Hi Chris,
Now, here's an odd thing ... when I built 2.4.11_004 there was no problem
with PDL::Graphics::PGPLOT (and there still is no problem with the
2.4.11_004 build of that module - all tests pass). The test suite reports:
t/pgplot.t .................. ok
However, with 2.4.11_005 I get:
t/pgplot.t .................. "my" variable $x0 masks earlier declaration in
same statement at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2712.
"my" variable $x1 masks earlier declaration in same statement at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2712.
"my" variable $y0 masks earlier declaration in same statement at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2712.
"my" variable $y1 masks earlier declaration in same statement at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2712.
"my" variable $self masks earlier declaration in same statement at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2807.
"my" variable $len masks earlier declaration in same scope at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2809.
"my" variable $self masks earlier declaration in same scope at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2833.
"my" variable $self masks earlier declaration in same scope at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2860.
"my" variable $self masks earlier declaration in same scope at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Graphics/PGPLOT/Window.pm line
2890.
t/pgplot.t .................. skipped: Skipped: PDL::Graphics::PGPLOT not
installed
(This is with MinGW build of 32-bit 5.16.0.)
AFAICT, neither Graphics/PGPLOT/Window.pm nor t/pgplot.t have changed.
Any pointers on where this failure might be originating ?
BTW, I'm going to stop building PDL::Graphics::PLplot as part of my PDL
builds - which means my PDL ppm packages will no longer include
PDL::Graphics::PLplot.
Instead PDL::Graphics::PLplot will have to be installed as a separate 'ppm
install'. This is probably a good thing anyway - as the separate ppm package
(which is the CPAN version of the module) is no older than the version that
ships with PDL and is, in fact, generally more recent.
There's an odd bug with building PDL::Graphics::PLplot as part of PDL. It's
probably a MinGW or gcc bug - far too obscure for me to deal with.
Hopefully, one day, kmx or Mark Dootson will come up against it ....
whereupon it will probably get resolved.
When I last built 2.4.11_004 (only a week or 2 back), I was able to beat the
bug on my 32-bit perls by rebasing
blib/arch/auto/PDL/Graphics/PLplot/PLplot.dll. Tonight, that has stopped
working - for both 2.4.11_004 and 2.4.11_005. I can't think of any reason
for the change in behaviour.
This has been going on for a good 18 months and I've had enough of banging
my head against it.
No problems with 2.4.11_005 and ActivePerl-5.16.0/MSVC++6.0, but with
earlier versions of ActivePerl/MSVC++6.0 I get lots of test failures.
They're probably all in relation to MatrixOps.pm ... but I haven't dug
through them all. Here's an example with ActivePerl 5.12.0:
C:\sisyphusion\PDL-2.4.11_005>perl -Mblib t/vaffine.t
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
430, near "0:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
435, near "1:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
436, near "0:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
853, near "0:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
863, near "$col:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
873, near "$whc("
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
874, near "(:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
893, near "1:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
1004, near "0:"
syntax error at C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm line
1014, near "1:"
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/MatrixOps.pm has too many errors.
Compilation failed in require at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Math.pm line 51.
BEGIN failed--compilation aborted at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Math.pm line 51.
Compilation failed in require at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Primitive.pm line 1794.
BEGIN failed--compilation aborted at
C:\sisyphusion\PDL-2.4.11_005\blib\lib/PDL/Primitive.pm line 1794.
Compilation failed in require at (eval 8) line 6.
BEGIN failed--compilation aborted at (eval 8) line 6.
BEGIN failed--compilation aborted at t/vaffine.t line 4.
Same thing seems to be happening with x64 ActivePerl and the Platform SDK
compiler - 5.16.0 is fine, but fine 5.10.0 throws up the MatrixOps.pm
errors.
There were no such problems with 2.4.11_004.
(These issues with older perls might also affect the MinGW builds. So far,
I've only built 2.4.11_005 for 5.16.0 with MinGW.)
I'll get to looking at these issues tomorrow morning. In the meantime, any
pointers that might save me some time would be most welcome.
In short - everything was hunky dory when I built 2.4.11_004, but it's a
different story now that 2.4.11_005 has arrived ;-)
Cheers,
Rob
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
