Re: [Gnu-arch-users] Re: abrowse and rbrowse (what rbrowse has that abrowse doesn't)

[James Blackwell]

In lists.arch.users, you wrote:
> address@hidden (James Blackwell) writes:
>> I'm confused. Are you worried that somebody is going to take abrowse
>> away from you and replace it with a tool that doesn't work quite as
>> good? If so, don't worry. abrowse isn't going anywhere in the near
>> future.  abrowse and rbrowse can *happily* co-exist for an eternity, 
>
> The talk is of merging them; maybe I misinterpreted this, but it sounded
> as if the resulting super-rbrowse would replace abrowse.

Huh? Who's merging rbrowse and abrowse? 

The plan is that some day, when rbrowse has everything that abrowse has
(both capabilities and performance), then, and only then, will abrowse 
go away.

Here's the history: 

I wrote rbrowse because Tom came to me and said 'Hey, James. I really
like the way regexes feel when running tla archives. Can you do
something similiar for abrowse?'. Diving into abrowse's code, I quickly 
realized that not only had abrowse's code already gotten pretty
complicated, it would rapidly get even worse if I started adding regular
expression code into there.

I started thinking of all of the things that a "browse" command would be
expected to do. Thats when I realized that there were really two jobs
here. One was to satisfy a slippery "regexes for abrowse" request. The
other was a need to write a very solid foundation that could be extended
upon in the future.

I also took a step back, and examined what people are most likely to
need when they're browsing and here's some facts I came up with: 

1. People would like a browse command that they can use to see "whats
   new"

2. When people are looking for "whats new" they are frequently doing 
   so with a snapfile

3. People that are interested in "whats new" are likely to be interested
   in certain categories in several archives. I.E. "Show me all the
   changes that happened in every "^tla--" category that I know about.

4. When people are generating a "whats new" list, they probably do 
   the same thing that I do: namely, they do a daily what's new in 
   a crontab and email it to themselves.

5. People would like to search for specific patches based on a variety
   of things, namely keyphrases

> If the intention on the other hand is a single command, then more
> attention should be paid to the details of making rbrowse as easy to use
> as abrowse in the common case.

[two pages deleted]

The problem is that when you look deeper into rbrowse, there's more than
one common case. Sometimes you're looking for a branch in your own
archive. Sometimes, you're looking to in another archive to see if your
latest patches have been accepted. Other times, you're looking at your
own patches to see what you've done lately. Yet more times (this one is
less common), you're looking for a specific patch.

When you get right down to it, its a matter of work habits and
preferences. I make no promises that you're particular work habits are
going to be the primary use case. In fact... well, never mind. 

Yeah, its not complete yet. One of the things I'm planning on doing is
adding archive_name/ syntax.


-- 
James Blackwell          Please do not send me carbon copies of mailing
Smile more!              list posts. Such mail is unsolicited. Thank you!

GnuPG (ID 06357400) AAE4 8C76 58DA 5902 761D  247A 8A55 DA73 0635 7400