mail.do_glyph_substitution" appears as "Mail/Display/Text Conversion/Glyph Substitution". The dialog could have a structure like the current Pref UI dialog: The Hierarchy is displayed on the left, right top the name as header and the description and right bottom the widgets, depending on the value type, and action buttons ("OK", "Cancel" and "Reset to default").
The hierarchy could be replaced by tabs containing both the hierarchy and an alphabetical list (of the names) of the options. This would simply references to these prefs, e.g. in the "Knowledge Base".
As both the UI and the datasource are in XML, the plan is to realize that using XSLT. The XUL for both the hierarchy and the "content" on the right are created on-the-fly, the first on opening the dialog and the latter on a click on an item in the hierarchy. That way, the implementation is just some XSL, XUL and a bit JS. The problem is, that the XSLT processor in Mozilla 1. seems to be unusable yet (not XP) and 2. doesn't support parameters, which we need for the creation of the right "content".
spam: in my testing realm, so reassigning qa contact to me, en masse.
Bulk move of all Pref UI component bugs to new Preferences component. Pref UI component will be deleted.
Component: Pref UI → Preferences
Status: ASSIGNED → RESOLVED
Last Resolved: 18 years ago
Resolution: --- → LATER
with the the Future milestone, i'm not sure what to do with Resolved Later bugs... reopen, and set the milestone to Future? that way it won't get ignored (which often happens to bugs which are resolved or verified). also adding helpwanted kw, if there are any [external or otherwise] takers out there. :-)
Status: RESOLVED → REOPENED
Resolution: LATER → ---
Target Milestone: --- → Future
I am not sure, this bug is still the right way to go. My current preference is the Deep Prefs Hierarchy(TM), see discussion in .prefs. This bug might be good for a "TweakUI"-like power-user package. In short: better than hand-editing prefs.js, but 2 different prefs dialogs are hardly good UI (opinion of Ben Bucksch at 2000-07-29 :) ).
Attempting a better summary than "advanced prefs".
Summary: Advanced Prefs → [RFE] Advanced users: Prefs Tree UI, edit all prefs w/ tree
Readding old name of this bug, thus keeping longer summary.
Summary: [RFE] Advanced users: Prefs Tree UI, edit all prefs w/ tree → Advanced Prefs - Edit *all* prefs using tree UI
*** Bug 64492 has been marked as a duplicate of this bug. ***
Changing personal priorities. Giving away most of my bugs :-( (reassigning to default owner). I will still track these bugs closely. If you need my input, feel free to ask me. New owner: Please do *not* close these bugs (as WONTFIX or whatever you may find) unless they are fixed. Rather, reassign to <firstname.lastname@example.org>, if you don't want to work on them.
Assignee: mozilla → matt
Status: REOPENED → NEW
Target Milestone: Future → ---
OT: I am very interested in seeing this bug implemented. It makes me wish i knew how to program. How about adding this to: prefs - advanced - miscellaneous?
This bug shou not be limited to just prefs.js settings; but to other settings as well. One setting I would really like to see implemented is to be able to turn off the dreaded cascading of new windows. I place my window EXACTLY where I want it, and want all new windows to appear there too.
Please also see (and CC and vote for) bug 68134 >>Need list of "tweakable settings" in advanced preferences<< for some good comments/arguments why this feature is important and how it can be implemented without confusing the "PC challenged" user (i.e. prefs - advanced - miscellaneous, should be clear and hidden enough to indicate to EVERYONE that these are non-required settings).
*** Bug 67996 has been marked as a duplicate of this bug. ***
Sorry, but that didn't insert my comment: For Boris: On the first place, thanks Boris. And if the implementation for bug 17199 is done right, then it might be helpful. I will make this a dup of that other bug, and i will add my concerns to that bug, which are backed up by a recent user poll by then. I only hope to protect my kids in a way. Also that bug is rather old by now. And belief me, this is all about marketing and pushing mozilla forward with an competitive edge. I'm not looking to make mozilla bloated, because that's out of the question. Powerful, Flexible and Customizable that's more like it. I'm also not looking for tweaks, but i do like to know more about pref. setting in general. But that's a completely different story. For Matthew: Matt, my comment was short, but that's just because i run into a long ammonia. For now my computer time is limited, and you know, we all have our bad days. Just bear with me Ok? Most friendly, HJ.
