On Tue, Jan 29, 2019 at 09:21:01AM +0000, dilyan.palau...@aegee.org wrote:
> For CGI I use thttpd and it tends to remove terminating slashes in the
> requests.
>
> cgit wants to have slash after about/ .
>
> So they do not work together, endless loop happens. Proposed fix:
>
> diff --git a/cmd.c b/cmd.c
> --- a/cmd.c
> +++ b/cmd.c
> @@ -40,16 +40,7 @@ static void atom_fn(void)
> static void about_fn(void)
> {
> if (ctx.repo) {
> - size_t path_info_len = ctx.env.path_info ?
> strlen(ctx.env.path_info) : 0;
> - if (!ctx.qry.path &&
> - ctx.qry.url[strlen(ctx.qry.url) - 1] != '/' &&
> - (!path_info_len || ctx.env.path_info[path_info_len - 1]
> != '/')) {
> - char *currenturl = cgit_currenturl();
> - char *redirect = fmtalloc("%s/", currenturl);
> - cgit_redirect(redirect, true);
> - free(currenturl);
> - free(redirect);
> - } else if (ctx.repo->readme.nr)
> + if (ctx.repo->readme.nr)
> cgit_print_repo_readme(ctx.qry.path);
> else if (ctx.repo->homepage)
> cgit_redirect(ctx.repo->homepage, false);
According to commit d703480 ("about: always ensure page has a trailing
slash") the trailing slash is required for easy embedding of links to
other /about/ pages.
I suspect it's possible to append the trailing slash without going round
a redirect loop, but I don't think this patch will work.
(I'm also inclined to agree that thttpd is broken here and should pass
the URL to CGI scripts as it is received.)
_______________________________________________
CGit mailing list
CGit@lists.zx2c4.com
https://lists.zx2c4.com/mailman/listinfo/cgit
