biology/stacks: Software pipeline for building loci from short-read sequences
ClosedPublic
Actions

Authored by jwb on Mar 28 2018, 5:25 PM.

Details

Reviewers

jrm
wen

Commits

rP466048: biology/stacks: Software pipeline for building loci from short-read sequences

Summary

biology/stacks: Software pipeline for building loci from short-read sequences
Approved by jrm (mentor) or wen (mentor)
Differential to be added to commit message

Test Plan

Passed pkglint -A, stage-qa, poudriere {10.3,11.1}-{amd64,i386}
Tested by researchers

Diff Detail

Repository

rP FreeBSD ports repository

Lint

No Lint Coverage

Unit

No Test Coverage

Build Status

Buildable 15905
Build 15901: arc lint + arc unit

Event Timeline

jwb created this revision.Mar 28 2018, 5:25 PM

Herald added a subscriber: mat. · View Herald TranscriptMar 28 2018, 5:25 PM

Harbormaster completed remote builds in B15854: Diff 40847.Mar 28 2018, 5:25 PM

I notice some perl and php files in pkg-plist. Looking at http://catchenlab.life.illinois.edu/stacks/manual-v1/#install they describe some other dependencies. How important is the database and web interface functionality? A run-time dependency on a web server is probably over the top, but maybe a note in pkg-message saying something like, "If you want support for these extras, install...".

biology/stacks/Makefile
19	`+=`? GNU_CONFIGURE=yes in the `Standard bsd.port.mk Variables` block with the LDFLAGS assignment?

Good idea. We don't use stacks directly here, only as a dependency for ddocent, which people run on our clusters. I'll take a look at how it might be used on a desktop machine and make some adjustments along the line you suggested.

mat added inline comments.Mar 30 2018, 5:09 PM

biology/stacks/Makefile
19	Technically, GNU_CONFIGURE includes USE_CONFIGURE, so having it in the USES block is all right.

Here's the best I can come up with short of setting up a web server and
testing, which I don't have time for. I'm not sure we have anyone here who
would know how to use it anyway. At least this points potential users in
the right direction.

Harbormaster completed remote builds in B15905: Diff 40934.Mar 31 2018, 12:29 AM

Regarding GNU_CONFIGURE: Yeah, I've always read this as "USES_GNU_CONFIGURE", so it seems logical to keep it in the USES section. Thanks...

Ok, just the += to = and I will sign off.

Are you referring to LDFLAGS+=? If so, I think that should be left as-is.

Mathieu's recent comment:

Old habit. I came under the impression long ago that using CONFIGURE_ARGS=
instead of CONFIGURE_ARGS+= would clobber args injected by the ports system,
and I transferred the same habit to cmake. This doesn't appear to be true
anymore, if it ever was.

Time to break the habit. Thanks for the heads-up.

The only variables that you must use += for are the ones you usually set
in make.conf like CFLAGS.

Seems to always be done this way as well:

<<<ROOT@cray.acadix>>> /usr/ports/biology/stacks 1017 # port-grep ^LDFLAGS
accessibility/caribou/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
accessibility/mousetweaks/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
accessibility/orca/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
accessibility/speech-dispatcher/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
archivers/advancecomp/Makefile:LDFLAGS+= -L${LOCALBASE}/lib -lzopfli
archivers/grzip/Makefile:LDFLAGS+= -L${LOCALBASE}/lib -lintl
archivers/libarchive/Makefile:LDFLAGS+= "-L${LOCALBASE}/lib"
archivers/lrzip/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
archivers/lzop/Makefile:LDFLAGS+= -L${LOCALBASE}/lib
archivers/p5-Compress-LZO/Makefile:LDFLAGS+= -L${LOCALBASE}/lib

Thanks...

Oh yes. I was just checking to make sure you were paying attention. :-P

I did actually test with make -C/usr/ports/biology/stacks -VLDFLAGS, with = and +=, but because I initially misunderstood @mat's comment, my test was flawed. I thought he meant when setting the value in make.conf use +=, but I now see that is wrong. Sorry.