Oh lord, what a begin, sorry for that but this had to be inserted for bug 67799.
*** Bug 67996 has been marked as a duplicate of this bug. ***
> One setting I would really like to see implemented is to be > able to turn off the dreaded cascading of new windows. I place > my window EXACTLY where I want it, and want all new windows to > appear there too. I would imagine this would be a prefs.js setting if implemented.
*** Bug 68134 has been marked as a duplicate of this bug. ***
Bug 46029 deals with a GUI for choosing a different UA string as 'seen' by webservers (e.g. turn us into IE5.5 according to the webserver). Ben B pointed us at this which I think is a better choice. CC self. What is the status of this? Are the technologies ready for use in Mozilla? If so, is it just a matter of discussing possible designs and implementing the best?
> What is the status of this? Untriaged. I bet, Netscape won't implement it. So, whoever is interested, take it (but double-check with current owner, to be sure). > Are the technologies ready for use in Mozilla? As for XSLT, I don't know. Might be. But there are other ways to implement it, without XSLT. > If so, is it just a matter of discussing possible designs and implementing > the best? Yes.
See also bug 74887. Gerv
Due to the high demand (see duplicates and their CC lists) the keywords: *nsCatFood*, and *MostFreq* should definetely be added. How about making the prefs appear in a seletion list. Below the selection list would be the data/variable selection area that changes according to what pref is selected. +-Miscellaneous Prfs-----------+ | | | Pref 1 | | Pref 2 ---s-e-l-e-c-t-e-d--- | | Pref 3 | | Pref 4 | | | +------------------------------+ | Brief Description of Pref 2 | | | | <x> Yes | | < > No | +------------------------------+
maybe buttons for: - "Reset *selected pref* back to Default value" and for - "Reset *all Miscellaneous prefs* back to default values" Shorter names for the buttons would be needed, of course.
or just have the default value in brackets behind each pref: +- Miscellaneous Prefs -----------+ | /\ | | Pref 1 |x|| | Pref 2 | || | Pref 3 ---s-e-l-e-c-t-e-d--- | || | Pref 4 |_|| | Pref 5 \/ | +---------------------------------+ +-----------------------------+ | Brief Description of Pref 3 | | Brief Description of Pref 2 | | | | | | [ 12 ] (Default = 8) | or | <x> Yes | | | | < > No (Default) | | | | | +---------------------------------+ +-----------------------------+
*** Bug 74887 has been marked as a duplicate of this bug. ***
Peter: what you propose is unworkable because it requires the UI to know more about each pref than the pref system does. We basically have the following information: - pref name - pref type (string, bool, int) - current pref value (if any) We also need the system to work with arbitrary numbers of preferences, and automatically incorporate new ones as they are created. There are far too many of them for a drop-down box. We either go with Ben's UI, and have a tree, or go with mine, and have a text box. My idea has the advantage of not requiring XSLT, and probably being simpler to implement (get a PrefsService, query for the given prefname string, update if it's found, suitable errors if not.) His has the advantage of being more user-friendly. On the other hand, given a) advanced users can cope with my UI (it's just a shortcut for them) b) non-advanced users can probably cope with an email/web page telling them "stick this value in here and that value in there and press OK and it'll do what you want". If we do mine, we can always do the other later. Unless anyone wants to step up to the plate for doing the tree view now... This bug is not mostfreq - it does not have 10 dupes. I very much doubt it's a Netscape priority either, and certainly not nscatfood. :-( Gerv
When prefs are created, they must exist (i.e. be documented) somewhere. All we need to do is add certain mandatory datafields (database) to that documentation (i.e. pref description in common terms, default value, pref type, current value). Then the prefs ui could retrieve those values and display them in the list i proposed. This would deal with an arbitrary number of prefs and automatically incorporate new ones as they are created since they would be retrieved from the datasource on startup or on prefs UI loading. Making users (especially novices) find some newsgroup and edit system files (prefs.js) is definetely always a very bad idea - let's forget that "solution" pronto. Having the number of dupes be the criteria for a *mostfreq* bug seems silly to me. It merely reflects the number of people who are too clueless to see if their bug has already been reported. Wouldn't it be better to base *mostfreq* on the number of persons in the CC list? I think so :)
