BUG FIXES
---------
* The code failed to compile under g++ 2.95 because it caught errors that
previous versions of g++ allowed to compile.
* There were a few mistakes in the section 1 manual pages to cover all the
changes to version 3.0.
Paul:
I setup a web page on an NT server that a client of mine runs using swish++
to index a set of files generated from a database. It works great, in fact
it's faster than the built-in microsoft index server; the only difference is
that index server runs all the time in the background indexing new documents
automatically as they're added to the server (uses up a lot of resources if
you ask me).
See:
http://www.mincava.umn.edu/trainingdb/index.asp
Anyhow, I tried to assemble something to index their whole site and found
that there is something wrong with the win32 version when I try to not use
the -r option. It will recurse through all the subdirectories when i omit
it, but I tried to selectively control which subdirectories should be
indexed using -r, and then specifying a set of directories, and it didn't
work.
I'll see if I can locate the source of the problem, but I can't do it right
away....
-----Original Message-----
From: Paul J. Lucas <pjl@...>
To: SWISH++ Discussion <swish@egroups.com>
Date: Monday, August 02, 1999 9:14 AM
Subject: [swish] SWISH++ 3.0.1 released
>BUG FIXES
>---------
>
>* The code failed to compile under g++ 2.95 because it caught errors that
> previous versions of g++ allowed to compile.
>
>* There were a few mistakes in the section 1 manual pages to cover all the
> changes to version 3.0.
>
>
>------------------------------------------------------------------------
>Click Here to apply for a NextCard Internet Visa and start earning
>FREE travel in HALF the time with the NextCard Rew@rds Program.
>http://clickhere.egroups.com/click/449
>
>
>eGroups.com home: http://www.egroups.com/group/swish
>http://www.egroups.com - Simplifying group communications
>
>
>
hi,
I just upgraded to swish++ 3.0.1. I was wondering if there is any way
to tell
index not to use more than x mb of ram during indexing.
I have a 80mb corpus. I'd like the indexing process to say under, say,
40mb.
I've tried lowering WordThreshold to 100000, but that didn't seem to
help much.
any suggestions?
btw, I'm indexing a large number of fairly small files (2-3kb each) if
that
makes a difference.
thanks
-- James
On Mon, 2 Aug 1999 pixel@... wrote:
> I was wondering if there is any way to tell index not to use more than x mb
> of ram during indexing.
If there were, there would either be a documented command line
option or configuration variable for it.
As stated in the README, the SWISH++ documentaion is complete:
if what you want isn't described there, then SWISH++ doesn't do
it.
In general, there is no way to tell any Unix program to confine
itself to a certain amount of memory.
- Paul
> If there were, there would either be a documented command line
> option or configuration variable for it.
>
> As stated in the README, the SWISH++ documentaion is complete:
> if what you want isn't described there, then SWISH++ doesn't do
> it.
ok, well, since you wrote the application, I was hoping you could
provide some advice on how to best configure swish++. That's why I
gave you my WordThreshold setting -- I hoped you could provide some
additional pointers.
swish++ is running my machine out of RAM/swap. I am surprised by this
behavior since the corpus is not that large. How large would you
expect the process size to be for a 80mb corpus? Is there a way to
minimize the memory footprint?
thank you
-- James
On Mon, 2 Aug 1999 pixel@... wrote:
> ok, well, since you wrote the application, I was hoping you could
> provide some advice on how to best configure swish++.
Whatever configuration advice is possible to be given by me has
already been given in the existing documentation. Why would I
intentionally withold configuration information by not placing
it in the documentation? Just so I could get asked these
questions over and over and over again?
> swish++ is running my machine out of RAM/swap. I am surprised by this
> behavior since the corpus is not that large.
Maybe the fault lies elsewhere. Maybe you don't have much
memory; maybe you have too little swap space; maybe you didn't
read the documentation for the -T option.
> How large would you expect the process size to be for a 80mb corpus?
I have no data that relates corpus size to memory usage.
> Is there a way to minimize the memory footprint?
