Does anyone know of a TextArea component with autocomplete? I understand Flextras.com's Autocomplete can be reskinned as a TextArea but wouldn't know where to begin.
If you just wanted to auto complete on each word, similar to (Ctrl-Space) in most IDEs, then you could try doing it yourself.
I would imagine that you would catch the change event on the text area then do some comparison (something like startsWith) of the characters preceding the carat against some pre-populated collection of words.
Flextras poses a good question, in how do you display suggestions. One idea would be to use a custom context menu, discussed here, containing your suggestions. There's also a decision to be made on when to display it. Do you display it constantly as the user types, or maybe when they use a key combination (Ctrl-Space).
It's an interesting challenge.
Related
I have in my application a larger textbox and offer user the possibility to enter text on several rows like this:
first piece of text
second piece of text
and so on...
Each row is a distinct piece of data that has a significance. The thing is I would like to use ajax auto complete extender functionality after the user types a certain character on a row, such as #, and filter the records in the database according to the first letter the user typed after #, moment when the auto complete suggestion list would appear. For example, third piece of text #Action1, means that after the user typed A he would be prompted with a list of values to choose from.
There are several items of concern for me: first, it is possible to use autocomplete extender only for part of the text in the textbox, and second does it work like this for multiple rows (i.e. one time - in each line behaviour)? Regarding its position, I guess I have no other choice than at the bottom of the textbox it refers to, despite I would have liked it to appear below the # symbol.
Thank you very much.
Update: I found something similar to what I am looking for here: Twitter-style autocomplete in textarea, but the author only explained briefly his solution. Any help much appreciated, thx!
Each row is a distinct piece of data that has a significance.
Why are you using a textarea then, and not multiple <input> fields?
Sounds to me like your error in UI design lies right there already, so the rest of my answer is rather theoretical; I probably won’t go into more details as the sensible fix for your problem is the one above.
first, it is possible to use autocomplete extender only for part of the text in the textbox,
Why shouldn’t it be? Cursor position in a textarea for example is readable (although it requires some working around cross-browser issues).
and second does it work like this for multiple rows (i.e. one time - in each line behaviour)?
So long as the line breaks are “hard” ones (made by the user themselves, using enter/return), splitting the actual textarea content by "\n" to have each row as a single value is no problem.
Regarding its position, I guess I have no other choice than at the bottom of the textbox it refers to, despite I would have liked it to appear below the # symbol.
You could try to roughly measure the #’s position, by line and col number it is on, and match that to the character width and line height, when using a monospace font. For other fonts, some more “magic” might be required to measure the actual width of the previous text before the # character.
I have created a Meteor package for this, which allows both free text and multiple autocomplete sources. Meteor's data model allows for fast multi-rule searching with custom rendered lists. If you're not using Meteor for your web app, (I believe) you unfortunately won't find anything this awesome for autocompletion.
https://github.com/mizzao/meteor-autocomplete
See the link for pictures of how it works. Fork, pull, and improve!
I have a database, which tables should viewed in a widget. Seems simple, but I can't decide what to du or use.
Each row of the table should be viewed as one list view item, for instance, imagine table with this fields: id, title, content, date, number.
I need to view it as a list view (not hard-coded, it may also be another thing, if it is possible or better), and the label of the list view item should be the title field. But when the user clicks or double clicks on that item it should open all the contents of the current row in a separate widget. The all of these can be implemented easily by me, but I can't understand what to use: QListView with its model, or QListWidget? Or maybe QSqlTableModel? The last one is unfamiliar to me, I can read about from documentation, but if you have heard or met some kind things/applications, please provide a better solution for described problem.
Hope I could explain my problem correctly,
Thanks in advance.
Have you read about model/view programming in Qt? Basically you should use some model (QSqlTableModel, QSqlQueryModel, QSqlRelationalTableModel or create your own) an then attach it to QListView or QListWidget.
I come from the Delphi school. I was really used and pleased with Delphi grids with the plus features of InfoPower. Not just me, my users loves all features that include grids. Some motives for that are:
With grids they have the ability to insert detail information using only the keyboard. Input, Tab, Input, Tab, Input Tab (automatically insert new row), input, tab, etc.
They were able to simply click in a row and edit it´s data.
They could navigate between records using only arrow keys.
