On Fri, 6 Jun 2025 at 16:28, Kristoffer Haugsbakk
<kristofferhaugsbakk@xxxxxxxxxxxx> wrote:
>
> On Fri, Jun 6, 2025, at 11:29, Johannes Schindelin via GitGitGadget wrote:
> > ## http.c ##
> > -@@ http.c: static CURL *get_curl_handle(void)
> > - die("curl_easy_init failed");
> > -
> > - if (!curl_ssl_verify) {
> > -- curl_easy_setopt(result, CURLOPT_SSL_VERIFYPEER, 0);
> > -- curl_easy_setopt(result, CURLOPT_SSL_VERIFYHOST, 0);
> > -+ curl_easy_setopt(result, CURLOPT_SSL_VERIFYPEER, 0l);
> > -+ curl_easy_setopt(result, CURLOPT_SSL_VERIFYHOST, 0l);
>
> Maybe I don’t understand range-diffs enough but it looked like this was
> using `0l` instead of `0L`.[1] However the patches do use `<num>L` instead
> of `<num>l` throughout. Which I like
>
> † 1: Or rather I don’t understand that this is showing `0l`
You can read "--" as "we no longer remove this" and "-+" as "we no
longer add this". In fact, this whole section begins with "-" (in the
outer diff) and we can read this as "we no longer touch this at all."
Where, crucially, it's "we no longer touch this *in this patch*". Patch
1/4 in this v2 does change this from "0" to "0L" and this updated patch
4/4 then leaves this spot as is. Unlike in v1, when this was a much bigger
patch that touched this spot and many others.
Martin
