I used a std::map below instead of a DenseHash because we need to retain a sorted order when enumerating.
Tue, Dec 11
- Remove new argument from relocateOne methods
- Introduce new helper function setMicroMipsBit and call it before the relocateOne
Fri, Dec 7
I don't feel enough comfortable with this Python code.
Sorry, it seems that my reply didn't go through.
bsdtar is libarchive (by BSD people too)
We might be able to change NetBSD's tar but I guess that takes very long time. This test is not very important in the sense that this tests just test a corner case. So I think we should just land this to make it work on NetBSD.
I thought NetBSD's tar is bsdtar because it's BSD... Anyways, I think I'm fine with this change, as the new test (which matches both foo\\.o and foo\.o) does not match a string that we don't want it to match.
In my opinion it's better to set UNSUPPORTED and align the NetBSD tar to GNU tar and libarchive.
Tue, Dec 4
It seems to me that just adding --start-lib to his command line can fix the issue, so I'm waiting for Robert's response. If it doesn't work for some reason, we can analyze why it doesn't work and then discuss what we can do for his problem.
Just for the record, this is not as simple as I initially thought. We should only be padding functions that originate from hotpatch-enabled OBJs. Currently, LLVM currently does not generate /hotpatch OBJs (at least the flag isn't set). In theory, x64 and ARM OBJs should be fine to have this flag enabled. I'll do more research and get back.
Sorry only getting around to this comment now. I think this is probably what you are looking for
Mon, Dec 3
Other then my one comment it LGTM.
Sun, Dec 2
This probably needs another approval based on the small changes I made to address the warning
update to address comments
Fri, Nov 30
But 2GB is perhaps still too big and I guess a large part of it can be for dead sections. If we fix this, I'd like to fix it in a proper way so that we can completely eliminate debug info for dead sections.
Commited as rL348001: [PDB] Quote linker arguments containing spaces (mimic MSVC)
After discussing ofline with Stefan, I will take ownership and will finish this patch.
Wed, Nov 28
If we decide to optimize DWARF garbage collection, something generic will be cool.
Thank you for the patch.
Tue, Nov 27
- Simplify handling of the readBfdName return value
As to defining a struct to store the parse result of parseBfdName, I don't think we need to define a new struct. That seems a bit too heavyweight to represent just a temporary value.