Short of reducing Word_Threshold or using the -T option (if it
applies to your OS), no. Again, if there were, it would be in
the documentaion.
- Paul
Hi there,
I've download ed Swish++-3.0.2 and am tried to compile it on RedHat
6.0
with g++2.95. Everything compiles except file_info.c and file_list.c
The warnings start in include/exception starting around 16
The errors in file_info.h:28
Any thoughts or even binaries for the i686 platform would be great
-Peter Crowley
Peter:
I've always managed to compile it without errors using 2.91 under RedHat
6.0. Did you install 2.95 yourself? It's possible that there is a compiler
related error here.
Yeah... somebody else on the list has suggested we prepare an RPM of the
source and binaries, but it's a matter of sitting down and figuring out how
to do that, I've never done it myself...
Rob
-----Original Message-----
From: pcrowley@... <pcrowley@...>
To: swish@egroups.com <swish@egroups.com>
Date: Tuesday, August 17, 1999 2:27 PM
Subject: [swish] Compiling 3.0.2 with G++2.95 on RedHat 6.0
>
>Hi there,
> I've download ed Swish++-3.0.2 and am tried to compile it on RedHat
>6.0
>with g++2.95. Everything compiles except file_info.c and file_list.c
>The warnings start in include/exception starting around 16
>The errors in file_info.h:28
>
>Any thoughts or even binaries for the i686 platform would be great
>
>-Peter Crowley
>
>
>
>------------------------------------------------------------------------
>Listen to Britany spears and more top artists
>now at audiohighway.com!
>http://clickhere.egroups.com/click/395
>
>
>eGroups.com home: http://www.egroups.com/group/swish
>http://www.egroups.com - Simplifying group communications
>
>
>
On Tue, 17 Aug 1999 pcrowley@... wrote:
> I've download ed Swish++-3.0.2 and am tried to compile it on RedHat 6.0 with
> g++2.95. Everything compiles except file_info.c and file_list.c. The
> warnings start in include/exception starting around 16
Since "include/exception" is not part of SWISH++, I can not
address that. From the INSTALL.unix file, note 1:
Don't complain to me or bother asking me for
help if you get either "No such file or
directory" for or errors in the standard C++
headers. It means that your C++ compiler
and/or libraries are improperly installed. I
know nothing about your OS or how your system
is (mis)configured. Complain to your sysadmin:
the person who botched the installation and
whose job it is to fix it.
> The errors in file_info.h:28
If the errors are caused by the improper g++ installation, then
you're on your own.
- Paul
>> I've download ed Swish++-3.0.2 and am tried to compile it on RedHat >> 6.0
with g++2.95. Everything compiles except file_info.c and
>> file_list.c. The warnings start in include/exception starting
>> around 16
>
> Since "include/exception" is not part of SWISH++, I can not
> address that. From the INSTALL.unix file, note 1:
>
> Don't complain to me or bother asking me for
> help if you get either "No such file or
> directory" for or errors in the standard C++
> headers. It means that your C++ compiler
> and/or libraries are improperly installed. I
> know nothing about your OS or how your system
> is (mis)configured. Complain to your sysadmin:
> the person who botched the installation and
> whose job it is to fix it.
What a lovely man Paul is.
Paul,
I did not complain or even ask you for anything. I mailed the
public swish egroup for assistance. I would classify this response
as noise, not sure why you bothered actually. Yes while it was a newbie question
I thought some user have an insight.
I solved the problem by removing the line '#include "fake_ansi.h".
The package compiles fine under g++2.95 now. I can a diff file
if you are interested.
> The errors in file_info.h:28
If the errors are caused by the improper g++ installation, then
you're on your own.
- Paul
---- End Included Message ----
Peter Crowley
WebZone Limited
106 Lwr Baggot St., D 2
01 676 9347
pcrowley@...
_____________________________________
Get your free E-mail at http://www.ireland.com
On Fri, 20 Aug 1999, Peter Crowley wrote:
> I've download ed Swish++-3.0.2 and am tried to compile it on RedHat >> 6.0
> with g++2.95. Everything compiles except file_info.c and file_list.c. The
> warnings start in include/exception starting around 16
> I solved the problem by removing the line '#include "fake_ansi.h". The
> package compiles fine under g++2.95 now. I can a diff file if you are
> interested.