In my flex apps I´ve being working arround this quite often, sometimes using a popup to insert data, sometimes using repeater instead of grid. But now I really really need some intuitive, fast way for my users to insert this data.
I am trying to use DataGrid (or AdvancedDataGrid) with inline editors and I am living a real hell to do that. For the last 3 days I am struggling, tweaking and workarrounding in order to try to mimic the features that I had with Delphi.
Main problems are:
DateField simply don´t work as an inline editor. It gets focus but never looses, only with pressing ESC, wich makes me loose the inputed data.
AdvancedAutoComplete: I have a component descendant from AdvancedAutoComplete with a custom browser. The custom browser is a popup, when I try to PopupManager.remove flex raises an exception of a null pointer inside it´s focus managing engines.
I couldn´t get to insert a new line focused in it´s first column when TABing in the last column of last row.
Having all these problems and not getting substantial material to solve them makes me think the (now obvious) fact that it is just not the way grids are intended to work in flex. My question is: what are the patterns used in flex applications to make sequencial, intuitive, fast, keyboard only insertions in a detail?
To answer your question bluntly, there are no specific established patterns to make editing or inserting data in the DataGrid fast and easy. Not being able to Tab between fields is one of the things I really wish Adobe had added to the DataGrid by default.
In order to get the functionality you want, you'll have to do some custom coding and extend/override the DataGrid's default functionality. Switch On The Code has an example up on tabbing between fields, but you would have to build upon their sample a bit further to include features like adding a new row when pressing tab in the last field of the last row in the DataGrid.
You may also want to consider Flexicious. Based on the Flexicious Ultimate demo, it looks like they have the functionality you're looking for (and then some!).
All,
As part of requirements for a new feature of "locking" a page, it is desired to have all elements on a locked page be non-selectable. This is not to be confused with disabled. All elements should appear as if the page were active, but not be selectable.
The current thought is to create a clear canvas and place it over the existing elements. With this thought, I have two questions:
if you can think of a better way to make all items non-selectable than applying a clear canvas element over the existing elements could you describe it?
if not, what is the best way to retrofit existing implementations to accept the overlayed canvas item? BTW, all .mxml pages inherit from a custom .as file.
Sorry if this is not very descriptive, however, I am new to Flex and have spent many days trying to figure this out.
Thanks,
Todd
You could also set the mouseChildren property of the page to false so the elements will not receive any mouseEvents
All,
To fully "lock" a screen from user manipulation, one must combine Chris Bos's and www.Flextras.com's answers: disable mouse input (mouseChildren) and disable keyboard focus (focusEnabled).
Todd
Would the focusEnabled property work for you?
Documentation says it only relates to "Tabbing", but my memory says it relates to all sorts of selection.
I've got a QTableView for which I want to display the last column always in edit mode. (It's a QComboBox where the user should be able to always change the value.)
I think I've seen the solution in the Qt documentation, but I can't find it anymore. Is there a simple way of doing it?
I think I could archive this effect by using openPersistentEditor() for every cell, but I'm looking for a better way. (Like specifying it only one time for the whole column.)
One way to get the automatic editing behaviour is to call the view's setEditTriggers() function with the QAbstractItemView::AllEditTriggers value.
To display the contents of a given column in a certain way, take a look at QAbstractItemView::setItemDelegateForColumn(). This will let you specify a custom delegate just for those items that need it. However, it won't automatically create an editor widget for each of them (there could in principle be thousands of them), but you could use the delegate to render each item in a way that makes it look like an editor widget.
There are two possibilities:
Using setIndexWidget, but Trolltech writes:
This function should only be used to
display static content within the
visible area corresponding to an item
of data. If you want to display custom
dynamic content or implement a custom
editor widget, subclass QItemDelegate
instead.
(And it breaks the Model/View pattern…)
Or using a delegate's paint method. But here you have to implement everything like enabled/disabled elements yourself.
The QAbstractItemModel::flags virtual function is called to test if an item is editable (see Qt::ItemIsEditable). Take a look at Making the Model Editable in the Model/View Programming documentation.
I can't see an easy way to do this, but you might be able to manage by using a delegate. I honestly don't know exactly how it would work, but you should be able to get something working if you try hard enough. If you get a proper delegate, you should be able to set it on a whole view, one cell of a view, or just a column or row.