Right. That was intentional. I assume by composing you mean all matches and not just the first.
You fine with the global fixture location I picked? Perhaps it's better under the test directory. On Sep 23, 2016 9:38 PM, "Bill Deegan" <[email protected]> wrote: Both options stop at the first matched directory, if I'm reading the code correctly. As opposed to composing them. Either way the code you have does add additional flexibility, and is good enough for now. So go ahead and add a pull request? On Fri, Sep 23, 2016 at 9:23 PM, William Blevins <[email protected]> wrote: > Yes. It should work like that for file_fixture and similarly for > dir_fixture but takes the whole directory obviously. > > On Sep 23, 2016 6:07 PM, "Bill Deegan" <[email protected]> wrote: > >> Dir 1: >> Files: >> a >> b >> c >> >> >> Dir 2: >> Files: >> a >> d >> >> Dir 3: >> Files: >> a >> d >> f >> >> >> If fixture dirs = [1,2,3] >> >> fixture dir should end up as: >> Files: >> a (dir 1) >> b (dir 1) >> c (dir 1) >> d (dir 2) >> f (dir 3) >> >> So like a series of overlays where first has priority. >> >> >> This is similar to how some MVC web frameworks work. >> >> >> >> On Fri, Sep 23, 2016 at 5:15 PM, William Blevins <[email protected]> >> wrote: >> >>> Bill, >>> >>> You might need to give me an example for the directory comment. I'm not >>> sure I follow. >>> >>> V/R, >>> William >>> >>> On Fri, Sep 23, 2016 at 4:13 PM, Bill Deegan <[email protected]> >>> wrote: >>> >>>> O.k so for files it looks good. >>>> >>>> For directories, there's no way to overlay a directory with files from >>>> the next directory in line in the list of fixture dirs, such that the >>>> created dir would be the set of all files in all listed fixture dirs? >>>> >>>> I'm not entirely sure this would be useful though.. >>>> >>>> On Fri, Sep 23, 2016 at 1:51 PM, William Blevins <[email protected] >>>> > wrote: >>>> >>>>> Any thoughts on the actual diff? >>>>> >>>>> On Fri, Sep 23, 2016 at 1:40 PM, Bill Deegan < >>>>> [email protected]> wrote: >>>>> >>>>>> ok. >>>>>> >>>>>> On Fri, Sep 23, 2016 at 1:35 PM, William Blevins < >>>>>> [email protected]> wrote: >>>>>> >>>>>>> Bill, >>>>>>> >>>>>>> The Bitbucket UI doesn't clearly distinguish between bookmarks and >>>>>>> branches. That diff is between SCons tip and my GLOBAL_FIXTURE bookmark. >>>>>>> >>>>>>> V/R, >>>>>>> William >>>>>>> >>>>>>> On Fri, Sep 23, 2016 at 12:52 PM, Bill Deegan < >>>>>>> [email protected]> wrote: >>>>>>> >>>>>>>> Please don't put these things on branches. >>>>>>>> Bookmarks should get you what you need here. >>>>>>>> When we merge the branches come to scons main repo and then we need >>>>>>>> to close them out. >>>>>>>> (this is one (annoying) difference between git and mercurial). >>>>>>>> >>>>>>>> -Bill >>>>>>>> >>>>>>>> On Fri, Sep 23, 2016 at 2:36 AM, William Blevins < >>>>>>>> [email protected]> wrote: >>>>>>>> >>>>>>>>> Team, >>>>>>>>> >>>>>>>>> As Bill and I briefly discussed in a PR, we thought it would be >>>>>>>>> nice if we could have various directories to pull fixtures. I made the >>>>>>>>> locations into a list, and added a global share directory for reusing >>>>>>>>> files >>>>>>>>> that get copy pasta lots. >>>>>>>>> >>>>>>>>> Please take a look at this diff which also includes first pass at >>>>>>>>> fixing Fortran tests in Python3 as an example. >>>>>>>>> >>>>>>>>> https://bitbucket.org/williamblevins/scons/branches/compare/ >>>>>>>>> williamblevins/scons:GLOBAL_FIXTURE%0Dscons/scons:default#diff >>>>>>>>> >>>>>>>>> V/R, >>>>>>>>> William >>>>>>>>> >>>>>>>>> _______________________________________________ >>>>>>>>> Scons-dev mailing list >>>>>>>>> [email protected] >>>>>>>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Scons-dev mailing list >>>>>>>> [email protected] >>>>>>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Scons-dev mailing list >>>>>>> [email protected] >>>>>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>>>>> >>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> Scons-dev mailing list >>>>>> [email protected] >>>>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>>>> >>>>>> >>>>> >>>>> _______________________________________________ >>>>> Scons-dev mailing list >>>>> [email protected] >>>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>>> >>>>> >>>> >>>> _______________________________________________ >>>> Scons-dev mailing list >>>> [email protected] >>>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>>> >>>> >>> >>> _______________________________________________ >>> Scons-dev mailing list >>> [email protected] >>> https://pairlist2.pair.net/mailman/listinfo/scons-dev >>> >>> >> >> _______________________________________________ >> Scons-dev mailing list >> [email protected] >> https://pairlist2.pair.net/mailman/listinfo/scons-dev >> >> > _______________________________________________ > Scons-dev mailing list > [email protected] > https://pairlist2.pair.net/mailman/listinfo/scons-dev > > _______________________________________________ Scons-dev mailing list [email protected] https://pairlist2.pair.net/mailman/listinfo/scons-dev
_______________________________________________ Scons-dev mailing list [email protected] https://pairlist2.pair.net/mailman/listinfo/scons-dev
