Closed tlienart closed 5 years ago
Merging #109 into master will increase coverage by
0.13%
. The diff coverage is100%
.
@@ Coverage Diff @@
## master #109 +/- ##
=========================================
+ Coverage 93.46% 93.6% +0.13%
=========================================
Files 10 10
Lines 490 500 +10
=========================================
+ Hits 458 468 +10
Misses 32 32
Impacted Files | Coverage Δ | |
---|---|---|
src/NLSolversBase.jl | 80% <ø> (ø) |
:arrow_up: |
src/objective_types/twicedifferentiablehv.jl | 100% <100%> (ø) |
:arrow_up: |
src/objective_types/incomplete.jl | 98% <100%> (+0.32%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 6af840c...71c6403. Read the comment docs.
Welcome to the frightening world of NLSolversBase :) I will have a look as soon as possible. Thanks
bump?
Great, looks good! Sorry for the long wait.
This is in view of https://github.com/JuliaNLSolvers/Optim.jl/issues/738 @pkofod
only_fg_and_hv
for a constructor where only these two things are providedonly_fghv
for a constructor that looks likeonly_fgh
and allows re-using computations for the construction ofHv
(a common trick using the construction of the gradient to getHv
, see issue quoted above).This may break Optim in that the call tofg
is now the "standard" format used elsewhere IE:(f, g, x)
instead of just(g, x)
; I'll open a PR for Optim soon too.PS: failure on nightly are due to the project toml not being up to date, see PR #108Edit marked as "do not merge" for now, the ambiguity of calling convention fornow ok, see also https://github.com/JuliaNLSolvers/Optim.jl/pull/742fg
either with (g, x) or (f, g, x) is causing me some trouble