> what you propose is unworkable because it requires the UI to know more > about each pref than the pref system does. I agree wtih Peter L. here: The intention was to build a (XML-based) database of available prefs and their values. Allowing a user to set 6 as value, while the pref is an enum and only 0-3 are defined, is not too useful. That is, when you wan to provide a comfortable UI. > We either go with Ben's UI, and have a tree, or go > with mine, and have a text box. Or both, one after the other. Please do not close this bug, after you implemented your short-term proposal in bug 74887. For the record: While I definitely prefer my inital suggestion of a tree with a comfortable UI, I have no (strong) objection, should somebody decide to implement Gerv's suggestion first. Peter L., the mostfreq keyword is mainly to stop too many dup reports, i.e. to help the bug triagers with their work. It is not so much a marker of the importance of a bug. (But it is an indirect one, just as the cc list is.)
> Allowing a user to set 6 as value, while the > pref is an enum and only 0-3 are defined, is not too useful. But going down this route will lead to a world of pain where the prefs database gets out of sync with reality, and it either lets you set invalid values or, worse, won't let you set valid ones. We must be getting on for 100 prefs, with more added every week. As you will note, I've closed the other bug, not this one. I have no intention of closing this one :-) Gerv
Gerv, you are right. But the current situation is a problem, and we need a authorative database of prefs anyway (apart from the default prefs). This bug would "encourage" programmers to keep the database current, because they like probably like the comfort of the UI as well, and have an interest themselves to set the values via the UI -> keep the database current.
As a side note, I doubt anyone could make an authoritative list of prefs without trawling the entire source base - AIUI, there's no central registry for them. This is another hurdle to cross :-( Gerv
If the "database" were dynamic (i.e. the prefs UI would retrieve /whatever/ is in the database), then we could gather together prefs gradually, but implement the UI now, with only the first set of available prefs.
Isn't that what all.js is for?
> If the "database" were dynamic (i.e. the prefs UI would retrieve /whatever/ > is in the database) That is the plan, that's what I meant with "on-the-fly" in my Dec 1999 comment. Note that we don't necessarily need XSLT for it, it canbe coded in JS, just that XSLT would be the most elegant or coolest solution. > Isn't that what all.js is for? all.js (et al) only ontains the default prefs. It does not contain (in a computer-readable form) all possible values or a description of them. Actually, for many prefs, there is no description at all currently. Also, it does not contain really all prefs (but almost all). For these missing prefs, we fall back to a hardcoded default value ...or misbehave :). In any case, missing entires in all.js (et al) are a bug. So, yes, creating such an database (for the current Mozilla) is work, but it is doable. And Peter is right, we don't need a complete database to implement or even check in the UI.
->mcafee [default prefs owner].
Assignee: matt → mcafee
mitch, this is the bug i was mentioning earlier... perhaps your webpage/script suggestion would be an alternative to this? feel free to ping me on irc to discuss...
Yes, we can do this with a webpage and some signed scripts (as soon as I get those working again). With a signed script, a webpage can display and modify prefs.
Problem: You want to keep a DB of prefs with all possible values for each pref. Idea: When the browser loads the prefs, if the prefs value is not in the DB then the setting is not valid. This would create a checks and balance system between programmers adding prefs and them also adding the possible values of that pref to the DB. Of course, this would be a startup speed issue, but I think the prefs are being compiled and cached for speed currently, so the speed issue would only be a factor the next time you started the browser. Not valid settings follow the current logic flow when a pref has an invalid value.
