Re: version control systems in meld
- From: Aaron Bentley <aaron bentley utoronto ca>
- To: stevek gnome org
- Cc: fds comnets rwth-aachen de, "Alex Collins \(Personal\)" <alex bigcatsystems co uk>, Marc Schinnenburg <msg comnets rwth-aachen de>, dk comnets rwth-aachen de, Adam Muegge <amuegge verano com>, meld-list gnome org, Markus Becker <mab comnets rwth-aachen de>
- Subject: Re: version control systems in meld
- Date: Fri, 17 Dec 2004 10:55:01 -0500
Stephen Kennedy wrote:
An extra state, 'illegal' would be very helpful.
What is "illegal"? For instance files with duplicate ids?
Duplicate file ids are illegal, but also, some filenames are disallowed.
High-bit characters are never allowed, and the user can also specify
other illegal types. I usually make .orig and .rej files illegal so I
can't commit until I resolve conflicts.
Can you give an example of how this would arise? How can arch detect a
modification if it is not in the inventory?
You can configure Arch so that adding files is optional, and some
projects do this (e.g. tla itself). So it's actually
source|unadded|modified.
Inventory status (illegal, unknown, missing, source), (added/unadded)
and
Modification status (normal, newly added, removed, modified, conflict)
However, the model you propose will probably work 95% of the time.
I don't think theres a problem in splitting the status into two fields
as arch does. Would that get us to 100%?
To get to 100%, we'd need three fields:
inventory status (illegal, unknown, missing, source)
add status (added/unadded)
modification status (normal, newly added, removed, modified, conflict)
OK, perhaps then the module has some constants which indicate that some
values of the stat result are always empty and the relevant tree column
can be hidden.
Sounds good.
Aaron
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]