Menu
▾
▴
htdig-dev
|
From: David B. <D.B...@vp...> - 2003-06-01 22:29:53
|
Thanks for input. I think I worked things out over the weekend. Did a search for for wordlist_cache_dirty_level and found it in htcommon/defaults.cc. It appears to be set to '1000', I changed that to 40 and all went very well! So I don't know if the need to go from 1000->40 is due to some wrinkle in our web site or because we are running it on an 64bit Itanium rather than the 32 bit system linux usually appears on. But either way it works. Maybe this needs to be a config parameter ? If you have any specific tests you would like me to run, I'll do so and you can add RedHat AdvancedServer on Itanium II to your supported platforms. We are quite happy to undertake tests as requested or before a release. Thanks for your help. ------------------------------------------------------------- David Bannon D.B...@vp... Phone 61 03 9925 4733 Fax 61 03 9925 4647 Mobile 0418 525687 http://www.vpac.org Systems Manager, Victorian Partnership for Advanced Computing ------------------------------------------------------------- ..... Humpty Dumpty was pushed ! > -----Original Message----- > From: Lachlan Andrew [mailto:lh...@us...] > Sent: Saturday, May 31, 2003 5:31 AM > To: D.B...@vp... > Subject: Re: HTDig on IA64 Linux > > > Greetings again, > > On second thoughts, there was a problem with that variable not being > set properly. To be safe, try this week's snapshot, due out on > Sunday or Monday (I don't remember which...). > > Cheers, > Lachlan > > On Fri, 30 May 2003 22:10, Lachlan Andrew wrote: > > > The snapshot would be about a week old. Try setting > > wordlist_cache_dirty_level=4 (the new default value) > > > On Fri, 30 May 2003 14:40, David Bannon wrote: > > > 3.2.04 CVS snapshots - indexes for far too long > > > and then produces an endless list of error > > > messages like this : > > > > > > .......... > > > Try setting wordlist_cache_dirty_level=40 in configuration file |
|
From: Lachlan A. <lh...@us...> - 2003-06-02 13:24:16
|
Greetings David, Thanks for agreeing to do some testing. I'll get back to you with the=20 tests once they pass on all platforms we have direct access to... Most systems break with w_c_d_l =3D 1000, and many with 40, hence the=20 current default of 4. Everything in htcommon/defaults.cc can be=20 overridden by a parameter of the same name in the file htdig.conf=20 (except for the default name of the configure file, of course :) Did=20 you try? If you did and it didn't work, that is a bug! Thanks again, Lachlan On Mon, 2 Jun 2003 08:26, David Bannon wrote: > Did a search for for wordlist_cache_dirty_level and found it in > htcommon/defaults.cc. It appears to be set to '1000', I changed > that to 40 and all went very well! > Maybe this needs to be a config parameter ? > > If you have any specific tests you would like me to run, I'll do so |
|
From: Neal R. <ne...@ri...> - 2003-06-04 21:54:20
|
Lachlan,
I'm working on the WIN32 port and reveiwing changes since my last port..
On 2002/12/30 12:42:59 you were hacking on with a commit message of:
"Forward-port of many 3.1.6 features"
These lines were removed new line 550:
else if (boolean && mystrcasecmp(word.get(), "+") == 0)
tempWords.Add(new WeightWord("&", -1.0));
else if (boolean && mystrcasecmp(word.get(), "-") == 0)
tempWords.Add(new WeightWord("!", -1.0));
Any explaination on this particular change? This looks to implement
+ (required word) and - (exclude word).
Thanks!
Neal Richter
Knowledgebase Developer
RightNow Technologies, Inc.
Customer Service for Every Web Site
Office: 406-522-1485
|
|
From: Lachlan A. <lh...@us...> - 2003-06-05 12:43:52
|
Greetings Neal,
These conditions were subsumed into the earlier tests:
if (boolean && mystrcasecmp(word.get(), "and") =3D=3D 0)
became
if (boolean && (mystrcasecmp(word.get(), "+") =3D=3D 0
|| mystrcasecmp(word.get(), boolean_keywords[AND]) =3D=3D 0))
and similarly for "-".
