(CC'd to the RCS experts) >>>>> On Wed, 26 Sep 2001 16:49:30 +0200, Ziya Suzen <[EMAIL PROTECTED]> said:
> Hi, > I am working on RCS file parsing. I had to do it cause I have 3,000 > files with about 1 million revisions and I have to check them out one by > one and collect information. Using existing Rcs module is very slow due > to system calls to co for every revision. There is VCS::Rcs > which I haven't tested but looks like it makes syscalls to 'co' too. > Any way, what i do is to parse all the revisions into core and process > them there. Which is about 10 times faster then using standart 'co' (or > Rcs.pm). Ofcourse if you are talking about checking out atleast 20-25 > revisions from the same revision file. > The modules I wrote -for my own usage- are: > XXX::Parser # this the main interface > XXX::YappRcsParser # produced from a Yapp grammar (Yapp::Parse) > XXX::Deltatext # impliments RCS deltatext algorithms > Would you please suggest; > 1- Shall I post them to CPAN, do you think others might use it? Definitely. > 2- If so, do you think RCS::Parser is a good name to use or not? So far the namespaces for RCS are unsatisfying: VCS::RCS, VCS::Rcs, and Rcs all exist, maybe it's time to clean that up. Maybe one of the RCS experts can chime in and voice an opinion? -- andreas > Thank you, > Ziya Suzen
