Re: RCS Programming Interface

[Tobias M.]

On 06/29/2013 02:06 PM, Thien-Thi Nguyen wrote:
> No, not at this time.  Could you please describe in more detail the
> algorithms you would like to use, given such a library?
In fact I just want to create a wrapper around rcs that offers a 
different user interface and maybe adds some highlevel features not 
implemented in rcs. I'm not yet sure about which features of rcs I will 
use. At least the basics like checking in/out, locking, showing history 
and diffs.
> If the internals are not suitable for wrapping presently, someone can
> (with reasonable effort) reshape them accordingly.  In fact, that is an
> ongoing process, the result depending largely on the kind of interaction
> client code desires.  That's why the first question above is important.
>
> Until now, the focus has been to keep those scripts happy in the face of
> massive internal changes introduced in 5.8.  I suppose one way to answer
> the (important!) question above is to survey those scripts to infer and
> categorize idioms.  So that's the second question: Have you done that?
I didn't survey much of there scripts as I first wanted to know if there 
is something like an API instead of the command line interface.

Of course for example to provide a different presentation of the log 
history, I could just parse the rcs file. But I thought there might be a 
more convinient way by using functions of rcs that do the parsing for me.
And for the user input (e.g. checking-in a file) I'd to avoid making an 
external call to the rcs binaries.