To my mind this is clearer, and it also should produce (marginally)=20
smaller code.
I hope the porting is going well. Is it going to change the main Unix=20
source, be a patch relative to the main source, or be a stand-alone=20
port?
Cheers,
Lachlan
On Thu, 5 Jun 2003 07:54, Neal Richter wrote:
> On 2002/12/30 12:42:59
> These lines were removed new line 550:
>
> else if (boolean && mystrcasecmp(word.get(), "+") =3D=3D 0)
> tempWords.Add(new WeightWord("&", -1.0));
> else if (boolean && mystrcasecmp(word.get(), "-") =3D=3D 0)
> tempWords.Add(new WeightWord("!", -1.0));
>
> Any explaination on this particular change? This looks to
> implement + (required word) and - (exclude word).
|
|
From: Neal R. <ne...@ri...> - 2003-06-05 14:31:26
|
Dooh! Good one. Need to read more carefully.. I looked at hundreds of
'diffs' that day....
You've done a lot of work recently as I crawl through diffs... EXCELLENT
JOB!
The port will change the main code-branch and will be invisible to Unix...
the differences in code are surrounded by '#ifdefs'.
It's 90% done from a compilation point of view. Next I'll test the
functionality. Native WIN32 has no 'alarm' and associated Unix functions,
so I have to revisit various code sections and implement a parallel
functionality in native WIN32 APIs.
There is also a separate Makefile set and db_config.h & htconfig.h for
WIN32. I am using cygwin and GNU-make environments, but use the msvc
compiler and linker for all code. I will probably create an msvc project
file as well.
Thanks.
On Thu, 5 Jun 2003, Lachlan Andrew wrote:
> Greetings Neal,
>
> These conditions were subsumed into the earlier tests:
> if (boolean && mystrcasecmp(word.get(), "and") == 0)
> became
> if (boolean && (mystrcasecmp(word.get(), "+") == 0
> || mystrcasecmp(word.get(), boolean_keywords[AND]) == 0))
> and similarly for "-".
>
> To my mind this is clearer, and it also should produce (marginally)
> smaller code.
>
> I hope the porting is going well. Is it going to change the main Unix
> source, be a patch relative to the main source, or be a stand-alone
> port?
>
> Cheers,
> Lachlan
>
> On Thu, 5 Jun 2003 07:54, Neal Richter wrote:
>
> > On 2002/12/30 12:42:59
> > These lines were removed new line 550:
> >
> > else if (boolean && mystrcasecmp(word.get(), "+") == 0)
> > tempWords.Add(new WeightWord("&", -1.0));
> > else if (boolean && mystrcasecmp(word.get(), "-") == 0)
> > tempWords.Add(new WeightWord("!", -1.0));
> >
> > Any explaination on this particular change? This looks to
> > implement + (required word) and - (exclude word).
>
Neal Richter
Knowledgebase Developer
RightNow Technologies, Inc.
Customer Service for Every Web Site
Office: 406-522-1485
|
|
From: Lachlan A. <lh...@us...> - 2003-06-16 14:01:20
|
Greetings David,
I think that the code is pretty much finalised now, and it would be=20
great if you could run some tests on the Itanium.
In particular, we have been testing with
(a) ./configure --enable-tests && make check
(b) Add "robotstxt_name: master-htdig" to htdig.conf and do a dig of
http://www.htdig.org
The second test produces a database of around 400MB.
Thanks again,
Lachlan
On Mon, 2 Jun 2003 08:26, David Bannon wrote:
> If you have any specific tests you would like me to run, I'll do so
> and you can add RedHat AdvancedServer on Itanium II to your
> supported platforms. We are quite happy to undertake tests as
> requested or before a release.
--=20
lh...@us...
ht://Dig developer DownUnder (http://www.htdig.org)
|
Thanks for helping keep SourceForge clean.
X