Comments (11)
I think it would be much better to not reanalyze existing entities in the parse
tree(s), because that would be
much more efficient...
Original comment by [email protected]
on 18 Apr 2010 at 2:51
from f2py.
Yes, modules should be analyzed only once.
Could you provide an example code that illustrates this issue?
Original comment by [email protected]
on 18 Apr 2010 at 3:26
from f2py.
oh.. hold on... i figured it out.. the assertion error happens actually not
because of multiple USEs but only
because the array is defined in a derived type defintion.
Here is a little example:
-----
MODULE test
TYPE t
integer, dimension(:,:), pointer :: a
END TYPE t
END MODULE test
-----
I have fixed the bug, see the diff in the block_statements.py file. The
original code calls first
BeginStatement.analyze(self) on the Type instance, which will call in turn the
analyze function on all elements
of content in this instance. But after analyzing the elements of spec the code
iterates again over content,
which causes the assertion failure. This assertion is only checked when the
field in the derived type is an array
(i.e. has the dimension keyword) So we either have to remove this assertion or
move it to another place in the
update routine where it would be checked for all variables, not only arrays.
I atteched a diff with a couple more changes, I caught 1 or 2 other things that
I think are bugs.
After applying those patches I still see one possible problem. Even though the
parser seems to be handling it
correctly, I still think that if a module A in used in modules B and C, and
both B and C are included in D, then
A is loaded twice.. that could be avoided, right? I've attached an example.. if
you read, parse and analyze
testd.f90 you'll see what I mean..
Thanks!
Original comment by [email protected]
on 18 Apr 2010 at 8:50
Attachments:
from f2py.
The attached test.zip file seems to be empty.
I have applied your patch to svn, except the readfortran.py that probably
contained only debugging code.
Original comment by [email protected]
on 20 Apr 2010 at 4:26
from f2py.
Previous comment is mine, not my wifes:)
Original comment by pearu.peterson
on 20 Apr 2010 at 4:28
from f2py.
Ok :) I assumed so..
the diff in readfortran.py is actually strange.. what I did is add source_only
= None in the base class
(FortranReaderBase) (in the init function) because find_module_source_file uses
source_only, without that I got an
exception when this function was called.
yepp.. the test.zip was broken.. I tried to pack those files again..
Original comment by [email protected]
on 20 Apr 2010 at 5:47
- Changed state: Started
Attachments:
from f2py.
When parsing testd.f90, I get warnings that entities a (defined in testa)
and x (defined in testb) are already defined in module c.
This happens when processing statement `use testc`.
After analyzing the code, here is how is I understand the situation:
The entities a and x are added to testd as a result of processing
the statement `use testb`. Note that at the same time testc provides
y (irrelevant here) in addition to a and x. So, when processing
`use testc`, the analyzer tries to add the content of testc to
testd, that is, add a,x,t to testd. Now a,x are already there
due to `use testb` and that triggers the warnings.
I am looking at the fix momentarily. The fix should not to give
a warning when adding equivalent entities from different modules
and warn only if different modules defines different entities with
the same name.
Original comment by pearu.peterson
on 20 Apr 2010 at 7:06
- Changed state: Accepted
from f2py.
I just committed a fix to svn.
Though there was a warning indicating
duplicate loading of a, it was not actually loaded twise.
The warning code just did not take into account that the
same entity could be loaded via several use paths.
Original comment by pearu.peterson
on 20 Apr 2010 at 7:15
from f2py.
Great!
One more thing..
One change I did that you did not include in the patch is adding '.f' into the
list of module_file_extensions. Would
it be a bug to have the .f in that list? the reason I added it, was that when
the reader searched for module files
(get_module_file) it didn't consider files ending with .f - however the fortran
project I am working on is in F90 but
still uses .f extensions for all fortran files. So, adding the .f was a quick
fix for that, but if that would cause
problems elsewhere then we maybe need a better solution...
Thanks!
Original comment by [email protected]
on 20 Apr 2010 at 10:55
from f2py.
It seems that using the Fortran file name extensions is
not so reliable way to determine the used standard and the
format of Fortran source code, unless it explicitely is
.f90, .f95, f03, .f08, or .f77. As it seems, .f
is becoming more popular for f90 and newer codes though
many programs still consider .f files as Fortran 77 files.
I think it is safe to add .f to module_file_extensions,
I must forgot it.
Original comment by pearu.peterson
on 21 Apr 2010 at 4:45
from f2py.
Done!
I will close this issue..
Original comment by [email protected]
on 21 Apr 2010 at 8:30
- Changed state: Fixed
from f2py.
Related Issues (20)
- Needs review: fparser does not parse SciPy's pyf files HOT 3
- Support redirection of stdout / stderr into CPython descriptors HOT 1
- [patch] fix occurrence of "if (F2PY_TUPLE_76) then" HOT 1
- fparser fails to parse interface blocks with parentheses in name, e.g. assignment(=) HOT 2
- [patch] Fix some regular expressions
- [patch] all_mod_provides should contain contents of module_interface
- [patch] find parent module before calling .check_private()
- [patch] Look for ForAllStmt before ForAllConstruct
- power operator in array dimension HOT 1
- STOP command kills entire python instance HOT 10
- [patch] Beginnings of shared-source Python 2/3 support HOT 1
- compiler dependent output
- Are INTERFACE blocks possible with f2py?
- f2py compile multiple files, Exception: Check the blocks HOT 2
- support PyPy HOT 1
- AssertionError when compiling dcuhre HOT 2
- Using user defined type HOT 12
- Fixed Form and AttributeError HOT 3
- Link to user's guide at f2py.com not connected
- local variable 'newline' referenced before assignment
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from f2py.