On Sunday, 12 April 2015 16:48:52 UTC+1, Conan Cook wrote: > On Saturday, 11 April 2015 18:46:31 UTC+1, Conan Cook wrote: > > On Saturday, 11 April 2015 18:37:51 UTC+1, Conan Cook wrote: > > > On Saturday, 11 April 2015 18:35:37 UTC+1, Conan Cook wrote: > > > > On Friday, 10 April 2015 18:23:52 UTC+1, David Powell wrote: > > > > > Wonder if something somewhere is calling java.io.FileDescriptor/sync > > > > > - this i uber slow on Windows (and some Linux filesystems) because it > > > > > causes the entire filesystem to sync, not just the selected file. > > > > > > > > > > > > > > > On Fri, Apr 10, 2015 at 6:08 PM, Shaun Mahood <[email protected]> > > > > > wrote: > > > > > Wow, that's crazy. On my main machine I get > > > > > > > > > > Successfully compiled "resources/public/js/app.js" in 5.545 seconds. > > > > > > > > > > on recompilation. > > > > > > > > > > > > > > > > > > > > Same project, making the same change to core.cljs, recompilation > > > > > using "lein figwheel" I get > > > > > > > > > > Successfully compiled "resources/public/js/app.js" in 0.14 seconds. > > > > > > > > > > > > > > > > > > > > Running Windows 10 Preview, processor is i7-4790K @ 4.00GHz, 32.0 GB > > > > > RAM and an SSD. I can try the same thing on a couple older machines > > > > > as well if you want. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Friday, April 10, 2015 at 9:38:47 AM UTC-6, Conan Cook wrote: > > > > > > > > > > > Just do: > > > > > > > > > > > > > > > > > > > > > > lein new chestnut speed-test > > > > > > > > > > > cd speed-test > > > > > > > > > > > lein repl > > > > > > > > > > > (run) > > > > > > > > > > > > > > > > > > > > > > That's enough to show it up for me; a change in core.cljs should > > > > > > trigger a recompilation. > > > > > > > > > > > > > > > > > > > > > > For me on a windows filesystem I get: > > > > > > > > > > > > > > > > > > > > > > Successfully compiled "resources/public/js/app.js" in 23.226 > > > > > > seconds. > > > > > > > > > > > > > > > > > > > > > > On a linux filesystem I get: > > > > > > > > > > > > > > > > > > > > > > Successfully compiled "resources/public/js/app.js" in 0.174 seconds. > > > > > > > > > > > > > > > > > > > > > > That's a pretty hefty difference, whatever your machine it should > > > > > > be obvious. > > > > > > > > > > > > > > > > > > > > > > On Friday, 10 April 2015 15:27:13 UTC+1, Shaun Mahood wrote: > > > > > > > > > > > > Conan, if you have a repo I can try it on a few windows machines > > > > > > > to compare. > > > > > > > > > > > > > > > > > > > > > > > > I did find that chestnut was really slow to refresh changes and > > > > > > > ended up switching to use figwheel instead, it is essentially > > > > > > > instant refresh now when I save a file. > > > > > > > > > > > > > > > > > > > > > > > > On Friday, April 10, 2015 at 5:19:58 AM UTC-6, Conan Cook wrote: > > > > > > > > > > > > > I've switched from OSX to Windows and a Ubuntu VM, and am > > > > > > > > experiencing extremely slow cljs compilation. I've got a few > > > > > > > > chestnut projects, and have tried the chestnut tutorial as > > > > > > > > well, all with the same results. > > > > > > > > > > > > > > > > > > > > > > > > > > I opened an issue over at Chestnut: > > > > > > > > https://github.com/plexus/chestnut/issues/114 > > > > > > > > > > > > > > > > > > > > > > > > > > The problem is that if I'm compiling files on a Windows > > > > > > > > filesystem, it's very slow, whether I'm running leiningen from > > > > > > > > inside Windows or inside Linux. If the files are on the Linux > > > > > > > > VM filesystem and I'm running leiningen in Linux, it's nice and > > > > > > > > snappy. > > > > > > > > > > > > > > > > > > > > > > > > > > Has anyone else had luck compiling cljs on Windows? > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > > Note that posts from new members are moderated - please be patient > > > > > with your first post. > > > > > > > > > > --- > > > > > > > > > > You received this message because you are subscribed to the Google > > > > > Groups "ClojureScript" group. > > > > > > > > > > To unsubscribe from this group and stop receiving emails from it, > > > > > send an email to [email protected]. > > > > > > > > > > To post to this group, send email to [email protected]. > > > > > > > > > > Visit this group at http://groups.google.com/group/clojurescript. > > > > > > > > That certainly sounds plausible, as taking a look at files in use > > > > during the compilation shows that core.js gets written a lot even when > > > > I've only made a change in another namespace. Surely it should be an > > > > incremental compilation? > > > > > > That said, there doesn't seem to be any mention of Filedescriptor in the > > > Figwheel codebase: > > > https://github.com/bhauman/lein-figwheel/search?utf8=%E2%9C%93&q=filedescriptor > > > > Figwheel uses a library called watchtower, which simply does a > > .lastModified call on a java.io.File, nothing fancy. > > > > https://github.com/ibdknox/watchtower/blob/master/src/watchtower/core.clj#L44 > > Having disabled all services I'm able to and killed all processes I can, I > still see the same poor performance. Maybe it's something to do with the > fact I'm running BitLocker on a MacBook Pro? I can't find any info about > that being a problem, however.
The problem is definitely that it's recompiling all the files. I've tried on another Windows 8.1 machine, this time a lenovo yoga 2 pro, with identical results. I've checked the compilation results and it's definitely compiling everything. Any ideas on how I could prevent this? -- Note that posts from new members are moderated - please be patient with your first post. --- You received this message because you are subscribed to the Google Groups "ClojureScript" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/clojurescript.