If it's a single-line diff, there isn't much point. As for it
compiling fine now, it doesn't makes sense since the file is
included anyway indirectly.
Simply removing it is not the answer since it *has* to be there
(directly or indirectly) since things in it are used in
file_info.c. The fake_ansi.h file is "included guarded," so its
extra inclusion shouldn't make any difference.
It's been reported that it compiles just find on other people's
gcc 2.95 (including my own).
Simply removing the line is not a correct solution.
- Paul
I have encountered an unusual problem with the SWISH++ feature that indexes=
the ALT tag of an IMG. Since the HTML pages in this particular website rea=
d ALT=” “, SWISH runs over it and indexes all the tags and their values that=
follow: for example, “WIDTH=”100” HEIGHT=”37” SRC=”../../shared/logo/gif." =
The descriptions on the results page are then cumbersome and awkward to read=
.
I would like to use the ExcludeClass feature in the configuration file to a=
void having SWISH index the IMG, and thus the ALT tag, but I am not at all c=
lear about the syntax to use. Just entering ALT and/or IMG on the ExlcudeCl=
ass line didn’t change the indexing behavior, nor did several variations I t=
ried of your example, <IMG SRC=”home.gif” ALT=”home” CLASS=”no_index”>. (I a=
m also unclear whether CLASS and no_index are metaphors for which I need to =
substitute other values.)
What would be the best approach to changing this indexing behavior?
On Mon, 23 Aug 1999 ehaymaker@... wrote:
> I have encountered an unusual problem with the SWISH++ feature that indexes
> the ALT tag of an IMG. Since the HTML pages in this particular website read
> ALT= , SWISH runs over it and indexes all the tags and their values that
> follow: for example, WIDTH=100 HEIGHT=37 SRC=../../shared/logo/gif."
Would you please resend your e-mail using exactly what is in
your web pages? The characters you typed in your e-mail are
not (or should not be) those used in yoru web pages. You typed
non-ASCII characters that can not be read except by others
using exactly the same software as you.
> I would like to use the ExcludeClass feature in the configuration file to
> avoid having SWISH index the IMG, and thus the ALT tag,
ExcludeClass has nothing to do with ignoring particular
elements. It has to do with ignoring *any* element that
contains a class attribute whose value contains at least one
excluded class.
> but I am not at all clear about the syntax to use.
The syntax is given, along with a real, working example, in the
manual page index(1) as well as in the HTML 4.0 specification
cited by the documentaion.
> Just entering ALT and/or IMG on the ExlcudeClass line didnt change the
> indexing behavior, nor did several variations I tried of your example, <IMG
> SRC=home.gif ALT=home CLASS=no_index>.
I think you need to go read the HTML 4.0 specification. The
SWISH++ ExcludeClass feature is based on standard HTML which you
are presumed to know. I didn't make this stuff up.
> (I am also unclear whether CLASS and no_index are metaphors for which I need
> to substitute other values.)
CLASS is not. Again, go learn HTML; "no_index" can be anything
you like: it was just an example. (If you knew how CLASS
worked, you wouldn't have asked this question, presumeably.)
Also, if "no_index" weren't a metaphor, then there would be
little point to the -C option.
- Paul
I have encountered an unusual problem with the SWISH++ feature that indexes
the ALT tag of an IMG. Since the HTML pages in this particular website read
ALT=" ", SWISH runs over it and indexes all the tags and their values that
follow: for example, "WIDTH="100" HEIGHT="37" SRC="../../shared/logo/gif."
The descriptions on the results page are then cumbersome and awkward to
read.
I would like to use the ExcludeClass feature in the configuration file to
avoid having SWISH index the IMG, and thus the ALT tag, but I am not at all
clear about the syntax to use. Just entering ALT and/or IMG on the
ExlcudeClass line didn't change the indexing behavior, nor did several
variations I tried of your example, <IMG SRC="home.gif" ALT="home" CLASS="
no_index">. (I am also unclear whether CLASS and no_index are metaphors for
which I need to substitute other values.)
