On Mon, 21 Mar 2016 11:29:12 -0700 Cedric BAIL <[email protected]> said:

> On Sat, Mar 19, 2016 at 8:17 AM, Jean Guyomarc'h
> <[email protected]> wrote:
> > Elementary (OSX) build failed one more time :/ with edje_cc failing
> > when compiling the default theme, with a very misleading error (more
> > details later). After some investigation, I found out that images
> > edje_cc attempts to load are never closed...
> > Linux is also affected, but this went unnoticed because the limit for
> > opened files seems to be higher by default than on OSX.
> 
> Oh, never though it would be a problem ! Hum, this is going to be
> interesting to solve. The patch is just turning on thread computation.
> Basically openning all file it can and starting to process them in a
> thread when a thread get available. I don't really see how I can fix
> easily this issue now. Is there a way to reliably detect that we have
> opened to much file ? I guess there is an error returned during open ?
> 
> This will also seriously complexify the logic here, as we will need to
> queue future request while waiting for a slot to become available.
> Very annoying :-(
> 
> So should I turn back off that option for everyone ? Just for OSX ? Or
> just on on Linux ? Opinion welcome.

back off for now until we have some more complex solutions, :(

> Cedric
> 
> > Bisecting pointed me out the guilty commit:
> >
> >>Author: Cedric BAIL <[email protected]>
> >>Date:   Thu Mar 10 16:29:56 2016 -0800
> >>
> >>    edje: seems like preloading is fixed and we can now run edje_cc in //
> >>
> >>diff --git a/src/bin/edje/edje_cc.c b/src/bin/edje/edje_cc.c
> >>index 9452fff..d4c7bb1 100644
> >>--- a/src/bin/edje/edje_cc.c
> >>+++ b/src/bin/edje/edje_cc.c
> >>@@ -38,7 +38,7 @@ int        no_save = 0;
> >> int        min_quality = 0;
> >> int        max_quality = 100;
> >> int        compress_mode = EET_COMPRESSION_HI;
> >>-int        threads = 0;
> >>+int        threads = 1;
> >> int        annotate = 0;
> >> int        no_etc1 = 0;
> >> int        no_etc2 = 0;
> >
> > After reverting it locally, everything went back to normal :-)
> > I didn't want to seem rude by reverting it upstream; I will let you
> > guys deal with this :p
> >
> > I have provided a fix for the misleading error I mentioned earlier
> > (branch devs/jayji/fixes). I am not sure this is the right way to
> > handle this... I would greatly appreciate some review on this patch.
> >
> > Best regards,
> >
> >
> > Jean Guyomarc'h
> >
> > ------------------------------------------------------------------------------
> > Transform Data into Opportunity.
> > Accelerate data analysis in your applications with
> > Intel Data Analytics Acceleration Library.
> > Click to learn more.
> > http://pubads.g.doubleclick.net/gampad/clk?id=278785231&iu=/4140
> > _______________________________________________
> > enlightenment-devel mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> >
> 
> 
> 
> -- 
> Cedric BAIL
> 
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> enlightenment-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
> 


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    [email protected]


------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
_______________________________________________
enlightenment-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to