Over to vishy to find an owner.
Assignee: mcafee → vishy
In addition to the current value of a pref, we also know the default value (if the pref is set in all.js et al. If this were for super-advanced folks we can leave it up to them to know what they're touching. We'll display the uneditable default value so they can get back to something sane if they screw up. This would be good for a QA audit, too: we could work down the list and find prefs that are in-use (exist, have a value) but which don't have a defined default value.
->samir. am also wondering if bug 98569 would be considered a dup of this...? esp now that about:config has been implemented. [yay!]
Assignee: vishy → sgehani
*** Bug 98569 has been marked as a duplicate of this bug. ***
what is with a regular expression to check the value. this could be implemented as a value in the xml source. to save runtime i prefer to check the values only after a change in the advanced prefs dialog. i think it is not a good idea to handle all.js for default values and a xml file as a database.
Couldn't we implement this as inline editing (ILE) of the about:config tree?
Summary: Advanced Prefs - Edit *all* prefs using tree UI → Advanced Prefs - Edit *all* prefs using tree UI (about:config?)
This bug was supposed to do much more than about:config does. It was planned that the user gets inline documentation about the pref in question, the possible input is restricted by the allowed options for the relevant prefs etc.. Making about:config editable might be a good workarond in the meantime, but I don't know, if it's worth the effort - the full-blown solution shouldn't be hard, now that Transformiix is supposed to work. That's what I wanted to use 2 years (!) ago.
Summary: Advanced Prefs - Edit *all* prefs using tree UI (about:config?) → Advanced Prefs - Edit *all* prefs using tree UI
*** Bug 103963 has been marked as a duplicate of this bug. ***
I don't know whether about:config is in outliner / tree or not, but if so, couldn't we leave the config items displayed as they are, except adding an arrow on the left of each? This, once "opened", would turn the fields editable and add buttons like "OK", "Cancel", "Reset to default" below. Wish I could find a good example screenshot at the moment.
It looks like another way to provide prefs editing is in bug 110090.
See also bug 14969, "[RFE] TweakUI-like Package to attract audiences that like to customize".
*** Bug 147750 has been marked as a duplicate of this bug. ***
I filed an enhancment request about this yesterday (bug 147750), without knowing that it was already there (sorry about that). I just want to say that I'd vote for Peter Lairo's suggestion in adding a description to the database (?) of mozilla preferences, so the UI could also display what a specific setting does. Since adding a description for each setting will take time, space could just be reserved for it, then we would add more and more descriptions of the most common settings as time went by. 'Hixie' Hickson (comment #46) also mentioned the about:config method. This actually lists exactly the features that we are talking about here. By just making that list editable and add an extra column (pref description), we would have the functionality already. Then we could just put that code into a section of the Preferences window.
This would be very similar to the Gruop Policy Editor in Windows 2000/XP. There you have the list of preferences, along with a description and default value. Since we're talking about more or less advanced features, it's not so important that it's *very* easy to use. It could, in my opinion, look very similar to the current about:config list, with the addition of being editable. maybe you could just double click on a true/false setting to toggle it? And string values would display an edit-box when double-clicking (like when you rename a file in Explorer).
I noted that I have some examples for the XML doc on my website, and I don't seem to have referenced them here. <http://www.bucksch.com/1/projects/mozilla/17199> As for the implementation, it's now 2.5 years later and XSLT in Mozilla still can't be used to generate XUL. Additionally, I have no idea if I can pass parameters to the XSLT via the URL. Maybe a hard-coded converter (implemented in JS) of the XML doc to XUL would be the more realistic way. It would read the XML file(s) into a DOM, generate the XUL tree for the categories out of it and when the user clicks on an item in the tree, it looks up the data in the XML DOM and generates the editing UI out of it. Reassign to nobody, to clarify that nobody is working on this.
