2) Added Dark(Beta) UI Theme option. Fixes#3741.
3) Fix an issue where a black arrow-kind image is displaying at the background of browser tree images. Fixes#4171
Changes include:
1) New theme option in preferences - Miscellaneous -> Themes. You can select the theme from the dropdown.
It also has a preview of the theme just below the dropdown. Note that, a page refresh is needed to apply changes.
On saving, a dialog appears to ask for refresh.
2) You can create your own theme and submit to hackers. README is updated to help you create a theme. Theme will be available only after the bundle.
3) Correction of SASS variables at few places and few other CSS corrections.
4) Added iconfont-webpack-plugin, which will convert all the SVG files(monochrome) used as icons for buttons to font icons.
This will allow us to change the color of the icon by using CSS color property.
5) All the .css files will bundle into a separate file now- pgadmin.style.css. This will help reduce the size of
theme CSS files as CSS in .css files will not change with the change of SASS variables.
Changes include:
1) Remove underscore-string and sprintf-js packages as we were using only %s. Instead, added a function to do the same. Also changed gettext to behave like sprintf directly.
2) backgrid.sizeable.columns was not used anywhere, removed. @babel/polyfill is deprecated, replaced it with core-js.
3) Moved few css to make sure they get minified and bundled.
4) Added Flask-Compress to send static files as compressed gzip. This will reduce network traffic and improve initial load time for pgAdmin.
5) Split few JS files to make code reusable.
6) Lazy load few modules like leaflet, wkx is required only if geometry viewer is opened. snapsvg loaded only when explain plan is executed. This will improve sqleditor initial opening time.
Reviewed By: Khushboo Vashi
Fixes#4701
Look 'n' Feel and implementation logic are inspired from
'http://explain.depsez.com'.
It now creates three tabs under the 'Explain' panel when executing a
query using the Explain Analyze/Explain button from the toolbar of the
Query tool.
Graphical
---------
-> Graphical Explain Plan
Analysis
--------
-> Table to show details of the explain plan analyse.
-> Each row represents the statistics per Explain Plan Node
-> It may contains columns like node information, exclusive timing
(time spent for this explain node excluding the child nodes),
inclusive timing, actual rows, plan rows,
rowsx (misestimation between planned vs actual rows), loop.
-> Background color of exclusive, inclusive, rows changes based on
their values.
i.e.
If Percentage of exclusive, and inclusive timings of total query time
is:
> 90 - Red Color
> 50 - Orange (Between Red & Yellow Color)
> 10 - Yellow color
If planner misestimation for the rows is
> 1000 times - Red Color
> 100 times - Orange (Between Red & Yellow Color)
> 10 times - Yellow Color
Also - if actual rows <= planned rows then it shows up arrow, else it
shows down arrow.
Statistics
----------
-> It contains a HTML table for the statistics per Node Type, and
a HTML table for the statistics per table.
Reviewed by: Akshay Joshi
workaround rendering speed issues in CodeMirror with very large scripts.
Fixes#4631.
Re-arrange editor options in the Preferences dialogue to tidy things up.
Attached is a tiny but very effective patch to improve the speed of
finding a node using path (used internally).
If you right click or just click on a node, internally the node is
traversed using its path. But currently, it compares with the path of
all the open nodes to find a match.
So if you 1000+ tables and the tables node is open and if you click on
a view, the view path is compared with all the 1000+ tables (and with
any other open nodes above) before arriving to path. You're at bad luck
if you have more open servers above.
Code is changed to check if the path of node to be found starts with the
current node path. If it doesn't match, why bother the children's of
current node.
This change will not show much effect for small data, but it does matter
for large servers.
One more change is to remove unnecessary calls to find node and use the
data available with Main Menu -> Object to enable/disable node context
menu items.
Ensure columns can be created when they are IDENTITY fields with the CYCLE option enabled. Fixes#4496
Ensure purely numeric comments can be saved on new columns. Fixed#4497
2) Fix an issue where Explain and Explain Analyze are not working, it's regression of #1760. Fixes#4484
3) Fix an issue where Filter toolbar button is not working in view/edit data, it's regression of keyboard navigation. Fixes#4485
When a query is run in the Query Tool, check if the source of the columns
can be identified as being from a single table, and that we have all
columns that make up the primary key. If so, consider the resultset to
be editable and allow the user to edit data and add/remove rows in the
grid. Changes to data are saved using SAVEPOINTs as part of any
transaction that's in progress, and rolled back if there are integrity
violations, without otherwise affecting the ongoing transaction.
Implemented by Yosry Muhammad as a Google Summer of Code project.
