On Wed, Jun 17, 2015 at 3:09 PM, Luis R. Rodriguez <mcgrof@xxxxxxxxxxxxxxxx> wrote: > As of coccinelle-1.0.0-rc24 Coccinelle got its own built-in > parallelism support which does not require use of the separate > instances to be spawned. Coccinelle's new built-in parallelism > uses the parmap library for this. > > Replace the pycocci script, which is no longer needed. I've > tested this with the 1.0.0 and 1.0.0-rc24 release and get > the same results for both releases. > > When comparing our old way of doing things (using pycocci) > vs using the new parallellism support (-j option for spatch) > I see a general performance boost of at least ~10 seconds > right on average now when generating code using gentree.py > on the backports build server. > > Cc: Xavier Clerc <xavier.clerc@xxxxxxxx> > Signed-off-by: Luis R. Rodriguez <mcgrof@xxxxxxxxxxxxxxxx> > --- > devel/pycocci | 174 ---------------------------------------------------------- > gentree.py | 23 +++++--- > 2 files changed, 16 insertions(+), 181 deletions(-) > delete mode 100755 devel/pycocci I ran into a snag with this feature, I reported it to the cocci mailing list [0], so for now we can't merge this. [0] https://systeme.lip6.fr/pipermail/cocci/2015-June/002140.html Luis -- To unsubscribe from this list: send the line "unsubscribe backports" in
