Ticket #9464 (closed defect: duplicate)
R depends on Fortran, but has no such dependancy in spkg/standard/deps
| Reported by: | drkirkby | Owned by: | GeorgSWeber |
|---|---|---|---|
| Priority: | minor | Milestone: | sage-duplicate/invalid/wontfix |
| Component: | build | Keywords: | |
| Cc: | rlm, jhpalmieri, mpatel | Work issues: | |
| Report Upstream: | N/A | Reviewers: | |
| Authors: | David Kirkby | Merged in: | |
| Dependencies: | Stopgaps: |
Description
In the sage-4.5.alpha4 code, the $SAGE_ROOT/spkg/standard/deps file has this entry for R:
$(INST)/$(R): $(BASE) $(INST)/$(PYTHON) $(INST)/$(ATLAS) $(INST)/$(ICONV)
$(INSTALL) "$(SAGE_SPKG) $(R) 2>&1" "tee -a $(SAGE_LOGS)/$(R).log"
Note, there is no Fortran dependency listed, yet R does have Fortran files:
drkirkby@hawk:~/sage-4.5.alpha4/spkg/standard$ find r-2.10.1.p2 -name '*.f' r-2.10.1.p2/src/src/library/stats/src/stxwx.f r-2.10.1.p2/src/src/library/stats/src/sgram.f r-2.10.1.p2/src/src/library/stats/src/bsplvd.f r-2.10.1.p2/src/src/library/stats/src/stl.f r-2.10.1.p2/src/src/library/stats/src/hclust.f r-2.10.1.p2/src/src/library/stats/src/sslvrg.f r-2.10.1.p2/src/src/library/stats/src/qsbart.f r-2.10.1.p2/src/src/library/stats/src/bvalue.f r-2.10.1.p2/src/src/library/stats/src/loessf.f r-2.10.1.p2/src/src/library/stats/src/lminfl.f r-2.10.1.p2/src/src/library/stats/src/kmns.f r-2.10.1.p2/src/src/library/stats/src/eureka.f r-2.10.1.p2/src/src/library/stats/src/sinerp.f r-2.10.1.p2/src/src/library/stats/src/bvalus.f r-2.10.1.p2/src/src/library/stats/src/ppr.f r-2.10.1.p2/src/src/library/stats/src/portsrc.f r-2.10.1.p2/src/src/appl/dchdc.f r-2.10.1.p2/src/src/appl/chol.f r-2.10.1.p2/src/src/appl/dqrsl.f r-2.10.1.p2/src/src/appl/dqrdc2.f r-2.10.1.p2/src/src/appl/eigen.f r-2.10.1.p2/src/src/appl/dpoco.f r-2.10.1.p2/src/src/appl/dposl.f r-2.10.1.p2/src/src/appl/dpbfa.f r-2.10.1.p2/src/src/appl/dtrco.f r-2.10.1.p2/src/src/appl/dtrsl.f r-2.10.1.p2/src/src/appl/ch2inv.f r-2.10.1.p2/src/src/appl/dpofa.f r-2.10.1.p2/src/src/appl/dqrutl.f r-2.10.1.p2/src/src/appl/dqrdc.f r-2.10.1.p2/src/src/appl/dpbsl.f r-2.10.1.p2/src/src/appl/dpodi.f r-2.10.1.p2/src/src/appl/dqrls.f r-2.10.1.p2/src/src/appl/dsvdc.f r-2.10.1.p2/src/src/modules/lapack/dlapack4.f r-2.10.1.p2/src/src/modules/lapack/vecLibg95f.f r-2.10.1.p2/src/src/modules/lapack/dlapack3.f r-2.10.1.p2/src/src/modules/lapack/dlapack2.f etc
I'll upload a patch and deps file
Dave
Attachments
Change History
Changed 3 years ago by drkirkby
-
attachment
deps
added
$SAGE_ROOT/spkg/standard/deps to force Fortran to build before R
Changed 3 years ago by drkirkby
-
attachment
deps.diff
added
Diff between $SAGE_ROOT/spkg/standard/deps and that is sage-4.5.alpha4
comment:2 follow-up: ↓ 3 Changed 3 years ago by jhpalmieri
This change to deps seems fine -- making the dependencies explicit seems like a good idea -- but note that fortran is a dependency of lapack, lapack is a dependency of atlas, and atlas is a dependency of R. So fortran should always get installed before R even without this change. Have you seen otherwise?
comment:3 in reply to: ↑ 2 Changed 3 years ago by drkirkby
- Status changed from new to needs_review
- Authors set to David Kirkby
Replying to jhpalmieri:
This change to deps seems fine -- making the dependencies explicit seems like a good idea -- but note that fortran is a dependency of lapack, lapack is a dependency of atlas, and atlas is a dependency of R. So fortran should always get installed before R even without this change. Have you seen otherwise?
I've personally not seen a problem.
I must admit I did not look carefully to see the Fortran/LAPACK/ATLAS/R dependency. But Robert thought it was the cause of an OS X problem (see #9435), so upgraded this ticket to blocker. I think you have rather dashed that hope however!
Dave
comment:5 Changed 3 years ago by leif
This change is now incorporated in #9761, which already has positive review, so this ticket can be closed as duplicate.
Mitesh, that's your job... ;-)

This looks like the cause of #9435