tech-pkg archive

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

Re: Please review wip/htslib



On 05/15/17 14:55, Johnny C. Lam wrote:
On Mon, May 15, 2017 at 10:48:04AM -0500, Jason Bacon wrote:
I believe this package should be about ready to commit, but I welcome any
suggestions.

This is the first of dozens of scientific packages we hope to commit in the
near future, so I'd like to get as much feedback as possible now and save
some needless chatter about future packages.

Note: I am aware that htslib 1.4 is available, but it has API changes that
make it incompatible with a number of dependent packages, so it will be a
while before it becomes mainstream in genomics research.
1) Why does the package Makefile set CPPFLAGS to -I${PREFIX}/include?
    If it's to find the zlib headers, you shouldn't need to do anything
    at all after including the zlib/buildlink3.mk file.

2) The package Makefile sets MAKE_ENV to various macros for installing
    files.  INSTALL_DIR should probably be set to ${INSTALL_LIB_DIR}.

3) patch-Makefile is incorrect.  The package should be respecting the
    value of ${PKGMANDIR} for the location of the manpages.

4) Why does the patch to configure.ac set the version to 1.3 while the
    package version is 1.3.2?

1) I honestly don't remember, but it doesn't appear necessary in 1.3.2, so I removed it.

2) I'll consider using INSTALL_LIB_DIR, INSTALL_PROGRAM_DIR, etc. This will require more extensive patching, though...

3) Good catch, thanks.

4) This is the risk of doing a SUBST on top of a static patch. When the patch was updated, it picked up the SUBST version change as well. This approach was adopted from the FreeBSD port and I generally avoid it myself. Unnecessary in this case, so I removed the patch file and do everything in the SUBST now.

Thanks for the feedback,

    JB

--
Earth is a beta site.



Home | Main Index | Thread Index | Old Index