Re: read-string performance C5 vs C6

On Mon Nov 17, 2025 at 3:46 AM CET, Diego via Chicken-users wrote:
> Hi all,
>
> I was working on porting chicken-update from CHICKEN 5 to 6, and noticed that something like the following took quite a bit longer (around ~1s -> ~5s on my machine):
>
> (http:call-with-input-request
>  "https://eggs.call-cc.org/rss-5.xml"
>  #f
>  (cut read-string #f <>))
>
> I thought it might be an http-client thing, but I replaced the cut read-string with void and the performance is similar, so I'm assuming read-string may be the issue here.
>
> Could this be due to utf-8 strings? Is there any way to speed this up? It's not at all performance critical here; I'm just wondering.

Yes, that's very likely and was to be expected. The impact should be
reasonable though, so I'll look at it, thanks for reporting this.


cheers,
felix