Reported by mayerch on 2010-12-03 14:53 UTC
Some browsers (esp. those for mobile devices) seem to be unable to scoll sub parts of a page - which means that the ComeVisu pages can't be scrolled.

(cf. e.g. post #147)

Reported by mayerch on 2010-12-03 14:53 UTC
Some browsers (esp. those for mobile devices) seem to be unable to scoll sub parts of a page - which means that the ComeVisu pages can't be scrolled.

(cf. e.g. post #147)

Reported by mayerch on 2010-12-03 14:27 UTC
The colorChooser plugin isn't initialisated by the current values

Reported by netzkind on 2010-11-15 09:23 UTC
When editing two widgets in a row, the edit-window for the second widget shows the GA of the first instead of its own.

Problem is with caching input-fields - they need to be reset on edit.

Reported by mayerch on 2011-03-09 21:52 UTC
Bug for the CometVisu rackage running on the WireGate:

The caching of the web server should be selectively disabled (e.g. for the config file)

Reported by netzkind on 2010-11-15 09:24 UTC
When setting mapping and style, javascript fails due to alleged missing GA.

Reported by mayerch on 2011-04-26 19:32 UTC
Sometimes adding a new element will sometimes create an "undefined" although the proper label was set during editing (cf. #590)

Reported by netzkind on 2010-11-15 09:24 UTC
When setting mapping and style, javascript fails due to alleged missing GA.

Reported by mayerch on 2010-12-03 14:57 UTC
It seems that the slieders aren't working on android and iPhone

(cf. post #172)

Reported by j-n-k on 2011-05-13 12:15 UTC
pre&post are displayed after the button when using the "switch-widget".

proposal: remove pre/post for all widgets.

Reported by netzkind on 2010-11-15 09:23 UTC
When editing two widgets in a row, the edit-window for the second widget shows the GA of the first instead of its own.

Problem is with caching input-fields - they need to be reset on edit.

Reported by mayerch on 2011-01-02 22:34 UTC
The widgets aren't properly initialised, so they might send their initialisation value once they receive them and thus create unnecessary bus traffic.

Reported by makki1 on 2011-02-27 20:19 UTC
eibread/write-cgi bails out with 500 Internal server error if local eibd not running
-> Should give a clear error-response

Reported by j-n-k on 2011-05-15 19:21 UTC
The diagram_popup-widget displays the content-text only but nor the preview-.image if used on sub-pages within the "pure"-design. everything woprks fine for the "discreet"-design.

Reported by makki1 on 2011-02-10 21:17 UTC
the clients should try to reconnect the XHR regularly on network-problems

Reported by makki1 on 2011-02-10 21:17 UTC
the clients should try to reconnect the XHR regularly on network-problems

Reported by j-n-k on 2011-05-15 19:21 UTC
The diagram_popup-widget displays the content-text only but nor the preview-.image if used on sub-pages within the "pure"-design. everything woprks fine for the "discreet"-design.

Reported by netzkind on 2010-11-15 09:26 UTC
If get_addresses.php fails on call (e.g. due to missing /etc/wiregate/eibga.conf et al), the methods for creating the select-entries for GA-selection fail. JS throws an error.

This needs to be catched, and an input-field should be presented to the user if no select-list can be created.

New error occurs in develop: If an "info" element contains no label the visu shows "undefined" instead.
In previous versions it was ok to leave the label empty, please see attached screenshots.

The xml is the same file:

        <info mapping="DoorOpenClose">
            <layout colspan="1"/>
            <label>Tor</label>
            <address transform="DPT:1.001" mode="read">3/2/5</address>
        </info>
        <info mapping="DoorOpenClose">
            <layout colspan="1"/>
            <address transform="DPT:1.001" mode="read">3/2/4</address>
        </info> 

Reported by mayerch on 2011-04-26 19:32 UTC
Sometimes adding a new element will sometimes create an "undefined" although the proper label was set during editing (cf. #590)

Reported by netzkind on 2010-11-15 09:24 UTC
When editing/inserting image-tag, the field text-content is mandatory. This is just plain wrong.

Reported by amindnich on 2011-02-07 20:15 UTC
Rendering issue (see attached screenshot)

1. embedded image
   An embedded image, such as the default weather forecast diagram is
   bigger than one column, therefore interfering with elements on the right
   hand column

2. diagram inline
   The inline diagram is too far left, therefore either missing numbers on the
   y-axis if in the left column, or overlapping across the middle, when on the
   right hand side

Experienced on both Chromium and Firefox 3.6.13 on Ubuntu

(deleted; wrong project...)

Reported by stefanwerner on 2011-02-28 10:37 UTC
The date format of the x-axis in ALL diagrams shows time in GMT instead of local time

Reported by mayerch on 2011-02-12 17:30 UTC
When the clients sends an AJAX request to the backend that contains an empty address (i.e. ...&a=1/2/3&a=&a=...) the server returnes a
500 Internal Server Error

=>

1. The backend should handle that gracefully
2. The client should'n send such a request (or the protocol definition should explicitly allow it)
3. The client shouldn't stop working afterwards...

