Firefox as plan B - just in case of unsatisfying ux

Question

We are on planning our migration from Windows to Maverick.
For several applications there is a plan B in preparation.
Just in case the given and Maverick included of current build software
does not enable satisfying productivity.
The lack of productivity will occur if customization/personalization/expendability
will be figured out.
Expendability in terms of availability of wide spectrum of downloadable extensions/add-ons.
Following features are asked
- sending a web page to file disk for purposes for later off-line viewing
- cookie management and control
- create desktop shortcut to open page
- zooming in/out/fitting to window embedded pictures
- download manager, simultaneous download of multitude of files, preselection of file types
  to be downloaded
- embedded flash, javascript, java control
- flexible, customizable tabs
- session management, save/restore browser sessions
- session recovery after browser crash
- reopen closed tabs
How well is the chance all these requirements are fulfilled by Safari?
There are mac guys advicing Firefox for reasons closed to expectations as
described above. There are also mac guys advicing Firefox to be run rather on virtual machine,
due to stability and/reliability issues.
If to take Maverick of current build which will be the best option - Firefox direct on Maverick, or Firefox on virt. machine ?

Answer 1

truell wrote:
Why does that happen?There are two things happening:
1. Number of times UDF is called
2. Different results
Number of times UDF is called depends on execution plan (as you already experienced) and on function being declared as deterministic. Results depend on UDF purity (you can read about it here). So if UDF, for example, reads/writes package state or sequence nextval or reads data from a table, etc. number of times UDF is called affects UDF return value. Another word "not pure" UDF can produce different results when called from SQL if optimizer decided to use a different plan.
SY.