Mats Wichmann

At least at my end, the effort of the `ifort.py` tool to add those extra recognized suffixes (`.i` and `.i90`) messes things up - you get into a place where...

Sigh... I'd swear I had been looking at that very page, must be blind.

There are these comments [in the wiki](/scons/scons/wiki/FortranDiscussion)

Does the normal trick for this kind of thing work here? Make the default be None, and if the value is still None when processing, set the "real" default.

SCons "knows" which source a value came from, since it walks through the list (cli param, SetOption, defaults from option definitions) in `SConsValues.__getattr__`, but it doesn't record or have a...

I see there was a kind of solution in the linked StackOverflow article. Is there anything from that we can bring back to SCons itself? Or alternatively add as a...

Can't remember if this got mentioned somewhere else. How about creating an options object that holds the "real" defaults, and put dummy ones (some kind of sentinel) in the `add_option`...

Oooh, this is an old issue. I've run into something similar... `-isystem` directives. The problem is once scons makes the committment to be "smart" it has to keep handling all...

This issue surfaced again on an scons chat recently. I'm trying to see if the relevant part of the patch in #1107 can be extracted. It doesn't include a test.

Adding a note with linkages - precompiled headers are a current topic of interest (see #1435 and #4026), and the GCC documentation on using precompiled headers (https://gcc.gnu.org/onlinedocs/gcc/Precompiled-Headers.html) describes using `-include`...