[Libreoffice-qa] Portable vs Parallel: which one is better?

[Pedro]

Hi all

In order to test regressions it is useful to have several versions
installed.

Since under Windows only one (stable) version can be installed as the
working version, the alternatives are: a Portable version or a parallel
installation.

The main advantage of the Portable version is that it is self-contained,
drive independent and anyone can install it. However not all versions are
available because the guys at PortableApps abandon a branch as soon as there
is a release from a newer branch (e.g. the final version in the 3.5 branch
is 3.5.5 so 3.5.6 and 3.5.7 were ignored)
This is not very good because ideally you would use the final version of
each branch to quickly locate in which branch the regression occurred (and
then check where it occurred within the branch)

A parallel install can be made with ANY installer so all versions are
available. However it requires more knowledge (even using Florian
Reisinger's SI GUI it's not trivial) and the install size is huge (because
all files are unpacked)

My questions are: 
1)If a bug occurs in a portable or parallel install can it be assumed to
occur in a standard install or some bugs can be due to installation method
(or inherent to the hacks needed to make the program parallel or portable)? 

2) Which one is more realistic when comparing functionality?

Cheers,
Pedro



--
View this message in context: http://nabble.documentfoundation.org/Portable-vs-Parallel-which-one-is-better-tp4056781.html
Sent from the QA mailing list archive at Nabble.com.