httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From C...@PROCESS.COM (Ken A L Coar)
Subject RE: [STATUS] Fri Oct 24 21:45:20 EDT 1997
Date Thu, 01 Jan 1970 00:00:00 GMT
>From the fingers of Jim Jagielski flowed the following:
>Let's all look over this. Very few people are
>bothering to vote... Boredom? In any case,
>there are some issues :/

    Hrmph.  I've been voting and commenting, you just haven't been
    recording them.

>As of:
>    Fri Oct 24 21:45:20 EDT 1997
>    * Dean's [PATCH] Break everything
>	<>
>	Status: Dean +1, Jim +1, Martin +1, Brian +1

    Untested, but a conceptual +1.

>    * Dean's protocol/1195: Bug in Authentication header (fwd)
>	<>
>	Status: Dean +1, Paul +1, Martin +1

    Was there an actual patch?  I thought this was a discussion about
    changing AuthName to use TAKE1 rather than RAW_ARGS so the "s would
    be swallowed, and living with the compatibility PR consequences.
    I'm +1 on that, because the current behaviour is non-compliant with
    RFC 2068.

>    * Dean's [PRE-PATCH] struct sockaddr_in error messages (ap_snprintf)
>      (feature? bug-fix?)
>	<>
>	Status: Dean +1, Ben +1, Jim 0, Martin 0, Brian -1(?)

    +1 if we decide to completely discard ever using an OS-bundled
    snprintf().  -1 otherwise because this breaks that.  Putting this
    stuff into aplog_error() is bogus, IMHO; ap_snprintf() is the
    correct place.

>    * Ken's [PATCH] for mod_autoindex: no-sorting, last-modified
>	<>
>	Status: Ken +1, Dean +1/-1 (see
>	<>

    I'm reworking this one into two separate patches: one for the
    IndexOptions keyword addition (simple), and one for the
    Last-Modified stuff (more complex, since it requires doing the
    HEADER/README stuff as real sub-requests rather than the current
    hack - and turning these into sub-requests has been avoided a long
    time).  So.. more to come on this.

>    * Manoj Kasichainula's <> libap implementation
>	<>
>	Status: Mahoj wants us to "hold off" on this patch, but
>	others could ``re-create'' it :)

    +1 on the theory.

>    * Martin's [PATCH] [FEATURE] Clickable Path Components in ftp dir header
>	<>
>	<>
>	Status:

    It's definitely a feature, but wicked cool.  A standing +1 for it
    whenever we allow features to creep in again.

>Needs patch:

    I wrote this is response to the last [STATUS], but I don't know
    where it stands on resolution:

>>    NameVirtualHost changes need to be clearly documented!  The
>>    directive itself needs to be added to directives.html and core.html.
>>    Currently only the new-for-1.3 and upgrading-to-1.3 documents
>>    mention it, and in nowhere sufficient detail.  I may make a stab at
>>    writing something for it, but that's extremely unlikely until Friday
>>    at the earliest.

>Open issues:
>    * Should we just use HSregex all the time?
>	Votes: Dean +1, Martin +1, Jim +1 (reserved), Brian -1
>	(why make Apache larger if the OS has a working regex?)

    +1 on always using HSRE.  Runtime memory consumption is *not* in the
    code portions.. besides, this lets us know any RE problems are
    *definitely* ours, saving cycles otherwise spent checking the OS.

>    * tarball issues:
>       o Remove .Z files? Dean says yes, others don't really care
>       o md5 or pgp or some combo of both?
>       o Brian says yes... Jim says +1

    MD5 the .gz files, then sign the MD5 hash.

>    * Lars Eilebrecht suggests removing HAVE_SYSLOG for Ultrix 4.4
>	1.3b3-dev won't compile with it 'cause syslog on Ultrix is weird.

    I'd prefer to make the appropriate #ifdefs for ULTRIX if it will
    work thus.

>    * Martin's SEGV, and Alexei's GP Fault under windows (probably the same)

    Related to PR#1257, maybe?

    An additional "open issue":

    * DDA's adding a tray application to the Windoze version for ease of

    I'm +1 on putting this in during the beta cycle.

    #ken    P-)}

View raw message