Opened 8 years ago

Closed 8 years ago

Last modified 8 years ago

#37498 closed defect (fixed)

libc-headers: archs?

Reported by: ryandesign (Ryan Schmidt) Owned by: mfeiri
Priority: Normal Milestone:
Component: ports Version: 2.1.2
Keywords: Cc:
Port: libc-headers

Description

I should have asked before committing r101028. I noticed the port only installed headers, therefore I added "supported_archs noarch". I removed "ARCHS=${build_arch}" from destroot.env because I couldn't imagine how a port that installs only headers could care about the arch. Now I think I may have made a mistake. Does this port generate different headers depending on the arch?

Change History (4)

comment:1 Changed 8 years ago by ryandesign (Ryan Schmidt)

Port: libc-headers added; libm-headers removed
Summary: libm-headers: archs?libc-headers: archs?

Make that libc-headers, though I did commit a similar change to libm-headers in r101026; I hope that's not bad as well.

comment:2 Changed 8 years ago by mfeiri

I cant point out a specific file, but there are actual buildscripts for many of the header files and some of the build scripts interpret the ARCHS variable. I simply decided to err on the side of caution, because it is a lot of work to verify non-existence of problems for the entire matrix of architecture combinations and platform versions. Same for the libm-headers. I'll undo the patch and bump the revisions.

BTW, the removal of "build.type bsdmake" in r101028 broke the port on darwin9 and bsdmake is indirectly required on darwin 10 and darwin11 as well. But bsdmake is indeed not needed for darwin12, so i'll put this in the relevant subsections.

comment:3 Changed 8 years ago by mfeiri

Resolution: fixed
Status: newclosed

I had a look at libm-headers and have come to believe that we can leave it as "noarch". For libc-headers I've commited an update in r101180. Thanks for your input.

comment:4 in reply to:  2 Changed 8 years ago by ryandesign (Ryan Schmidt)

Replying to mfeiri@…:

I cant point out a specific file, but there are actual buildscripts for many of the header files and some of the build scripts interpret the ARCHS variable. I simply decided to err on the side of caution, because it is a lot of work to verify non-existence of problems for the entire matrix of architecture combinations and platform versions. Same for the libm-headers. I'll undo the patch and bump the revisions.

Sounds good. Sorry for the trouble!

BTW, the removal of "build.type bsdmake" in r101028 broke the port on darwin9 and bsdmake is indirectly required on darwin 10 and darwin11 as well. But bsdmake is indeed not needed for darwin12, so i'll put this in the relevant subsections.

I removed it because I thought that was the default, and obviously it's not. I'll pay more attention and be more careful in the future.

Note: See TracTickets for help on using tickets.