Differences between revisions 1 and 2
Revision 1 as of 2017-02-27 00:31:02
Size: 2242
Editor: JunWu
Comment: Initial commit
Revision 2 as of 2017-02-27 00:44:33
Size: 2245
Editor: JunWu
Comment:
Deletions are marked like this. Additions are marked like this.
Line 8: Line 8:
Email: `'@'.join(reversed(['fb.com', 'quark']))` Email: `'\x40'.join(reversed(['fb.com', 'quark']))`

Jun Wu

Working at Facebook on source control.

Email: '\x40'.join(reversed(['fb.com', 'quark']))

IRC: junw

Some (interesting) contribution

Random items in wishlist (in no particular order)

  • In-memory DAG writes (waiting for smf)
  • Sane obsolete markers cycle handling
  • Cleaner (decoupled from Python.h) changelog index in C, so it could be better optimized
  • Phase/head-based repoview. (see the preliminary idea)

  • Unique linkrev per revision (by having some salt coming from the commit node)
  • A separate blob layer (file nodes store blob IDs instead of file contents)
  • More flexible graphlog (make things like issue5434 solvable, and allow more possibilities like diffing two DAGs)
  • Backport a more correct yet efficient diff algorithm (bdiff outputs suboptimal results for large json files)
  • Word-level merge (see a hacky POC)

  • UUIDs in manifest to track renames (and give filelogs better names)


CategoryHomepage

JunWu (last edited 2017-02-27 00:44:33 by JunWu)