Closed yihui closed 12 years ago
Agreed. Overhauling the process used to locate TeX compliers has been on the TODO list---it is slow and doesn't scale well. Now that support for LuaTeX is on the horizon, we need something better.
Lately I just haven't had the time to work on it.
Just cut a 0.6.2 release to push a bugfix and clear up some issues with CRAN checks on R-devel. This release removes require(filehash)
from zzz.R
.
Version 0.7.0 will contain new code for locating TeX compilers.
I've played around with this a little bit. Here is the time it takes to load v0.6.2:
> system.time(require(tikzDevice, quietly=TRUE))
...
user system elapsed
0.220 0.033 0.262
If I completely remove all the compiler checking code, I get:
> system.time(require(tikzDevice, quietly = TRUE))
...
user system elapsed
0.200 0.007 0.209
The compiler checks definitely need to be re-factored to be more generic and maintainable, but I don't think it is worth putting time into a speed-up that is on the order of hundredths or even tenths of a second.
0.26 sounds great. Thanks!
I have two quick suggestions for zzz.R since I see it is slightly slow to load tikzDevice:
require(filehash)
as long as it is stated in DESCRIPTION.Rprofile
), tikzDevice may just go with them and do not go through the process of checking pdflatex and xelatex