What would be the best approach to changing this indexing behavior?
Since this e-mail is going into an archive, please don't give
it a Subject like "Resending Message." I've given this thread
a better Subject.
Now then...
On Mon, 23 Aug 1999, Eric Haymaker wrote:
> Since the HTML pages in this particular website read ALT=" ", SWISH runs over
> it and indexes all the tags and their values that follow: for example,
> "WIDTH="100" HEIGHT="37" SRC="../../shared/logo/gif."
You're not making this easy, are you? Give me an *EXACT*
example <IMG ...> element for which you believe SWISH++ is
doing the wrong thing.
> The descriptions on the results page are then cumbersome and awkward to read.
I'm confused: SWISH++ does not generate descriptions in its
results. It generates a single line per result as described in
search(1).
If you are using WWW::extract_description(), then *SAY* you are
using it. It makes a *BIG* difference as to where to look for
the problem.
> I would like to use the ExcludeClass feature in the configuration file to
> avoid having SWISH index the IMG, and thus the ALT tag, but I am not at all
> clear about the syntax to use.
I've already addressed this point in my earlier message.
- Paul
On Mon, 23 Aug 1999, Paul J. Lucas wrote:
> If you are using WWW::extract_description(), then *SAY* you are
> using it. It makes a *BIG* difference as to where to look for
> the problem.
Assuming you are using WWW::extract_description(), change line
103 from:
$s =~ s/<[^>]+?ALT\s*=\s*(['"])([^>]+)\1[^>]*?>/$2/gi;
to:
$s =~ s/<[^>]+?ALT\s*=\s*(['"])([^>]+?)\1[^>]*?>/$2/gi;
and let me know if the problem goes away.
- Paul
1) I EDITED LINE 103 OF WWW AS YOU SUGGESTED, WITH VASTLY IMPROVED
RESULTS, THOUGH WIDTH IS STILL BEGIN INDEXED. FOLLOWING IS WHAT APPEARS IN
THE "TOY" CGI SCRIPT.
100% Trash, heat, and ash (18.83203125)
" WIDTH= " WIDTH= Trash, HEAT, and ash The French are turning household
waste into energy, while developing technologies to control the residue. By
Michael Valenti, Senior Editor The French city of Nantes is the birthplace
of Jules Verne, the 19th century visionary whose novels predicted futuristic
wonders such as
85% Standing up to the test (12.5439453125)
" WIDTH= " WIDTH= Standing up to the test Crosslinking polyethylene can
greatly reduce wear in hip implants. By John DeGaspari, Associate Editor
Artificial hip joints must perform reliably for many years of use and
millions of cycles. A typical patient with a hip replacement joint may take
as many
2) ONE EXAMPLE OF THE HTML SOURCE THAT I INDEXED IS AS FOLLOWS:
<TR>
<TD align=right valign=top width=100><A
HREF="http://www.memagazine.org/index.html"><IMG
BORDER="0" ALT="" WIDTH="100" HEIGHT="37"
SRC="../../shared/littlelogo6.gif"></A></TD>
<TD align=left width=380><MAP name="navContentsMap">
<AREA shape="rect" coords="0,0,93,24" HREF="../feat_toc.html">
<AREA shape="rect" coords="94,0,185,24"
HREF="../../departments/dept_toc.html">
<AREA shape="rect" coords="187,0,278,24"
HREF="../../marketplace/market_toc.html">
<AREA shape="rect" coords="277,0,369,24"
HREF="http://www.memagazine.org/contents/newsdigest/weekly.html""">
</MAP>
<IMG BORDER="0" SRC="../../shared/navContents.gif"
USEMAP="#navContentsMap"></TD>
</TR>
<TR>
<TD align=right valign=top width=100><IMG ALT="" WIDTH="100" HEIGHT="16"
SRC="../../shared/feature2.gif"></TD>
<TD align=left width=380><BR>
<FONT size=6><B><font color="#0085B0">Trash,</font><font
color="#CE0025"> HEAT,</font><font color="#0085B0"> and ash</font>
</B></FONT>
<BR>
<BR>
<FONT size=4 color="#000000">The French are turning household waste
into energy, while developing technologies to control the residue.
