You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Rodent of Unusual Size <Ke...@Golux.Com> on 1999/04/13 00:41:09 UTC
Changing Windows dev platform to MSVC++ 6?
I've been told that the output MSVC++ version 6 compiler is
considerably better optimised over V5. Measurably. What's
the general feeling about moving the base Apache development
platform on windows forward from V5 to V6?
--
#ken P-)}
Ken Coar <http://Web.Golux.Com/coar/>
Apache Group member <http://www.apache.org/>
"Apache Server for Dummies" <http://Web.Golux.Com/coar/ASFD/>
Re: follow up handler problem
Posted by Ben Laurie <be...@algroup.co.uk>.
tser wrote:
>
> i l(ove)(ife) to be ignored :o)
>
> *sits relaxed in corner of the mailing list*
> *puts hands in the air*
> "may i have for 2 seconds your attention"?
> (hears mumble of shaking heads in the background)
> No ?-)
> *turns around*
> *talks to the wall* "hi wall!"
>
> "When some (dummy, mostly, somebody like reinder), user makes a handler
> misconfiguration inside his httpd.conf, should he
>
> a) Be blamed, by not getting any errors in his error log, just because a
> file, similair to the requested locationn exist, or
> b) Be warned about his foolishness, by both giving the surfing user a
> (any number) error, and a bravely report that he might have a typo
> c) Be Warned that the Handler Doesn't exist, and a file, similair to the
> similair to the requested location doesn't exist neither.
>
> *stares to the wall* "a and c are the current situation,wall"
>
> *writes in the ground* "Hello, WorldWideWall "
> - Reinder.
Hmm. Well, I'm not going to ingore you, but I can't say I understand you
either. I will note, though, that when I use mod_perl, and have _all_ my
Perl done as an Action, then I get a warning about the handler for Perl
not being found, with great regularity. One of these days I'll get
around to finding out why, and then I may understand what you are
saying.
Cheers,
Ben.
>
> tser wrote:
> >
> > If' seen a lot of similair Reports as i encountered with the dual
> > warning/error about handlers in the bug database.
> >
> > with either warnings about missing files after a wrong handler, or files
> > still shown after a fake missing handler, or abusivly reports the
> > handler isn't there
> >
> > 2751 general open Handler problem false, should not be logged.
> > 3349 general open spurious "handler xxx not found" messages
> > 4090 general open Handler checkup generates warnings with mod_jserv
> > and possibly others
> >
> > these might be related :
> > 3436 general open Continuous handler XXXX not found messages
> > 2584 mod_acti analyzed Seeing alot of these: "handler "cgi-wrapper" not
> > found for" since upgrading
> > 3548 mod_acti open Seeing a lot of "handler "cgi-wrapper" not found
> > for:" Same as PR2584
> >
> > To fall Back on my previous message
> >
> > 1) Should Apache when it doesn't find the handler, Stop processing the
> > request and report a 404 error ?
> > 2) Should Apacke place a warning in the errorlog if the hanlder failes,
> > but the file deliviry succeeds ?
--
http://www.apache-ssl.org/ben.html
"My grandfather once told me that there are two kinds of people: those
who work and those who take the credit. He told me to try to be in the
first group; there was less competition there."
- Indira Gandhi
Re: follow up handler problem
Posted by tser <te...@worldonline.nl>.
i l(ove)(ife) to be ignored :o)
*sits relaxed in corner of the mailing list*
*puts hands in the air*
"may i have for 2 seconds your attention"?
(hears mumble of shaking heads in the background)
No ?-)
*turns around*
*talks to the wall* "hi wall!"
"When some (dummy, mostly, somebody like reinder), user makes a handler
misconfiguration inside his httpd.conf, should he
a) Be blamed, by not getting any errors in his error log, just because a
file, similair to the requested locationn exist, or
b) Be warned about his foolishness, by both giving the surfing user a
(any number) error, and a bravely report that he might have a typo
c) Be Warned that the Handler Doesn't exist, and a file, similair to the
similair to the requested location doesn't exist neither.
*stares to the wall* "a and c are the current situation,wall"
*writes in the ground* "Hello, WorldWideWall "
- Reinder.
tser wrote:
>
> If' seen a lot of similair Reports as i encountered with the dual
> warning/error about handlers in the bug database.
>
> with either warnings about missing files after a wrong handler, or files
> still shown after a fake missing handler, or abusivly reports the
> handler isn't there
>
> 2751 general open Handler problem false, should not be logged.
