Many Webkit layout tests use DRT (DumpRenderTree), which dumps the layout object tree and compares against reference text files. These became difficult to maintain as the layout engine evolved and such tests are not portable to any other engine. Testing display list dumps sounds like it would have similar downsides.
Unsurprisingly I recommend making reftests fuzzy if that's what you need. I think our approach has worked pretty well: the annotation fuzzy(A, B) means that at most B pixels differ, and those pixels differ by at most A in any color channel. This seems to be a good mix of convenience and precision. I resisted allowing fuzziness for a long time for roughly the same reason as Lars --- too tempting --- but I haven't been able to say "I told you so!". Rob -- lbir ye,ea yer.tnietoehr rdn rdsme,anea lurpr edna e hnysnenh hhe uresyf toD selthor stor edna siewaoeodm or v sstvr esBa kbvted,t rdsme,aoreseoouoto o l euetiuruewFa kbn e hnystoivateweh uresyf tulsa rehr rdm or rnea lurpr .a war hsrer holsa rodvted,t nenh hneireseoouot.tniesiewaoeivatewt sstvr esn _______________________________________________ dev-servo mailing list dev-servo@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-servo
