[Gnu-arch-users] Re: Obsoleting abrowse

[Matthieu Moy]

Mikhael Goikhman <address@hidden> writes:

>> BTW, "browse" has a --versions option, but not with the same output
>> 
>> I realize I should display this format only with --no-tree-view.
>
> Yes, I see the value in both views.

This is done.

>> >     these are needed to greately speed up the output and not needlesly
>> >     traverse revision (patch-log) directories
>> 
>> With tla 1.0 archives, yes. With baz flattened archives, it shouldn't
>> make any difference.
>
> To generate the "base-0 ... patch-399" line you should enter into
> every revision container directory.

Then, the --versions option should be sufficient (but here also, I
realize that my implementation of --version is inefficient and does
compute "base-0 ... patch-399" although it doesn't display it. But
this is fixed, now)

>>   -r                       Reverse patch logs
>
> Please restore --reverse.

Done.

>>   --show-sealed            Do not hide sealed branches
>
> This is a change in behaviour

I didn't touch this in rbrowse actually.

> ((arguably serious). Since I want to use the old behaviour, some
> shortcut like -S may be nice. (Actually, I would rather use a more
> intuitive --all for this, but it is already taken...)

Sorry, I don't understand what you mean. Can you clarify what you mean
by "the old behaviour"?

>>   -m, --modified FILE      Show revisions modifying FILE
>
> --affecting, --touching?

Why not. I've stolen the option name from fai, but I'm not 100%
satisfied with it.

> I mean, are the file renames included here?

No, they aren't.

>>   -k, --kind               show each revision kind (import, changeset or id)
>
> s/id/tag/?

Yes, done (id was comming from 1.4's abrowse, actually)

>>   -C, --cacherevs          show cached revisions
>>   --desc                   implies -s -c -D -k -C
>
> Hmm, this last bit is a change in behaviour too.

?

I've cloned abrowse on that point. What's the difference?

-- 
Matthieu