I'm afraid I don't. I'm still picking it up. But I did narrow it down to an issue parsing unicode characters.
https://github.com/kschiess/parslet/issues/73 is a much more succinct example. I'm slowly picking up ruby-prof and learning how to use the output to point a finger at misbehaving code. My next step is to install a patched version of 1.9.3 so I could profile the memory to see if the GC is going haywire at some point. -Zach On Mar 19, 2013, at 1:35 AM, Nigel Thorne wrote: > I'm sorry no-one has responded yet to your well constructed question. I > don't have the knowledge to answer you, I'd have to profile the parser to > work out the possible cause. > > Have you any experience profiling Ruby? > > Cheers > Nigel > > On Mar 18, 2013 10:31 AM, "Zach Moazeni" <[email protected]> wrote: > Ah shoot. I realized I forgot to mention some specs. > > I'm using parslet v1.5.0 on ruby 1.9.3-p385. I'm on a Macbook Air, but I > don't think the machine is the issue. > > -Zach > > On Mar 17, 2013, at 7:21 PM, Zach Moazeni wrote: > > > Hi folks, > > > > I'm working on a course grained css parser, and I've noticed that parsing > > css comments swamps my performance. Can anyone give me a hand and point out > > if I'm doing something silly? As well as any tips on how to approach > > optimizing a parser. > > > > https://gist.github.com/zmoazeni/9208f594bdf61aed1419 > > > > https://gist.github.com/zmoazeni/9208f594bdf61aed1419#file-file-with-comments-css > > (file with comments) takes ~20 seconds to parse on my machine while > > https://gist.github.com/zmoazeni/9208f594bdf61aed1419#file-file-without-comments-css > > (file without comments) takes ~3 seconds. > > > > I'm surprised that the comments ruin the performance so dramatically. This > > obviously gets much worse as the css file gets larger. > > > > I appreciate any help, > > > > Thanks! > > -Zach >
