[prev in list] [next in list] [prev in thread] [next in thread] 

List:       fossil-users
Subject:    Re: [fossil-users] Using Fossil SCM with Master Repositories held on aFlash Drive ?
From:       Artur Shepilko <nomadbyte () gmail ! com>
Date:       2017-01-31 7:31:43
Message-ID: CAMCMYR43WaLdMdMfLCcM95Se8ZBgkp4u_cRwP44XRiP5rovXTg () mail ! gmail ! com
[Download RAW message or body]

> So, if everything goes according to plan, I think I could do this
> indefinitely without ever needing to push, pull or sync to a Master
> Repository ?  Is this valid and reasonable, or am I missing something
> important here ?  Artur Shepilko suggested having Master Repositories on my
> Home PC and syncing the repositories on the Flash Drive into these, but (in
> a single developer non-networked environment) I do not understand if this
> offers any benefit over simply copying the repository files from the Flash
> Drive to my Home PC, and knowing that the "backup copies" of my repository
> files at home should always be "byte-for-byte identical" to the "live"
> versions on my Flash Drive ?
>

Sounds like you are trying to validate a file-copy approach that you
have in mind. Perhaps file-storage instills in you more trust and
consistency than Fossil sync.
An easy way to advance your choice is to try out the 'fossil sync'
route in some test set up, so you could gain more trust in how it
operates. It is fairly seamless yet robust. Its design goal is
change-consistency, so it off-loads such decisions from the user.

I would not consider a Flash drive as main Reference storage, it's a
gamble.  More of "backup" flash drives -- just a protracted gamble.
Flash drive can be a temp workspace or means of transit, that is in
the absense of a more reliable transport e.g. network.
So I would try to minimize such transit-time, this mitigates the risks
for anything that I would care not to accidentally lose.

Designating a Reference place/storage in this case would simplify
logistics of sync and backup. Be it your "trusted"-PC or someplace
online, the field-updates must find their way into it as soon and
reliably as possible. Flash-repos then would be considered
"repos-in-transit", in general are clones of Reference. There's no
practical need to have these repos copied onto a field-PC, since you
want to leave it clean afterwards anyway. Also, you would probably do
some development on your own PC -- here it is, this needs to sync
somehow too.

Copying the repo files manually is just another way of sync, just in
such case you yourself become the utility that does it. So the success
would depend on your discipline about what/where/when/how. Minimizing
the need for decisions like that would free up attention and
brain-cells for other tasks.

For consistency I would rely on "fossil sync" to do the job of keeping
those Reference repos up-to-date. Isn't it easier to blame a utility
in case it fails, than to assume the blame on self :)  Fossil sync
protocol is tried and tested, file-copy bypasses it altogether.

As for the working copy, Richard's suggestion to use RAM-disk also has
a bonus to possibly speed up the builds or anything that does frequent
writes; flash writes are annoyingly slow at times, but from experience
are quite brisk for commits.
_______________________________________________
fossil-users mailing list
fossil-users@lists.fossil-scm.org
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

[prev in list] [next in list] [prev in thread] [next in thread]