</FONT><BR>
<BR>
</TD>
3) THE ORIGINAL RESULTS WERE AS FOLLOWS:
100% Trash, heat, and ash (18.83203125)
" WIDTH="100" HEIGHT="37" SRC="../../shared/littlelogo6.gif " WIDTH="100"
HEIGHT="16" SRC="../../shared/feature2.gif Trash, HEAT, and ash The French
are turning household waste into energy, while developing technologies to
control the residue. By Michael Valenti, Senior Editor The French city of
Nantes is the birthplace of Jules Verne, the 19th century visionary whose
novels predicted
85% Standing up to the test (12.5439453125)
" WIDTH="100" HEIGHT="37" SRC="../../shared/littlelogo6.gif " WIDTH="100"
HEIGHT="16" SRC="../../shared/feature2.gif Standing up to the test
Crosslinking polyethylene can greatly reduce wear in hip implants. By John
DeGaspari, Associate Editor Artificial hip joints must perform reliably for
many years of use and millions of cycles. A typical patient with a hip
replacement joint
-- END --
On Tue, 24 Aug 1999, Eric Haymaker wrote:
> 1) I EDITED LINE 103 OF WWW AS YOU SUGGESTED, WITH VASTLY IMPROVED
> RESULTS, THOUGH WIDTH IS STILL BEGIN INDEXED.
1. Do not type in all capital letters.
2. Let's get the terminology straight: "WIDTH" is *NOT* being
"indexed." You can use WWW::extract_description() on *ANY*
HTML file. Whether said file has been indexed is
irrelevant. Therefore, there is *NOT* a bug in the indexer.
Are we clear yet?
Try changing line 103 to:
$s =~ s/<[^>]+?ALT\s*=\s*(['"])([^>]*?)\1[^>]*?>/$2/gi;
- Paul
Thank you for your kind assistance!
Changing line 103 to:
$s =~ s/<[^>]+?ALT\s*=\s*(['"])([^>]*?)\1[^>]*?>/$2/gi;
works fine, and I've learned a lot.
-- Eric
I have managed to get swish++ compiled under Irix using g++. I have several problems so far for which I haven't found answers in the docs. Perhaps someone on the group can point me to the right direction for fixing them or what to check.
1) if I try to uncomment and use ExcludeClass index will crash with a "Segmentation Fault: core dump" message.
2) Even if I do not use -r, index seems only to index the files in the top directory. I have the binaries in ~/public_html/cgi-user/swish and I am running it from there using (index ../..) It only indexes the files in the top directory (public_html). FWIW, swish++.conf has RecurseSubdirs set to "yes"
3) A request since this doesn't seem to be in swish++... but it would be nice to be able to filter what files are indexed by substrings in their names as well as by extension. For instance... avoid all test*.htm files.
Let me know what additional details are useful if you can help.
I have managed to get swish++ compiled under Irix using g++. I have several problems so far for which I haven't found answers in the docs. Perhaps someone on the group can point me to the right direction for fixing them or what to check.
1) if I try to uncomment and use ExcludeClass index will crash with a "Segmentation Fault: core dump" message.
2) Even if I do not use -r, index seems only to index the files in the top directory. I have the binaries in ~/public_html/cgi-user/swish and I am running it from there using (index ../..) It only indexes the files in the top directory (public_html). FWIW, swish++.conf has RecurseSubdirs set to "yes"
3) A request since this doesn't seem to be in swish++... but it would be nice to be able to filter what files are indexed by substrings in their names as well as by extension. For instance... avoid all test*.htm files.
Let me know what additional details are useful if you can help.
Do not submit questions in HTML. Please use only plain text.
On Tue, 24 Aug 1999 jwilkinson@... wrote:
> 1) if I try to uncomment and use ExcludeClass index will crash with a
> "Segmentation Fault: core dump" message.
