[Bug 1200384] Review Request: ocaml-config-file - Configuration file management for OCaml

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



https://bugzilla.redhat.com/show_bug.cgi?id=1200384



--- Comment #6 from Tomas Heinrich <theinric@xxxxxxxxxx> ---
(In reply to Richard W.M. Jones from comment #4)

Thanks for the comments.

> In this very specific instance (but not in OCaml libraries in
> general) the ocaml-config-file package contains only bytecode.
> However we have found in the past that bytecode isn't completely
> non-arch-specific, so I'd be very dubious about making the
> subpackage noarch.  It would require you to build on arm/x86/x86-64
> and then manually compare the files to check there are really
> no differences.
> 
> The -devel subpackage has a *.cmxs file which is really a shared
> library of native code, so that's certainly not noarch.
> 
> I would not advise making either subpackage noarch.

OK, just want to have this documented here.

> > > rpms/x86_64/ocaml-config-file-debuginfo-1.2-2.fc22.x86_64.rpm
> > > ocaml-config-file-debuginfo.x86_64: E: debuginfo-without-sources
> > > 1 packages and 0 specfiles checked; 1 errors, 0 warnings.
> > 
> > Can't really say why this one is happening.
> 
> Have you got the filelist handy?  Since OCaml 4, the compiler
> supports fairly good DWARF generation, but our debuginfo
> tooling doesn't think *.ml is a source file.

$ rpm -qlp rpms/x86_64/ocaml-config-file-debuginfo-1.2-2.fc22.x86_64.rpm
/usr/lib/debug
/usr/lib/debug/.build-id
/usr/lib/debug/.build-id/49
/usr/lib/debug/.build-id/49/e2bfe8f16e0685485f0c8fa717cc2a8b61feb1
/usr/lib/debug/.build-id/49/e2bfe8f16e0685485f0c8fa717cc2a8b61feb1.debug
/usr/lib/debug/usr
/usr/lib/debug/usr/lib64
/usr/lib/debug/usr/lib64/ocaml
/usr/lib/debug/usr/lib64/ocaml/config-file
/usr/lib/debug/usr/lib64/ocaml/config-file/config_file.cmxs.debug
$ find build/config-file-1.2/ | sort
<...>
build/config-file-1.2/config_file.cmi
build/config-file-1.2/config_file.cmo
build/config-file-1.2/config_file.cmx
build/config-file-1.2/config_file.cmxs
build/config-file-1.2/config_file.ml
build/config-file-1.2/config_file.mli
build/config-file-1.2/config_file.o
build/config-file-1.2/config_file_parser.ml4
build/config-file-1.2/debugfiles.list
build/config-file-1.2/debuglinks.list
build/config-file-1.2/debugsources.list
build/config-file-1.2/elfbins.list
build/config-file-1.2/example.ml
<...>

> Also you may have to change the invocation of ocamlc & ocamlopt
> to pass -g everywhere.  Typically upstream OCaml packages
> don't do this consistently.  If you're not passing -g to
> everything, then you'll end up with empty/broken debuginfo
> which might be what's happening here.

Looking at the Makefile, I guess it would have to be patched to pass "-g"
everywhere.

> > Looks benign, but I guess it won't hurt to add an explicit --libdir=...
> 
> It depends if the configure script supplied by upstream is a
> real autotools configure, or something else.  It might choke
> on --libdir.

Checked the configure script and tried a buildrun with the flag, didn't notice
any issues.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
_______________________________________________
package-review mailing list
package-review@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/package-review





[Index of Archives]     [Fedora Legacy]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [KDE Users]     [Fedora Tools]