***catonano_ is now known as catonano
***bsima1 is now known as bsima
***lavaflow_ is now known as lavaflow
<zig>ArneBab: are you sure? did you benchmark the release or git master? <zig>I did no do extensive benchmark yet. <lloda>pls remind me how to adjust the number of columns in backtrace output? <lloda>I wrote @{vec} instead of @var{vec} in a docstring <lloda>I got >Throw to key `parser-error' with args `(#f "Expecting @end for " deffn ", got " (END . #f))'.< <lloda>then I spent the next hour going through texinfo.scm until I saw the error T_T <lloda>it's really too bad that C99 doesn't have optional arguments :-/ <lloda>makes extending an API harder ***ng0_ is now known as ng0
<lloda>so the problem with extending bytevector-fill! from (... bv fill) to (... bv fill start end) is that the C function scm_bytevector_fill_x becomes incompatible. So you'd have to go through the deprecation cycle :-/ <lloda>and all for a function that isn't all that useful in C anyway <lloda>so I'd just leave scm_bytevector_fill_x as it is and extend only the Scheme function. <ArneBab>zig: I just ran a guile-fibers server and attacked it with wrk. I can drive the memory consumption quite high with that. I’ll ask my employer whether I can release the tool as free software. <zig>ArneBab: it would be nice.