Dear maintainer(s),

Your package has an autopkgtest, great. However, I looked into
the history of your autopkgtest [1] on amd64 and arm64 (because it is
blocking perl) and I noticed it fails regularly, while a rerun passes. I
copied some of the output at the bottom of this report. The last warning
has a different number every time.

Because the unstable-to-testing migration software now blocks on
regressions in testing, flaky tests, i.e. tests that flip between
passing and failing without changes to the list of installed packages,
are causing people unrelated to your package to spend time on these
tests.

Please do get in touch if we need to dive into this together. Or if you
want to discuss this issue. I noticed that all the failed runs I checked
were done on the same worker. Could the problem be a timing issue? (The
worker has a spinning disk and is slower than our other workers).

Paul

https://ci.debian.net/data/autopkgtest/testing/arm64/t/trinityrnaseq/10170685/log.gz
https://ci.debian.net/data/autopkgtest/testing/amd64/t/trinityrnaseq/10058739/log.gz
https://ci.debian.net/data/autopkgtest/testing/amd64/t/trinityrnaseq/10172707/log.gz
https://ci.debian.net/data/autopkgtest/testing/amd64/t/trinityrnaseq/10194744/log.gz
https://ci.debian.net/data/autopkgtest/testing/amd64/t/trinityrnaseq/10216335/log.gz

Warning messages:
1: Removed 55 rows containing missing values (geom_point).
2: Removed 1 rows containing missing values (geom_text).
GOBubble, bckgrd color
Warning messages:
1: Removed 55 rows containing missing values (geom_point).
2: Removed 1 rows containing missing values (geom_text).
GOCircle
Error: Breaks and labels are different lengths
Backtrace:
     █
  1. └─GOplot::GOCircle(circ)
  2.   └─gridExtra::grid.arrange(c, table, ncol = 2)
  3.     └─gridExtra::arrangeGrob(...)
  4.       └─base::lapply(grobs[toconv], ggplot2::ggplotGrob)
  5.         └─ggplot2:::FUN(X[[i]], ...)
  6.           ├─ggplot2::ggplot_gtable(ggplot_build(x))
  7.           └─ggplot2:::ggplot_gtable.ggplot_built(ggplot_build(x))
  8.             └─ggplot2:::build_guides(...)
  9.               └─ggplot2:::guides_train(...)
 10.                 ├─ggplot2::guide_train(guide, scale, output)
 11.                 └─ggplot2:::guide_train.colorbar(guide, scale, output)
 12.                   └─scale$get_labels(breaks)
 13.                     └─ggplot2:::f(..., self = self)
Warning messages:
1: Removed 6 rows containing missing values (geom_point).
2: Removed 6 rows containing missing values (geom_point).
Execution halted
Error, CMD:
/usr/lib/trinityrnaseq/Analysis/DifferentialExpression/GOplot.Rscript
--EC_david ./EC.david --EC_genelist ./EC.genelist --pdf_outfile
goplot.pdf died with ret 256 at
/usr/lib/trinityrnaseq/Analysis/DifferentialExpression/prep_n_run_GOplot.pl
line 246.
make[1]: *** [Makefile:4: test] Error 25
make[1]: Leaving directory
'/tmp/autopkgtest-lxc.9tclu2at/downtmp/autopkgtest_tmp/trinityrnaseq/trinity_ext_sample_data/test_GOplot'
make: *** [Makefile:18: test] Error 2
autopkgtest [00:08:05]: test run-tests: -----------------------]

On Thu, 4 Feb 2021 19:01:37 +0100 Paul Gevers <elbrus@debian.org> wrote:

 > Dear maintainer(s),
 >
 > Your package has an autopkgtest, great. However, I looked into
 > the history of your autopkgtest [1] on amd64 and arm64 (because it is
 > blocking perl) and I noticed it fails regularly, while a rerun passes. I
 > copied some of the output at the bottom of this report. The last warning
 > has a different number every time.
 >
 > Because the unstable-to-testing migration software now blocks on
 > regressions in testing, flaky tests, i.e. tests that flip between
 > passing and failing without changes to the list of installed packages,
 > are causing people unrelated to your package to spend time on these
 > tests.
 >
 > Please do get in touch if we need to dive into this together. Or if you
 > want to discuss this issue. I noticed that all the failed runs I checked
 > were done on the same worker. Could the problem be a timing issue? (The
 > worker has a spinning disk and is slower than our other workers).

Trinityrnaseq has a very complicated testsuite, and there are probably
undocumented dependencies between steps.

Since the testsuite does pass often, and perl did migrate, I am reducing
the severity of this issue.

Perhaps that worker needs to be retired.