On Thu, Jan 8, 2015 at 7:16 PM, <[email protected]> wrote: > 2. I was hoping for a followup question to: > "Fossil does not perform well with very large repo's or histories > 15 > years." > How is the performance hit quantified? 1day or 1hour / 1GB repo / > commit? > What is the limiting factor? > Is it SQLite or the fossil approach itself? >
The number of files is the primary factor, if i'm not sorely mistaken. Building up a manifest of (say) 50000k+ files takes up a large amount of memory and time, and calculating the R-card for it is _tremendously_ expensive (computationally and memory costs). IIRC, the BSD package repo was the initial reason the R-card was eventually made optional (see the 'repo-cksum' option). The mtime-changes option can also speed up access on large repos, but 50000 files in a checkin still requires stat()ing 50000 files. Is there a path to improve this performance similar to the SQLite speed > gains in the last 2 years? > Not a concrete one, but we are always happy when people point out new efficiency improvement possibilities. Since we have literally the world's foremost expert writing much of the SQL, we get the benefit of his ability to write SQL which he knows sqlite can handle most efficiently. Not that _all_ of the SQL is written by Richard, but certainly the core-most SQL is/was primarily his. FWIW (and this is my personal opinion): i've never seen Fossil as having been in the market to compete for users or draw users away from any given SCM. "It is what it is." Fossil serves a particular niche very well (small/medium projects with small, focused teams), and is (if i may say so) an absolute godsend for single-developer projects (its CGI support, though archaic-sounding, is what first got me to take a close look at Fossil). It doesn't support other niches well (massively projects: Kernels, OpenOffice, BSD package super-repository, etc.). i've never seen that as a bad thing. i.e. never felt that fossil _should_ necessarily aim to scale up to those types of projects. That said: it'd be interesting to hear ideas about how it might be (re-)made to. -- ----- stephan beal http://wanderinghorse.net/home/stephan/ http://gplus.to/sgbeal "Freedom is sloppy. But since tyranny's the only guaranteed byproduct of those who insist on a perfect world, freedom will have to do." -- Bigby Wolf
_______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
