On Jan 22, 2010, at 10:52 AM, Stephen De Gabrielle wrote: > Hi someone recently mentioned to me that they were uncomfortable with > their entire repository in a single file. I am happy with it, but why > would this be a problem? > -Large projects? > -Big blobs in your source? > -filesystem limits > -lots of submitters (eg git for the Linux kernel) > > I'd like to come up with a good response > but I couldn't find on the FAQ/q's&criticisms.
I don't really understand the objection. Why would you want your data spread out over multiple files? Seems to me, the more file you have, the more likely you are to misplace one or damage one or forget to copy one when you are doing a backup. Could this be a case of "we've never done it that way before"? Every VCS uses a database. CVS, Git, and Hg all happen to use a pile-of- files database of their own design. Fossil uses a well-tested relational database (SQLite) for local storage. Seems to me that outsourcing data storage to a well-tested and reliable database engine rather than growing your own pile-of-files database would yield improved reliability. Why would anybody think otherwise? Please help me to understand. > > Any suggestions appreciated. > > Thanks, > > Stephen > > > > > -- > > -- > Stephen De Gabrielle > [email protected] > Telephone +44 (0)20 85670911 > Mobile +44 (0)79 85189045 > http://www.degabrielle.name/stephen > _______________________________________________ > fossil-users mailing list > [email protected] > http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users D. Richard Hipp [email protected] _______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
