Hi Kazunobu! On Do, 30 Mär 2017, Kazunobu Kuriyama wrote:
> 2017-03-29 23:06 GMT+09:00 Dominique Pellé <[email protected]>: > > Bram Moolenaar <[email protected]> wrote: > > > Christian Brabandt wrote: > > > >> On Sa, 25 Mär 2017, Bram Moolenaar wrote: > >> > >> > > > https://github.com/chrisbra/vim/commit/3b79bae9a9d1198b28dc606c9a64b > >> > > >> > Thanks. Let's give that a try. > >> > >> That seems to work. I wonder however, why the coverage is less than on > >> coveralls... > > > > Looks like Coveralls is back. You could compare a file to see where the > > difference comes from. > > > I compared coveralls and codecov stats for 3 files at random > (main.c, mbytes.c and normal.c). > > main.c > ======================================================== > #lines #relevant #covered #missed %cov > coveralls 4247 1433 860 635 60.01% > ======================================================== > #line #line #lines #lines %cov > tracked hits partial misssed > codecov 1412 777 0 635 55.02% > ======================================================== > > > mbytes.c > ======================================================== > #lines #relevant #covered #missed %cov > coveralls 6581 1476 991 602 67.14% > ======================================================== > #line #line #lines #lines %cov > tracked hits partial misssed > codecov 1473 871 0 602 59.13% > ======================================================== > > > normal.c > ======================================================== > #lines #relevant #covered #missed %cov > coveralls 9627 3917 3115 1102 79.53% > ======================================================== > #line #line #lines #lines > tracked hits partial misssed %cov > codecov 3843 2741 0 1102 71.32% > ======================================================== > > All files have different coverage statistics. > > The number '#missed lines' in coveralls seem to be > always be equal to the number '#lines missed' in codecov. > This is the most important number in my opinion, as it shows > where we should improve tests. > > Other numbers are always different. I would have expected that > '#relevant' lines in coveralls to be equal to ''#line tracked' > in codecov but they seem be counting lines different. No idea why > yet. I don't think it matters much anyway, as long as we have > a metric which allows to see improvement to code coverage, and > as long as they show the same uncovered lines, which > seems to be the case. > > > Today, following a link given at https://github.com/vim/vim/pull/857# > issuecomment-289990690 (a codecov report on one of Christian's PRs), I found a > page which explains how codecov calculates those numbers: https:// > docs.codecov.io/docs/about-code-coverage . > > According to that, the calculations are configurable to some extent. That > sounds like the digits themselves are not meant to be regarded as absolute, > but > are waiting for our own analysis or evaluation on them. > > In addition, some explanations why coverage sometimes unexpectedly changes are > given at https://docs.codecov.io/docs/unexpected-coverage-changes and https:// > docs.codecov.io/docs/coverage-offset . > > I'm wondering if coverall has similar pages explaining their metrics. If they > have any, that will help us to give an exact comparison between codecov and > coverall. We could probably ask support, however I am not sure it matters much. > > > > I find Codecov web interface nicer and faster to use > than coveralls. With coveralls, browsing through source > files was sometimes jumping up and down for no apparent > reasons, which was quite annoying. No such problem > with codecov. > > > Exactly. +1 Best, Christian -- Auspuffgase: der Weihrauch der Zivilisation. -- -- You received this message from the "vim_dev" maillist. Do not top-post! Type your reply below the text you are replying to. For more information, visit http://www.vim.org/maillist.php --- You received this message because you are subscribed to the Google Groups "vim_dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
