Hi Charles, The docs clearly state that the dependent targets are always executed (first):
"Note: the dependencies of a target are always executed before testing the target's condition." Source: http://nant.sourceforge.net/release/latest/help/fundamentals/targets.html This makes sense, you actually may want to use the dependent targets to set a property which controls whether the conditional target is executed. <target name="clean-registry" depends="determine-platform-capabilities" if="${platform.supports.registry}"> ... </target> In case of the NAnt build file: the proper way to deal with this is to fix the performance issue on Mono. But I'm afraid that'll take some time (find the problem, fix it, and then wait for the fix to be included in a release). Regards, Gert -----Original Message----- From: Charles Chan [mailto:cchan...@users.sourceforge.net] Sent: vrijdag 5 maart 2010 4:26 To: Ryan Boggs Cc: Gert Driesen; nant-developers@lists.sourceforge.net Subject: Re: [nant-dev] Task/planning for 0.87 release Hello Ryan, Thanks for looking into this. Yeah I follow what you are saying (but I couldn't really test it ... see my other post.) The question basically comes down to this: If the target has a conditional expression. What order is it evaluated in? (I checked the NAnt doc. but it did not specify this clearly.) eg. <target name="install-windows" depends="userdoc" if="${platform::is-win32()}"> Base on your comments, right now, the depends target is executed BEFORE the condition is tested on a Linux system. Logically, however, if the condition expression is false, then the current target is not going to be executed, and in this case, neither should the depends target. So ... to ensure .build scrips are executed as efficiently as possible, yeah, I agree in this case it probably *should* be fixed. ie. only execute the depends and target IFF if/unless attribute tests is true. On the other hand, since I don't know the history behind this, I would like to listen to what the senior developers have to say first. This behavior might be used in production and my concern is that changing it now might affect these users during an upgrade -- even if it was a bug in the software. Regards, Charles --- On Fri, 3/5/10, Ryan Boggs <rmboggs.obsd.po...@gmail.com> wrote: > From: Ryan Boggs <rmboggs.obsd.po...@gmail.com> > Subject: Re: [nant-dev] Task/planning for 0.87 release > To: "cchan...@users.sourceforge.net" <cchan...@users.sourceforge.net> > Cc: "Gert Driesen" <gert.drie...@telenet.be>, "nant-developers@lists.sourceforge.net" <nant-developers@lists.sourceforge.net> > Received: Friday, March 5, 2010, 1:35 AM > Hi, > > Me again. I noticed something in the current > NAnt.build file that may > be causing the massive *nix delay, or at least making it > worse. The > install target on line 822, which I believe is being called > by > (n)make, lists two depending targets called install-windows > and > install-linux. Looking at these targets, only the > install-windows > depends on the userdoc target which takes 30 minutes to run > on > Mono/*nix. The install-linux target only depends on > the build target. > As a test, I removed the install-windows target from the > install > target depends list and reran on Linux. NAnt was > installed in less > than 5 minutes. This makes me think that NAnt is > building the target > depends before it is checking the target's if > statement. If that is > true, is this functionality intended? I would think > that the if > statement would need to be checked prior to running the > depended > targets. > > Hope this makes sense. > > Thanks, > Ryan __________________________________________________________________ Looking for the perfect gift? Give the gift of Flickr! http://www.flickr.com/gift/ No virus found in this incoming message. Checked by AVG - www.avg.com Version: 9.0.733 / Virus Database: 271.1.1/2722 - Release Date: 03/04/10 20:34:00 ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ nant-developers mailing list nant-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nant-developers