Assignee: sgehani → nobody
I _might_ be willing to tackle this as "Baby's First Mozilla Project". I am an amateur Borland "Delphi" programmer with no non-DOS or non-Windows experience. With Delphi, I am more than capable of writing an applet that can 1.) Display the contents of all.js, prefs.js, and user.js in a browsable and searchable interface. 2.) Allow the user to make changes to the values. 3.) Write the changes back to either prefs.js or user.js. As to what I can do with programming tools/languages I'm not familiar with? As well, it must be noted that all.js, prefs.js, and user.js are NOT a comprehensive list. Example: if you don't already know about browser.bookmarks.file you are not going to find it in one of those three .js files.
Is this a duplicate of Bug 110090 ?
oh, duh! I just found out about XUL templates and that they might be even more appropriate than XSLT. I feel a bit like an idiot. I'd like to try it out, but unfortunately, I have no time right now.
Assignee: nobody → ben.bucksch
*** Bug 170661 has been marked as a duplicate of this bug. ***
*** Bug 173465 has been marked as a duplicate of this bug. ***
Here is a proposed solution that requires very little in the way of changes to the current prefs system and shouldn't be too hard to implement: Make each pref have its own URL (for example about:config/browser/popups/showPopupBlocker). At this URL have a page that allows the user to set the value of the pref according to its type (bool, string, whatever). It doesn't have to have any extra documentation other than the pref name, because most people won't go browsing the jillions of prefs. Instead this makes it possible to make a *link* to a pref. Now when someone hears about a whiz-bang new Moz feature, all they have to do is click the link given in the Slashdot comment (or whatever) and they get a page that allows them to set it right away. No mucking with text files. I think this will solve the #1 problem this bug is trying to solve - people hear about a cool new moz feature but don't like having to find text files and edit them and restart mozilla and all that. No extra documentation (taking up valuable memory space and download time) is needed since this is not meant to be a way of discovering new prefs. This just lets people have easy access to any pref they personally know about, and be easily told about any pref in Mozilla. If you *really* can't live without the ability to browse the prefs with documentation, that could fit in nicely with this solution too. Just make each pref group a directory (about:config/browser) showing each pref and have documentation on each individual pref page. But I think that the descriptions aren't really necessary; the problem is mostly setting specific prefs that you have heard about, not surfing the prefs to find new cool ones, and besides, the pref names are quite descriptive already by themselves.
I have just posted an XPI to the mozdev site (http://preferential.mozdev.org) that provides this very functionality. Eventually, I plan to add documentation on each bug from an RDF datastore, but for now, the XPI simply allows access to a preference tree that can have entries added, edited and deleted on the fly. The current results are very similar to that proposed by comment 62.
The Preferential Mozilla extension (http://preferential.mozdev.org) is now at version 0.3.1. It has been successfully tested as providing a tree UI for all Mozilla preferences on Windows 98 for both Mozilla 1.2b and Phoenix 0.4. Any feedback on this tool would be welcome.
There's no way to load this pref viewer in the content area, or using URLs, is there? That could be dangerous.
This is sorta covered by the new editable about:config screen now - no?
*sigh*, no, read up.
QA Contact: sairuh → nobody
Is there any hope for this feature request? Perhaps the author of Preferential should be asked if he'd be willing to create a fix for this bug. Anyways, this bug has been inactive a rather long time...
QA Contact: nobody → mozillamonks
Sure, if preferential made it so that it could provide the same functionality through about:config by overriding chrome://global/content/config.xul, I don't see why not as long as preferential were allowed to be included as a default extension. I have some complaints about preferential, though: *Filter doesn't work (filter does in about:config) *Needs a root tree element so you can truly Expand All *Whereas preferential is tree-only (it seems) and about:config is list-only, we'd want a combination of the two If we went this route, preferential should probably supply a slightly different UI for its about:config replacement than it does in a window since standard menus shouldn't be inside a tab. I like the UI for about:config, and this could be expanded upon. Perhaps clicking "Advanced Preferences" could simply always load about:config in a tab.
