biology/fasttree: update to version 2.1.9

[jwb]

Summary

biology/fasttree: update to version 2.1.9

Changes in version 2.1.9:

Uses Le-Gascuel 2008 model instead of (default) Jones-Taylor-Thorton 1992 model

Test Plan

portlint -AC: Ok
testport: OK (poudriere: 9.3-RELEASE-p45, i386)
testport: OK (poudriere: 9.3-RELEASE-p45, amd64)
testport: OK (poudriere: 10.3-RELEASE-p7, i386)
testport: OK (poudriere: 10.3-RELEASE-p7, amd64)
testport: OK (poudriere: 11.0-RC3, i386)
testport: OK (poudriere: 11.0-RC3, amd64)

[AMDmi3]

I'd silence MKDIRs and do-extract, otherwise LGTM.

[mat]

biology/fasttree/Makefile

23–24

This is not needed.

[jrm]

biology/fasttree/Makefile

23–24

Without that line there is an error. http://pkg.awarnach.mathstat.dal.ca/data/11amd64-default/2016-10-12_16h27m25s/logs/errors/FastTree-2.1.9.log

[jrm]

Silence mkdir in do-extract as recommended by AMDmi3

[swills]

biology/fasttree/Makefile

23–24

In that case you want NO_WRKSUBDIR=yes

[jrm]

Do not explicitly create WRKDIR and add NO_WRKSUBDIR=yes

[jrm]

Silence copy in do-extract target

[swills]

Approved

[jrm]

Chase r426566

[swills]

Approved

[jrm]

I am waiting before committing this because I am seeing frequent core dumps. The problem, however, is not new to this version.

[jwb]

I don't think we have anyone using fasttree presently, but if you send me a test case that produces the core dump, I'll try and reproduce it.

[jwb]

Update to new release 2.1.10.

Holding off on commit pending core dump investigation.

[jrm]

We are not using FastTree either, so maybe I'm doing something stupid.

1. Download alignments at http://www.microbesonline.org/fasttree/#16S
2. FastTree < COG1028.sim.p > blah
3. core dump at https://ftfl.ca/misc/FastTree.core

biology/fasttree/Makefile

6

DISTVERSION ?

7

remove wip

10

Update address?

24

I'm not sure why I left this comment?

[jwb]

Remove wip category

[mat]

biology/fasttree/Makefile

24

Remove please :-)

[jwb]

Fix core dump issue. Code was assuming malloc() aligns to 16 byte boundaries. May be a Linuxism.

Also removed the forced -O3 as I've never seen it make a measurable difference in performance.

[jwb]

Add Makefile. Will send upstream along with malloc patch.

[mat]

biology/fasttree/Makefile

5

You should consider switching to DISTVERSION here as the version changes. In ports Makefiles, you should always try to use DISTVERSION first, as it will make sure to derive a PORTVERSION that is correct per our versioning scheme, and only use PORTVERSION (and possibly DISTNAME) if the final version is wrong.

See 5.2.2. Versions, DISTVERSION or PORTVERSION and the examples in that section.

[mat]

biology/fasttree/Makefile

32

No need to copy it:

MAKEFILE= ${FILESDIR}/Makefile

[jrm]

Fantastic!

[jwb]

Clean up

[jwb]

Thanks, I hadn't given it a full review since commandeering...

The Makefile itself is a dependency for FastTree (so a rebuild happens if Makefile is altered, useful outside ports during development), so it has to be in ${WRKSRC}, or I'd have to pass more args or env from the port Makefile.