R devel - Jan 2025 - Development version of R fails a regression test

G'day all,

I have daily scripts running to install the patched version of the
current R version and the development version of R on my linux box
(Ubuntu 24.04.1 LTS).

The last development version that was successfully compiled and
installed on my machine is "R Under development (unstable) (2024-12-31
r87500)". Since then the script always fails on one of the regression
tests; details are in the attached file R-CheckLog.txt.

I also include a file (R-Script.txt) which shows why this test fails on
my machine using the last R development version that did install on
my machine.  It seems the command 'si$LAPACK <- ""'
introduces an extra
line in the output of 'capture.output(si)' which then leads to an error
message in 'which(osi != osi.noLA)'.  

Sorry, but I do not have any good idea/suggestion how this issue could
be fixed.

Cheers,
	
	Berwin



-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: R-CheckLog.txt
URL:
<https://stat.ethz.ch/pipermail/r-devel/attachments/20250121/e937617a/attachment.txt>

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: R-Script.txt
URL:
<https://stat.ethz.ch/pipermail/r-devel/attachments/20250121/e937617a/attachment-0001.txt>

>>>>> Berwin A Turlach 
>>>>>     on Tue, 21 Jan 2025 13:23:17 +0800 writes:
    > G'day all, I have daily scripts running to install the
    > patched version of the current R version and the
    > development version of R on my linux box (Ubuntu 24.04.1
    > LTS).

    > The last development version that was successfully
    > compiled and installed on my machine is "R Under
    > development (unstable) (2024-12-31 r87500)". Since then
    > the script always fails on one of the regression tests;
    > details are in the attached file R-CheckLog.txt.

    > I also include a file (R-Script.txt) which shows why this
    > test fails on my machine using the last R development
    > version that did install on my machine.  It seems the
    > command 'si$LAPACK <- ""' introduces an extra line
in the
    > output of 'capture.output(si)' which then leads to an
    > error message in 'which(osi != osi.noLA)'.

    > Sorry, but I do not have any good idea/suggestion how this
    > issue could be fixed.

    > Cheers,
    > 	Berwin

Thank you very much, Berwin,
for the report!

This is a regression/functionality/consistency test I added with
the new smarter printing of the Lapack library version  in
R-devel with svn revisions
- 87492 (sessionInfo() code);
- 87506 for the test (to work on Windows and elsewhere).

and it looks like the test/check is really *not* working on your platform.
To debug and notably fix the issue, I'd be glad if you could
send me the console output on that platform  of

    dput( extSoftVersion()[["BLAS"]] )
    dput( La_library() )
    dput( La_version() )


I definitely will get back to this, within a few days,
but not yet today.

Martin