gemini_parse_reply: return code and don't close connection clang static analyzer found a possible use after free: if the code is not 2X in gemini_parse_reply we call close_conn. Then, in net_read we set req->done_header, but req may have been free'd! Actually, this is almost impossible to trigger. close_conn never ends up calling free(req) on its first try, because tls_close takes a while and is rescheduled by libevent. (The check req->ctx != NULL in close_conn is always true if it's a gemini request.) Nevertheless, it's clear to move close_conn out of gemini_parse_reply and simply return the response code: it feels wrong that a "parsing function" takes logic decisions.

declare data as uint8_t *; pointer arith on void * is a GNU extension

add some more warnings to the mix

fix tests for build out-of-tree now `make distcheck' works as advertised

add test of parsing and deserialization of gophermaps

move load_page_from_str to telescope.c I'd like to use parser{,_gophermap,_gemtext}.c in standalone testing programs, and this function is getting in the way since it depends on ui.c

switch the cache to the new parser_serialize API

parsers: add serialize function Add a parser_serialize function and an optional function `serialize' for parsers. The default implementation just prints the text line, which is suitable for parser_textplain and parser_textpatch; gemtext and gophermap uses their own specific unparse functions. This is intended for an incoming cache refactoring: we'll cache the page serialized in the correct format, not once it has been translated to gemtext. This will allow in the future to provide commands such as ``save page to disk'' or interface with an on-disk cache.

add fn pointer to the parser init function to struct parser will come in handy in the future cache refactoring.

use parser_init instead of directly calling the parser init fn

correctly save scrolling position Before we didn't save the scrolling position before a refresh (nohist=1) and that's annoying. This saves the scrolling position only in the correct case, which is when changing from a page to another, expect when switching to TAB_LAZY off.

move TAB_LAZY flipping from switch_to_tab in load_url load_url needs to know when a tab passes from lazy to not lazy to correctly saves the scrolling position (see follow-up commit.)

typo in parse_session_line without +4 strtonum sees top=X and cur=Y which are invalid numbers, so the scroll position is lost.

sync changelog

don't reset scroll position if the cursor changed line If the user managed to move the cursor before the page load, don't restore the scroll position. This is to avoid unnecessary jumping and causing possible confusion on the user.