> 3349 general open spurious "handler xxx not found" messages
> 4090 general open Handler checkup generates warnings with mod_jserv
> and possibly others
>
> these might be related :
> 3436 general open Continuous handler XXXX not found messages
> 2584 mod_acti analyzed Seeing alot of these: "handler "cgi-wrapper" not
> found for" since upgrading
> 3548 mod_acti open Seeing a lot of "handler "cgi-wrapper" not found
> for:" Same as PR2584
>
> To fall Back on my previous message
>
> 1) Should Apache when it doesn't find the handler, Stop processing the
> request and report a 404 error ?
> 2) Should Apacke place a warning in the errorlog if the hanlder failes,
> but the file deliviry succeeds ?
follow up handler problem (6 similair bug reports found)
Posted by tser <te...@worldonline.nl>.
If' seen a lot of similair Reports as i encountered with the dual
warning/error about handlers in the bug database.
with either warnings about missing files after a wrong handler, or files
still shown after a fake missing handler, or abusivly reports the
handler isn't there
2751 general open Handler problem false, should not be logged.
3349 general open spurious "handler xxx not found" messages
4090 general open Handler checkup generates warnings with mod_jserv
and possibly others
these might be related :
3436 general open Continuous handler XXXX not found messages
2584 mod_acti analyzed Seeing alot of these: "handler "cgi-wrapper" not
found for" since upgrading
3548 mod_acti open Seeing a lot of "handler "cgi-wrapper" not found
for:" Same as PR2584
To fall Back on my previous message
1) Should Apache when it doesn't find the handler, Stop processing the
request and report a 404 error ?
2) Should Apacke place a warning in the errorlog if the hanlder failes,
but the file deliviry succeeds ?
Regards,
Reinder Kraaij.
Ps :there might be somethign wrong with this passage in the
http_config.c :
-----
/* Pass one --- direct matches */
for (handp = handlers; handp->hr.content_type; ++handp) {
if (handler_len == handp->len
&& !strncmp(handler, handp->hr.content_type, handler_len)) {
result = (*handp->hr.handler) (r);
if (result != DECLINED)
return result;
}
}
if (result == HTTP_INTERNAL_SERVER_ERROR && r->handler) {
ap_log_rerror(APLOG_MARK, APLOG_NOERRNO|APLOG_WARNING, r,
"handler \"%s\" not found for: %s", r->handler, r->filename);
}
/* Pass two --- wildcard matches */
for (handp = wildhandlers; handp->hr.content_type; ++handp) {
if (handler_len >= handp->len
&& !strncmp(handler, handp->hr.content_type, handp->len)) {
result = (*handp->hr.handler) (r);
if (result != DECLINED)
return result;
}
}
Apache on Win98 / with MSVC++ 6? using nmake / Crash questions. /handler
not found, not report 404 imediatly
Posted by tser <te...@worldonline.nl>.
I today compiled the apache on a win98 / Visual studio 6 machine.
I used nmake, not the GUi, to compile the source, and it went OK.
(expect tons of Signification lost "here and there" warningmessages)
It runs's fine & fast and a standard - no modifications - runs reliable
also.
I placed my own module in the core, and tried to recompile it, and after
some error digging, Finaly, the module worked inside.
Apache -l bravely showed my module (yes, i was happy :o))
I trow up the handler, and made a typo which brings me to the first
question.
In my error log, i got the following output :
[Sat Apr 17 14:24:20 1999] [warn] [client 127.0.0.1] handler
"chatshow" not found for: e:/w/apache/htdocs/chat
[Sat Apr 17 14:24:20 1999] [error] [client 127.0.0.1] File does not
exist: e:/w/apache/htdocs/chat
That looks like that, if the handler does not exist, apache tries to
find the real location of the URL. I created a directory at
e:/w/apache/htdocs/chat and Indeed, Apache then shows that directory.
Without posting *any* message in the error.log, not even a warning.
about the handler.
isn't it safer to report directly a 404 Error, insteed of trying to
lookup the file location of your not-availeble handler ? And Reporting
an Error, Rather then a warning ? e.g :[error] Handler not found
The Second Thing i walked up, was a Crash in apache, (in my own module,
don't worry :o)), It popud up with a well-known Windows-Crash Window,
Asking me for input (eg, do you want to debug, ok..etc).
It was pretty cool apache continued to life !, But i was wondering if
anybody knows how to supress that window, and just write in the
errorlog, that a thread died.
You know, if oneday, a real bug is found, and spread among the net like
fire, i don't want to see my screen full of CrashWindows :o)
Regards,
Reinder Kraaij
Re: Changing Windows dev platform to MSVC++ 6?
Posted by David Whitmarsh <da...@dial.pipex.com>.