Assigned to Makki to fix point 1)

Reported by j-n-k on 2011-05-18 18:15 UTC
STR:

1. 'pure' design, flavour different from default
2. use editor
3. save config
4. flavour reset to default

Reported by netzkind on 2010-11-15 09:24 UTC
When editing/inserting image-tag, the field text-content is mandatory. This is just plain wrong.

Reported by makki1 on 2010-12-05 23:50 UTC
Opened (flot) diagrams are not getting refreshed.
tested in FF & Chromium.

Reported by stefanwerner on 2011-02-28 10:37 UTC
The date format of the x-axis in ALL diagrams shows time in GMT instead of local time

Reported by mayerch on 2011-02-20 12:42 UTC
When editing a widget that has quite a few elements and attributes that editor pane / frame can become quite long - so long that a short browser window won't show the OK and Cancel buttons.

=> The editor should be scrollable

Reported by makki1 on 2010-12-05 23:50 UTC
Opened (flot) diagrams are not getting refreshed.
tested in FF & Chromium.

Reported by j-n-k on 2011-05-12 09:43 UTC
on very wide screens the buttons are arranged not in 2x2 grid but three in a row and one below which looks really bad

Reported by mayerch on 2010-12-03 14:11 UTC
The client is currently transmitting the KNX datatype. This is not in the protocol and shouldn't be done as it's highly KNX specific.
(note: this might need a change in the current backend)

Reported by j-n-k on 2011-05-12 09:43 UTC
on very wide screens the buttons are arranged not in 2x2 grid but three in a row and one below which looks really bad

Reported by mayerch on 2010-12-03 14:11 UTC
The client is currently transmitting the KNX datatype. This is not in the protocol and shouldn't be done as it's highly KNX specific.
(note: this might need a change in the current backend)

Reported by mayerch on 2011-03-09 21:52 UTC
Bug for the CometVisu rackage running on the WireGate:

The caching of the web server should be selectively disabled (e.g. for the config file)

Reported by mayerch on 2010-12-03 14:57 UTC
It seems that the slieders aren't working on android and iPhone

(cf. post #172)

Reported by mayerch on 2011-01-02 22:34 UTC
The widgets aren't properly initialised, so they might send their initialisation value once they receive them and thus create unnecessary bus traffic.

Reported by mayerch on 2010-12-03 14:27 UTC
The colorChooser plugin isn't initialisated by the current values

Reported by amindnich on 2011-02-07 20:15 UTC
Rendering issue (see attached screenshot)

1. embedded image
   An embedded image, such as the default weather forecast diagram is
   bigger than one column, therefore interfering with elements on the right
   hand column

2. diagram inline
   The inline diagram is too far left, therefore either missing numbers on the
   y-axis if in the left column, or overlapping across the middle, when on the
   right hand side

Experienced on both Chromium and Firefox 3.6.13 on Ubuntu

Reported by mayerch on 2011-02-20 15:56 UTC
The editor needs to take care of the align attribute of the page widget.

Currently it deletes it even if it was manually set...

Reported by mayerch on 2011-02-20 15:56 UTC
The editor needs to take care of the align attribute of the page widget.

Currently it deletes it even if it was manually set...

Reported by mayerch on 2010-11-20 18:03 UTC
The fallback for unknown widgets (called "unknown") should'n be offered in the editor (i.e. filtered out)

Reported by mayerch on 2011-02-20 12:42 UTC
When editing a widget that has quite a few elements and attributes that editor pane / frame can become quite long - so long that a short browser window won't show the OK and Cancel buttons.

=> The editor should be scrollable

Reported by j-n-k on 2011-05-13 12:15 UTC
pre&post are displayed after the button when using the "switch-widget".

proposal: remove pre/post for all widgets.

Reported by makki1 on 2011-02-27 20:19 UTC
eibread/write-cgi bails out with 500 Internal server error if local eibd not running
-> Should give a clear error-response

Reported by netzkind on 2010-11-15 09:26 UTC
If get_addresses.php fails on call (e.g. due to missing /etc/wiregate/eibga.conf et al), the methods for creating the select-entries for GA-selection fail. JS throws an error.

This needs to be catched, and an input-field should be presented to the user if no select-list can be created.

Reported by j-n-k on 2011-05-18 18:15 UTC
STR:

1. 'pure' design, flavour different from default
2. use editor
3. save config
4. flavour reset to default

Reported by mayerch on 2010-11-20 18:03 UTC
The fallback for unknown widgets (called "unknown") should'n be offered in the editor (i.e. filtered out)

Reported by mayerch on 2011-02-12 17:30 UTC
When the clients sends an AJAX request to the backend that contains an empty address (i.e. ...&a=1/2/3&a=&a=...) the server returnes a
500 Internal Server Error

=>

1. The backend should handle that gracefully
2. The client should'n send such a request (or the protocol definition should explicitly allow it)
3. The client shouldn't stop working afterwards...

Assigned to Makki to fix point 1)