Page 1 of 1
gbi problem
Posted: Thu May 12, 2016 9:05 am
by ASX
There is a function in gbi, partition.py, meant to delete some config files: revertChange(self, widget)
the above function, among other things, delete some config files in /tmp/.gbi.
That's fne, as far as I can see.
~~~
There is however a case where you can select manual partitioning ... do something, then decide to click 'back' button and next choose 'full disk' install.
In this case, some config files previously saved in /tmp/.gbi are not deleted, creating ambiguity to the program who will put in effect the selected partitioning.
In my opinion those config files should be always deleted upon entering the window where you choose the type of partitioning (ufs vs zfs vs. custom), most likely by calling the above function revertChange(), but I will leave to Eric the choice about how to modify the program.
~~~
Additionally, there is a type at line 443 of partition.py: "Faulse" instead of "False".
Re: gbi problem
Posted: Sat May 14, 2016 9:26 am
by ericbsd
Yup this need to be handle, will do so before 10.3 Alpha2.
Re: gbi problem
Posted: Sun May 15, 2016 9:19 pm
by ericbsd
I did a change this need to be tested.
Re: gbi problem
Posted: Tue May 17, 2016 5:41 pm
by ASX
I marked this as solved, by error, it is not solved, or least not tested enough.
Re: gbi problem
Posted: Thu May 19, 2016 6:54 am
by ASX
ASX wrote:There is a function in gbi, partition.py, meant to delete some config files: revertChange(self, widget)
There is however a case where you can select manual partitioning ... do something, then decide to click 'back' button and next choose 'full disk' install.
In this case, some config files previously saved in /tmp/.gbi are not deleted, creating ambiguity to the program who will put in effect the selected partitioning.
I do not see commits in ghostbsd/gbi related to the above, unsure if it was addressed somewhere else.
However, there is another corner use case that came to my mind, it is about to launch gbi, perform some custom partitioning, and then abandon/exit gbi completely and restart anew, possibly resulting in a similar scenario as above.
Re: gbi problem
Posted: Thu May 19, 2016 10:27 am
by ericbsd
I have forgotten that one.
Re: gbi problem
Posted: Mon May 23, 2016 3:31 am
by ASX
ASX wrote:ASX wrote:There is a function in gbi, partition.py, meant to delete some config files: revertChange(self, widget)
There is however a case where you can select manual partitioning ... do something, then decide to click 'back' button and next choose 'full disk' install.
In this case, some config files previously saved in /tmp/.gbi are not deleted, creating ambiguity to the program who will put in effect the selected partitioning.
I do not see commits in ghostbsd/gbi related to the above, unsure if it was addressed somewhere else.
However, there is another corner use case that came to my mind, it is about to launch gbi, perform some custom partitioning, and then abandon/exit gbi completely and restart anew, possibly resulting in a similar scenario as above.
Quoting myself, just as a reminder, because I noticed you are going to release ALPHA2.
Let me disagree, this is only significative bug still pending. Once this this fixed, you can safely go for BETA1 instead of ALPHA2, in my opinion.
All is needed here is to assure that every time gbi enter "custom partitioning", all tmp file eventually created in a previous run are being cleared/deleted.
Re: gbi problem
Posted: Mon May 23, 2016 7:15 pm
by ericbsd
Yea I think this one is now fix and I want to release an alpha2 only because we might have problem with uefi, only to be on the safe side.