On Mon, 12 Apr 1999 18:41:09 -0400, you wrote:
>I've been told that the output MSVC++ version 6 compiler is
>considerably better optimised over V5. Measurably. What's
>the general feeling about moving the base Apache development
>platform on windows forward from V5 to V6?
I've been building, debugging, using, and generally hacking apache on
my 95 box with MSVC6 for some weeks with no problems, at least none
due to any difference between MSVC5 and MSVC6. I don't see any big
issue here.
David
************************************
David Whitmarsh
Sparkle Computer Co Ltd
Sybase C C++ perl UNIX NT
************************************
Re: Changing Windows dev platform to MSVC++ 6?
Posted by Marc Slemko <ma...@znep.com>.
On Tue, 13 Apr 1999, Rodent of Unusual Size wrote:
> Marc Slemko wrote:
> >
> > Do we really care about a better optimized compiler on Win32 at this
> > point?
>
> Sorry, a word got lost there in my message. It's not the operation
> of the compiler that's much improved, it's the output from it.
> That is, there appears to be an excellent chance that something
> compiled with V6 will run measurably faster than something compiled
> with V5.
I don't think that is too likely to impact Apache too much. Try it. I
think you will find that the performance increase by using V6 is trivial.
Re: Changing Windows dev platform to MSVC++ 6?
Posted by Rodent of Unusual Size <Ke...@Golux.Com>.
Marc Slemko wrote:
>
> Do we really care about a better optimized compiler on Win32 at this
> point?
Sorry, a word got lost there in my message. It's not the operation
of the compiler that's much improved, it's the output from it.
That is, there appears to be an excellent chance that something
compiled with V6 will run measurably faster than something compiled
with V5.
If performance is one of our nags on Win32, it might be worth it.
Also, people with MSDN subscriptions have V6 by now, and anyone
who doesn't already have MSVC and buys it at this point will
almost certainly be getting V6.
So I suppose part of the question is also balancing the compatibility
of existing developers against the tools available to new ones.
I'm very fortunate in having both (V5 of my own, and V6 through my
employer), so I'm not really urgent either way. However, after
discovering that there are a couple of migration issues going V5->V6
and no chance of going the other way, I thought I'd bring it up.
--
#ken P-)}
Ken Coar <http://Web.Golux.Com/coar/>
Apache Group member <http://www.apache.org/>
"Apache Server for Dummies" <http://Web.Golux.Com/coar/ASFD/>
Re: Changing Windows dev platform to MSVC++ 6?
Posted by Marc Slemko <ma...@znep.com>.
On Mon, 12 Apr 1999, Rodent of Unusual Size wrote:
> I've been told that the output MSVC++ version 6 compiler is
> considerably better optimised over V5. Measurably. What's
> the general feeling about moving the base Apache development
> platform on windows forward from V5 to V6?
If I had 30 spare minutes, and I had reason, I could do something on the
Win32 code.
If you upgrade to VC6, I can't.
Should you care about me? Probably not. However, I think there are a
number of developers like me which may add up to enough...
Why can't things work right? We support gcc 2.8 and gcc 2.7 etc. Sigh.
Do we really care about a better optimized compiler on Win32 at this
point?
Re: Changing Windows dev platform to MSVC++ 6?
Posted by Pierpaolo Fumagalli <p_...@fumagalli.org>.
Brian Behlendorf wrote:
>
> Availability of MSVC6 - I haven't bought a compiler, uh, ever, so some
> notion of how expensive it is would be appreciated.
>
Around here (in Italy) it's about 1000$ (that's why mod_jserv has so many
troubles to be compiled for windows... MSVC costs too much for us...)
> I guess backwards compatibility for people using V5 is too much to ask...
>
I think it's a must, with the cost of MSVC...
What about a port to the CYGWIN? It's freely available...
Pier
--
Pierpaolo Fumagalli - Java Apache Project - <http://java.apache.org/>
E-Mail <ma...@fumagalli.org> <ma...@iname.com>
---------------------------------------------------------------------
Re: Changing Windows dev platform to MSVC++ 6?
Posted by Brian Behlendorf <br...@hyperreal.org>.
On Mon, 12 Apr 1999, Rodent of Unusual Size wrote:
> I've been told that the output MSVC++ version 6 compiler is
> considerably better optimised over V5. Measurably. What's
> the general feeling about moving the base Apache development
> platform on windows forward from V5 to V6?
Benchmarks.
Availability of MSVC6 - I haven't bought a compiler, uh, ever, so some
notion of how expensive it is would be appreciated.
I guess backwards compatibility for people using V5 is too much to ask...
Brian