On Thu, Jun 8, 2017 at 5:10 PM Nicolas Goaziou <
address@hidden> wrote:
> I remember reading that someone else too saw the fortran errors.. Is it
> because I don't have fortran on my system?
No, I don't think so. OTOH, I have no idea about where this comes from.
Did little digging, TIL that the gfortran binary is part of GCC.. as gfortran binary. So I do have fortran, but I never use it.
Also found out why fortran tests were failing for me.. by default the gcc version on my system (RHEL 6.6) is 4.4.7. Using GCC 6.1.0 passes the full "make test" on emacs 25.2.
For example, this is one of the ob-fortran examples that was failing:
* matrix
:PROPERTIES:
:ID: 3f73ab19-d25a-428d-8c26-e8c6aa933976
:END:
Real matrix as input
#+name: fortran-input-matrix1
| 0.0 | 42.0 |
| 0.0 | 0.0 |
| 0.0 | 0.0 |
#+name: fortran-input-matrix2
| 0.0 | 0.0 | 0.0 |
| 0.0 | 0.0 | 42.0 |
#+begin_src fortran :var s=fortran-input-matrix1 :results silent
write (*, '(i2)'), nint(s(1,2))
#+end_src
#+begin_src fortran :var s=fortran-input-matrix2 :results silent
write (*, '(i2)'), nint(s(2,3))
#+end_src
Doing C-c C-c on the first source block gives this error with gcc 4.4.7:
/tmp/babel-6872M6V/fortran-src-6872rcR.F90:4.27:
real, parameter :: s(3,2) = transpose( reshape( (/(/0.0, 42.0/), (/0.0, 0.0/),
1
Error: transformational intrinsic 'transpose' at (1) is not permitted in an initialization _expression_
/bin/sh: /tmp/babel-6872M6V/fortran-bin-68724mX: Permission denied
That _expression_ is apparently legal in newer fortran, is my guess.
> About the other 3 errors:
Since our buildbot doesn't report any problem, this seems specific to
Emacs master.
Thanks. I confirm that make test on org master passes with emacs 25.2.
> Test test-org-export/file-uri condition:
> (ert-test-failed
> ((should
> (equal "file://address@hidden:papers/last.pdf"
> (org-export-file-uri "/address@hidden:papers/last.pdf")))
> :form
> (equal "file://address@hidden:papers/last.pdf"
> "file:///address@hidden:papers/last.pdf")
> :value nil :explanation
> (arrays-of-different-length 40 41
> "file://address@hidden:papers/last.pdf"
> "file:///address@hidden:papers/last.pdf" first-mismatch-at 7)))
This one should be easy to debug since
(org-export-file-uri "/address@hidden:papers/last.pdf")
is easy to reproduce. Could you investigate where the spurious "/" comes
from?
Will do.
> FAILED 646/734 test-org/custom-properties
> FAILED 668/734 test-org/forward-paragraph
These one are related to invisible text. I don't what is going to change
in this area in next Emacs release.
I'll have to leave this for a future debug.. Someone else wants to help out with this?
You can run, e.g.,
BTEST_RE="foo" make test
Thanks!