Pedro J. Aphalo

@markbneal While trying to get EIV working I ended implementing support for `nlme:gls()`, i.e, generalized least squares, in `stat_poly_eq()` and `stat_poly_line()`. So now there is support for variance covariates through...

Raised issue for 'refitME' https://github.com/JakubStats/refitME/issues/5

Hi @markbneal The author of 'refitME' has provided a temporary fix to get 'refitME' working with 'ggpmisc'. He will also try to find a proper fix and update his package....

Yes, `stat_poly_eq()`, `stat_ma_eq()` and `stat_quant_eq()` add a `method.label` to the returned `data` object. For example, `mapping = use_label("eq", "method")` does the trick. ``` r library(ggpmisc) ggplot(mpg, aes(displ, hwy)) + geom_point()...

@markbneal I forgot to mention that method labels work correctly only in the version now in R-Universe or here at GitHub. In the CRAN version they were not set when...

@markbneal I see that the fitted model object returned by `lqs()` has a field named `bestone` with indexes to observations. However, when I run the code with an example, `bestone`...

@markbneal It does not look like `bestone` is of any use. However, `robustbase::ltsReg()` seems to be similarly resistant and does return `lts.weights` that actually make sense. I updated the code...

@markbneal I haven't made progress with EIV, but I noticed while implementing support for 'smatr' (#59) in `stat_poly_line()` and `stat_poly_eq()` that `ma()` and `sma()` functions from 'smatr' accept as input...

This is a preliminary implementation based in existing code, not yet addressing #56 and #58.

I have come across this warning rather frequently, but the simple solution in most cases is to increase max overlaps (Inf is accepted). There is parameter `verbose` already and `show_warning`...