best way to add unknowns from the command line?

[Chris Hecker]

Sorry I'm late to my own thread, got busy...

> For example, one thing we've certainly had a bug open on is having
>  'bzr status' when in a subdir give the relative paths. Which would
>  make it easier to cut-and-paste them to the 'bzr add' command. Is 
> that a better solution for Chris?

A command line option to print relative paths would certainly help,
yes!  I just want to be able to cut and paste from the display, or do
a trivial grep based filter, or whatever.  I don't want to have to
sed/perl/awk the results to pipe them from one bzr tool to another.

> I'll also mention that I pretty much never sit in a command line
> in a subdirectory.

Yeah, I _only_ sit in subdirectories, I'm never in the root.  It's
just a style thing, I'd guess.

> Though I would also say it depends on how 'hygenic' people are with
>  their directories and their ignore rules. A plain 'bzr add' is 
> almost never a mistake in my directories. Everything I don't want
> to add is already ignored.

I ignore stuff I will never want to add, but I usually have a couple
directories and a couple files that I keep unversioned and unignored
because I'm trying to figure out what I want to do them, and I like
them showing up in bzr st, which I run constantly.  It's kind of like
a little todo list.

Chris


On 2012/03/22 02:58, John Arbash Meinel wrote:
> On 3/22/2012 10:40 AM, Alexander Belchenko wrote:
>> Vincent Ladeuil пишет:
>>>>>>>> Alexander Belchenko <bialix at ukr.net> writes:
>>> 
>>>> Chris Hecker пишет:
>>>>> bzr qadd takes forever to launch because it expands entire 
>>>>> unknown directories on load, which is silly.
>>> 
>>>> I don't think it's silly.
>>> 
>>> Me neither but hey, some people just talk this way (figure of 
>>> speech and stuff like that [1])...
>>> 
>>> On the other hand, an option to stop at the unknown directory 
>>> level could probably help (such directories could then be
>>> either ignored or explored one level at a time or in full under
>>> user control).
> 
>> The whole idea of add is "add unknown files and directories".
>> Where we should stop? Why? Why don't user ignore directories with
>> huge amount of garbage? Where is balance?
> 
> 
> bzr status intentionally does not descend into unknown or ignored 
> directories. We try not to walk them at all, because many trees
> have a "build" directory with thousands of files that we would have
> to walk for no purpose.
> 
> On the other hand, 'bzr add' does recursively add everything.
> 
> A potentially interesting tradeoff would be to have 'qadd' be able
> to collapse directories, and default to collapsing unknown
> directories. Selecting them could still recursively add all files
> underneath.
> 
> Though I would also say it depends on how 'hygenic' people are
> with their directories and their ignore rules. A plain 'bzr add' is
> almost never a mistake in my directories. Everything I don't want
> to add is already ignored.
> 
> As such, I think it would be good to get some more background
> about how they want to use the tool and why that method is
> valuable.
> 
> Certainly *I* don't work that way, so I don't have the knowledge
> about how to make that experience good for the user.
> 
> 
> For example, one thing we've certainly had a bug open on is having 
> 'bzr status' when in a subdir give the relative paths. Which would 
> make it easier to cut-and-paste them to the 'bzr add' command. Is
> that a better solution for Chris?
> 
> I'll also mention that I pretty much never sit in a command line in
> a subdirectory. Which is also different from how some people
> structure their code. I might occasionally descend, but I don't
> really ever need to add a file from a different parent dir.
> 
> I'm curious to learn a bit more about how Chris uses the tool.
> 
> John =:->
> 
>