Summary: Advanced Prefs - Edit *all* prefs using tree UI → Advanced Prefs - Edit *all* prefs using tree UI in about:config
Geez. Read up. This bug is *not* about <about:config>, but a specific GUI idea. It's not "Preferential" either. Resetting summary.
Ben: The discussions in this bug are old and its hard to follow exactly what people are talking about in the beginning because a lot has changed. Also, most these comments were prior to the inclusion of about:config for editing prefs, and the creation of Preferential. Please recap what this bug is about because I think there is a lot of confusion in the last year's worth of comments, and that is what people generally go by because the older discussion is out of date and also there are conflicting statements and it also appears incomplete. Whoever orignally CCed themselves has probably forgot a lot of the original discussion that occurred within newsgroups and IRC. Can you recap what the most popular ideas are and _exactly_ what this bug is about? Is this basically a discussion for a UI that shows the documentation for preferences that are contained within the source code? I remember something to that affect, and people mentioning that the documentation could be within .properties files, etc. If so, why is tree UI mentioned in the summary. Is a tree necessary to show documentation for the preferences? It almost seems like 3 different things are discussed in this bug. Please clarify.
What is the status of this bug as of now? Is it dead? There were no activity since 2004!
Since there was no clarification, what this bug really is about and the functionality to edit all prefs is given by about:config, I'd suggest to mark it wontfix.
Whiteboard: [se-radar] → [se-radar] [CLOSEME INVA/WONT?]
Bruno, as said above, the idea was expanded on the newsgroup a bit, then drowned into a discussion with Matthew Thomas. <http://groups.google.com/group/netscape.public.mozilla.prefs/browse_frm/thread/68dd94274bbcb58e/a76398d6c86a0d97> The idea was: > the prefs should be designed (semantically) well. But the real solution > to the problem is to organize prefs well. > > In the long term, I'd like to organize them in a deep hierarchy, with > the "frontpage" (the panel you get by clicking an a parent node) of a > subtree not being a collection of misc prefs (which didn't well fit in a > subpanel), but the most important ones. That way, a user with limited > needs wouldn't have to care about the deep hierarchy. If a user misses a > pref on the "frontpage", he can explore the hierarchy in more depth. > > (See also bug 17199.) Robert O'Callahan wrote: <http://groups.google.com/group/netscape.public.mozilla.prefs/msg/12e66c73ba72391a> > the maximally usable set of preferences is different for different kinds > of users. Most people can only handle a small set of preferences. > On the other hand, I just love programs that give me scrolling lists > with hundreds of checkboxes. So, I would like to have lots of preferences > that are usually concealed. about:config is a hack, not a solution. 1) There are absolutely no descriptions what the prefs mean. The description is very much needed. 2) It's not discoverable, not browsable. Unless you remember the pref name, it's very hard to find. 3) It only lists prefs that are in the default prefs file. The infamous "browser.dom.window.dump.enabled" for example - I *always* have to look it up, even now, because I can't remember it and it's not in the defaults, because default prefs are claimed to slow things down. That's not usable in my opinion, I am angry every time I have to set this pref. This bug would fix all of them. It would have a database of prefs, including descriptions and allowed values, as described at the top and at the URL. They are also categorized, so you can browse. Each category page summaries the most important prefs for that category, so you naturally get a nice UI, where you don't have to drill down deep for commonly needed prefs, but you can drill down as deep as you need to. This solves the problem that the amount of usable prefs UI, and the number of needed prefs, varies a lot between users and there's no clear cut-off point. The point is different for each user. Given that you can dive deeper, per brach, as much or little as you want, and it's always gentle, this solves the problem.
Whiteboard: [se-radar] [CLOSEME INVA/WONT?] → See comment 74 for description
You need to log in before you can comment on or make changes to this bug.