SCons
CXXPCH and CCPCH, and the PDB-as-SideEffect
This issue was originally created at: 2005-03-16 10:50:07.
This issue was reported by: dobesv.
dobesv said at 2005-03-16 10:50:07
I had a bit of trouble disentangling the diffs for CPPFORCEINCLUDE, CXXPCH and CCPCH, and the PDB-as-SideEffect fixes so they're here in one patch.
I didn't run the test cases, I don't really know how.
issues@scons said at 2005-03-16 10:50:07
Converted from SourceForge tracker item 1164702
stevenknight said at 2006-05-20 19:56:15
Eliminate example SourceForge keyword.
stevenknight said at 2006-05-26 04:14:00
Created an attachment (id=23) Patch, copied from SourceForge tracker.
stevenknight said at 2008-10-02 10:43:24
See also issue #3064, where Greg Noel suggests introspecting on the $*FLAGS variables for the -include option. I think this approach fits better with other SCons functionality, but have left it up to Greg whether he wants to dup that issue to this one.
gregnoel said at 2008-10-02 16:34:32
*** Issue 8 has been marked as a duplicate of this issue. ***
gregnoel said at 2008-10-02 17:13:22
I think the name CPPFORCEINCLUDES is too long; how about just CPPINCLUDES? That is mostly because you'd also need a CPPINCLUDES{PRE,SUF)FIX as well to make it configurable and with FORCE in there, those names would be really long.
Since the way that I discovered this flag was in a *-config file, MergeFlags() will need to be extended to deal with -include file in the input and do the right thing with it.
And there's a question of where to put it on the command line. Does this mean that CPPFLAGS will now have a magic component by default, or will the reference go in one of the command-line variables? And if the latter, should it go in something like _CCCOMMON, or should that be broken up into pre-processor flags and compile flags?
gregnoel said at 2008-10-05 07:59:08
Note that, as the OP wrote, this is an entangling of several issues. Maybe we should break them out into separate issues? Or reopen issue 8 for the -include sub-issue, at least, which is really distinct from the others.
gregnoel said at 2009-02-06 06:42:25
Bug party triage. Issue #8 has been reopened for the -include command-line aspects. PCH may already be supported; someone knowledgable about it should look at that part of the patch. Adding a .pdb file as a side-effect on compiles based on an Environment also needs to be evaluated; it may need to be broken out into a separate issue.
Review and triage this issue at the next bug party.
stevenknight said at 2009-02-22 07:40:46
Assigning target milestone per bug party triage.
garyo said at 2009-05-02 05:43:08
*** Issue 2411 has been marked as a duplicate of this issue. ***
stevenknight attached forceinclude.patch at 2008-03-25 11:08:15.
Patch, copied from SourceForge tracker.
garyo said this issue is duplicated by #2411 at 2009-05-02 05:43:09.
This bugreport is kind of irritating: it seems to pick up after whatever the problem report was, as the first bit of information is "I had a bit of trouble disentangling the diffs" - why was this diff generated? If there missing history?
lost to the annals of time.. ;) I think PCH and PDB are handled. so really it's just --include?
I think both work, but there were some extras in that patch... the pch stuff is split into separate bits for C and C++ and the pdb stuff (not Python debugger, sigh) had some sort of extra thing - none of those are in the current codebase, but the bug has no evidence why they're needed. You mean you don't remember stuff from 15 years ago? :)
also - "issue 8" isn't issue 8 in the current system, as that links to an unrelated PR.
also - "issue 8" isn't issue 8 in the current system, as that links to an unrelated PR.
Found and fixed. Was issue #3064