Hi @stephano-meschiari,
I like {latex2exp}, but as a basic tool possibly used in different contexts, it should limit its strong dependencies to a bare minimum. {magrittr} is really not required since code could be easily rewritten without the pipe. Also {stringr} brings other packages with him : {stringi}, {rlang}, {cli}, {lifecycle}, {vctrs}, and {glue}. While this is OK for people using the {tidyverse}, it is not for those that stick on base R.
Also note that many base functions equivalent to {stringr} ones are now faster in recent R versions.
Here, I propose a version of {latex2exp} that does not depends on {magrittr} or {stringr}. It passes all tests, produces correct latex2exp_examples() and compiles both vignettes, appart for the bug with \pmod that is also in the original version (and it is not the purpose of this pull request to correct existing bugs).
Hi @stephano-meschiari, I like {latex2exp}, but as a basic tool possibly used in different contexts, it should limit its strong dependencies to a bare minimum. {magrittr} is really not required since code could be easily rewritten without the pipe. Also {stringr} brings other packages with him : {stringi}, {rlang}, {cli}, {lifecycle}, {vctrs}, and {glue}. While this is OK for people using the {tidyverse}, it is not for those that stick on base R. Also note that many base functions equivalent to {stringr} ones are now faster in recent R versions. Here, I propose a version of {latex2exp} that does not depends on {magrittr} or {stringr}. It passes all tests, produces correct
latex2exp_examples()
and compiles both vignettes, appart for the bug with\pmod
that is also in the original version (and it is not the purpose of this pull request to correct existing bugs).