httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ben Laurie <...@algroup.co.uk>
Subject Re: Whoa! It works!
Date Tue, 15 Jul 1997 11:59:52 GMT
Alexei Kosut wrote:

> On Mon, 14 Jul 1997, Ben Laurie wrote:
>
> > > IMHO, this is a bad idea. There are a fair number of functions
> that
> > > are
> > > simply not used by our standard modules, but are definitely part
> of
> > > the Apache module API.
> >
> > Possibly so. Now, give me a list ;-)
> >
> > > By tagging as needed, you may very well miss some of these. And
> then,
> > > six months down the line, when, say, mod_php gets ported to
> Windows,
> > > we find out we forgot to export bputs().
> >
> > Uhuh, and 10 seconds later, it'll be fixed.
>
> But it'll be a huge mess. "ten seconds later" won't fix the half a
> million copies of 1.3 already out there.

Mmm, yes - I do agree we need to get this sorted out before a final
release. Its OK for an alpha, tho.

> I think it's better to go through the .h files, and tag the ones that
> we
> think are API functions, and then if we miss any, fix them when a
> module
> needs them.

Feel free.

> Here's a list of functions not tagged with API_EXPORT that I think
> should
> be (by the way, you *are* going to finish your
> makefile-reorganization,
> right?), organized by header file.
>
> It's amazing how many symbols make up the darn thing... And this list
> doesn't include the ones you've already tagged, or the many
> "functions"
> that are actually preprocessor macros:
>
> alloc.h:
> clear_pool()
> bytes_in_pool()
> bytes_in_free_blocks()

Are bytes_in_* global APIs?

> buff.h:
> bcreate()
> bpushfd()
> bsetopt()
> bgetopt()
> bsetflag()
> bonerror()
> bread()
> blookc()
> bskiplf()
> bputs()
> bvputs()
> bpritnf()
> vbprintf()

+1

> http_conf_globals.h:
> server_argv0

+1

> http_config.h:
> set_flag_slot()

And there really, really, should be more of these!

> add_module()
> add_named_module()
> clear_module_list()
> find_module_name()
> find_linked_module()

Glurk!

> http_log.h:
> log_error()
> log_unixerr()
> log_printf()

+1

> http_main.h:
> sync_scoreboard_image()

???

> http_protocol.h:
> set_content_length()
> set_last_modified()
> send_fd()
> send_fd_length()
> rputc()
> rputs()
> rwrite()
> rvputs()
> rprintf()
> rflush()
> set_byterange()
> each_byterange()
> client_to_stdout()
> note_auth_failure()
> note_basic_auth_failure()
> note_digest_auth_failure()
> get_basic_auth_pw()

+1

> http_request.h:
> sub_req_lookup_uri()
> sub_req_lookup_file()
> run_sub_req()
> destroy_sub_req()
> internal_redirect()
> is_initial_req()

+1

> httpd.h:
> month_snames
> get_gmtoff()
> ht_time()
> gm_timestr_822()
> getword_nc()
> getword_white()
> getword_white_nc()
> getword_nulls()
> getword_conf_nc()
> get_token()
> find_token()
> find_last_token()
> unescape_url()
> no2slash()
> getparents()
> escape_path_segment()
> escape_html()
> construct_server()
> construct_url()
> escape_shell_cmd()
> count_dirs()
> make_dirstr_prefix()
> make_dirstr_parent()
> make_dirstr()
> make_full_path()
> is_matchexp()
> strcmp_match()
> strcasecmp_match()
> uudecode()
> ind()
> rind()

Aren't these two macros?

> uname2id()
> gname2id()
> is_directory()
> restart_time
> log_assert()

+1

> multithread.h:
> ??? Most of these functions, probably; depends if we want 1.3 modules
>     to be able to be internally multithreaded. Or we want to restrict
>     that to 2.0 modules?
>
> scoreboard.h:
> sync_scoreboard_image()
> get_scoreboard_info()
> exists_socreboard_image()

???

> util_md5.h:
> md5()
> md5contextTo64()
> md5digest()

+1

> util_script.h:
> find_path_info()
> scan_script_header_err()

+1

> send_size()

??Cheers,

Ben.

--
Ben Laurie [Normal service will be resumed as soon as possible]



Mime
View raw message