[Pkg-javascript-devel] Bug#919588: Bug#919588: nodejs: FTBFS nodejs

[Jérémy Lal]

Le mar. 12 mars 2019 à 11:13, Santiago Vila <sanvila at unex.es> a écrit :

> Hi.
>
> For the record, I have been unable to build nodejs at all since 2018-04.
> This is my build history for buster:
>
> Status: successful  nodejs_8.9.3~dfsg-12_amd64-20180226T123733Z
> Status: failed      nodejs_8.11.1~dfsg-2_amd64-20180418T105030Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180521T234811Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180608T120305Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180614T234712Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180620T121016Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180630T103408Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180709T222124Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180729T033115Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180812T235312Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180830T211131Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20180921T132811Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20181012T152154Z
> Status: failed      nodejs_8.11.2~dfsg-1_amd64-20181207T133517.964Z
> Status: failed      nodejs_10.15.0~dfsg-8_amd64-20190116T104234.702Z
> Status: failed      nodejs_10.15.0~dfsg-10_amd64-20190126T113009.661Z
> Status: failed      nodejs_10.15.0~dfsg-10_amd64-20190131T043316.958Z
> Status: failed      nodejs_10.15.1~dfsg-2_amd64-20190204T103205.172Z
> Status: failed      nodejs_10.15.1~dfsg-5_amd64-20190211T131413.708Z
> Status: failed      nodejs_10.15.1~dfsg-5_amd64-20190301T191242.213Z
> Status: failed      nodejs_10.15.2~dfsg-1_amd64-20190303T150622.376Z
>
> I've put my failed build logs here:
>
> https://people.debian.org/~sanvila/build-logs/nodejs/
>
> I am using sbuild + schroot + eatmydata. Will now try without eatmydata
> to make sure eatmydata is not to blame for the failures.
>
> On Thu, 17 Jan 2019, Jérémy Lal wrote:
>
> > It's building all right - the problem you see is that there are some
> tests (4) failures:
> > not ok 1 parallel/test-assert
> > not ok 1428 parallel/test-repl-preprocess-top-level-await
> > not ok 1447 parallel/test-repl-top-level-await
> > not ok 2181 parallel/test-net-listen-after-destroying-stdin
> >
> > repl and net failures might happen depending on your build env.
> > If i were you i'd simply mark them as FLAKY (there is a debian patch
> doing similar changes).
> >
> > Build tests failures are ignored (see debian/rules) if you build with
> experimental distribution
> > or UNRELEASED,
> > so i suggest you do the build like that, then see if, once installed,
> > node test/parallel/test-assert.js
> > reports the same error or not. Likewise for the other errors.
> > If it doesn't then just shut the failure using the technique in the
> patch.
> > It it still does then it's a good idea to report it upstream.
>
> I don't think this reply is good enough for a FTBFS bug in a supported
> architecture (amd64).
>
> Packages should build for everybody, including the tests, and the end
> user should never be told that they have to try the build in
> experimental, or disable the tests, or, in general, jumping through
> hoops. The build should just work (with 0 exit status, indicating
> success) without doing anything "special".
>

Agreed.

However, if users expect not doing anything special to achieve a build,
maintainers expect users to not do anything special to the build
environment.

In your case (which is different than the original bug report),
you're using eatmydata and that might very well be the source of the
"test-fs-error-messages" test failure.

For the nodejs 8 failures i can't say.
Some tests were flaky and either i disabled them or upstream fixed them.

For the original bug report:
- i am not sure the build environment has the same constraints, and reporter
does not answer
- reporter should try with DEB_BUILD_OPTIONS=parallel=1

Jérémy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-javascript-devel/attachments/20190312/14cebe09/attachment.html>