Mats Wichmann
Mats Wichmann
> also prepends the path elements of the source file And that's not correct, I presume. The model is that a reference to a file or directory name is interpreted...
Trying to get back to this with one of a couple of patchsets I have... these tests don't seem to fail on unmodified SCons head, which I expected. What am...
Just an update here, sorry it's been so long... I'm fiddling with Fortran module stuff which is a similar problem, though not identical... scanners are not properly accounting for the...
OK, I'm wrong... looks like the scanner does deduce the correct path: ``` scons: Building targets ... looking for 'part.di' in 'hws' ... looking for 'part.di' in 'build/include' ... ......
Added note: I'd really like to see this intentional exclusion ("it's probably a system header, so it's okay we didn't find it in our limited search") documented, in code comments...
Hmmm, anything we want to do with this, nearly 12 years later?
I suspect this is fixed, there's been a lot of care to make sure rawstrings are used everywhere where Python is launched. Assign to self to check later.
Is there something concrete we can do here? The analysis was if source/target guessing left us with source==target, then there's a problem. Other than "don't do that", can that scenario...
Updated to modern `raise` syntax. Is this something worth pursuing?
Copying over a comment from duplicate #2870 which originally proposed the same simple (one-liner) fix: > Unfortunately, the fix I proposed is not enough: different versions of python on different...