At Wed, 06 Feb 2013 16:16:16 -0700, Neil Toronto wrote: > > On 02/06/2013 03:28 PM, stamo...@racket-lang.org wrote: > > stamourv has updated `master' from bb216d142c to 117c81e2a6. > > http://git.racket-lang.org/plt/bb216d142c..117c81e2a6 > > > > =====[ One Commit ]===================================================== > > Directory summary: > > 100.0% collects/meta/ > > > > ~~~~~~~~~~ > > > > 117c81e Vincent St-Amour <stamo...@racket-lang.org> 2013-02-06 17:22 > > : > > | Increase timeout for TR random testing. > > : > > M collects/meta/props | 1 + > > Have you been getting timeout messages from DrDr too?
I have. > I've been getting them for "matrix-tests.rkt" for the last few days, and > just got one for "array-tests.rkt". Both are TR programs. From the > graph, for the latter, it looks like it's been slowing down over the > last week or so with a steep ramp in the last few days that just got > over 90 seconds. > > http://drdr.racket-lang.org/26222/collects/math/tests/array-tests.rkt > > I haven't changed `math/array' or `math/matrix' or their tests in the > last 16 days. Were there any changes to TR that could have slowed down > expansion? The union for `Number' has recently gotten bigger with the inclusion of types for imaginaries, which can slow things down. Typechecking of calls to `=' has also gotten slower, now that its type is used in cases where a special typing rule (faster, but much less precise) was used. Both changes were bug fixes. Vincent _________________________ Racket Developers list: http://lists.racket-lang.org/dev