1. Which version of SWISH++? If it's anything earlier than
3.0.2, upgrade to 3.0.2 first.
2. Assuming you are running 3.0.2, you *MUST* provide a sample
HTML file that causes the core dump.
>2) Even if I do not use -r, index seems only to index the files in the top
>directory. I have the binaries in ~/public_html/cgi-user/swish and I am
>running it from there using (index ../..) It only indexes the files in the top
>directory (public_html). FWIW, swish++.conf has RecurseSubdirs set to "yes"
1. Repeat of #1 above.
>3) A request since this doesn't seem to be in swish++... but it would be
>nice to be able to filter what files are indexed by substrings in their names
>as well as by extension. For instance... avoid all test*.htm files.
There is no need to provide this feature. Reread the use of
'-' to read files off of standard input.
- Paul
I just downloaded swish++, so it is v3.0.2.
On further testing to narrow things down to one html file that causes it to
crash on ExcludeClass, I'm finding something interesting.
My website root is
/disk2/people/central/public_html
I have the binaries, index file and conf file in
/disk2/people/central/public_html/cgi-user/swish
if I run from the swish subdir and use
index ../..
or index ../../*.*
or index ./../*.* -r
it will run ok without ExcludeClass, but will crash with it in swish++.conf
if I run from there and use
index ../../*.htm
or index ../../*.html
it will run fine, even with ExcludeClass.
At the level of the top directory (public_html) there are 9 htm files, 1
html file and a robots.txt file. I have IncludeExtension set to only htm
and html files. I'm not sure what it's seeing that is causing the crash.
I'll attach some of the pertinent files.
Either way I do this, it still also refuses to recurse subdirs.
--
Jeff Wilkinson
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<<swish++.conf>> <<config.mk>> <<CONTACT.HTM>> <<CHRIST.HTM>>
<<CoreValues.htm>> <<Index.html>> <<IntOrganist.htm>> <<ROBOTS.TXT>>
<<STAFF.HTM>> <<staffpic.htm>> <<Visiting.htm>> <<WELCOME.HTM>>
<<WhoAreWe.htm>>
The Mission of Central Presbyterian Church is:
MOVING PEOPLE TOWARD CHRIST
by being a community of faith which loves,
encourages, and equips them in Christ,
sending them out to serve.
A community of faith based on the Lordship of
Jesus Christ and the Authority of Scripture.
Welcome to the website of the Central Presbyterian Church
in Baltimore, Maryland, an active, warm, growing, evangelical
church, striving to be faithful to Jesus Christ, our Lord and Savior!
7308 York Road - Baltimore, MD 21204
Phone:(410) 823-6145 - Fax:(410) 494-1169
11:00 Service Broadcast Live on AM1230, WITH
Some additional info about my problem with recursion of subdirs.
If I index using -v4, it gives me a line for each subdir saying
"(skipped: not plain file)"
I'm using SGI Irix v5.3. Is there some reason it should be
unable to tell that a directory is a directory?
(this is all on www.centralpc.org)
-----Original Message-----
From: jwilkinson@... <jwilkinson@...>
To: swish@egroups.com <swish@egroups.com>
Date: Wednesday, August 25, 1999 6:30 PM
Subject: [swish] Re: ExcludeClass core dumps
>Some additional info about my problem with recursion of subdirs.
>If I index using -v4, it gives me a line for each subdir saying
> "(skipped: not plain file)"
>
>I'm using SGI Irix v5.3. Is there some reason it should be
>unable to tell that a directory is a directory?
>
I came across a similar problem a while back... it was related to the stat()
function, if I'm not mistaken. This was not on a Unix box, however, I
encountered it while working on the win32 port of swish++.
Rob
Note: you are reporting TWO SEPARATE PROBLEMS with a SINGLE
Subject. Bad. I have therefore given THIS problem a new
Subject.
On Wed, 25 Aug 1999 jwilkinson@... wrote:
> Some additional info about my problem with recursion of subdirs.
> If I index using -v4, it gives me a line for each subdir saying
> "(skipped: not plain file)"
>
> I'm using SGI Irix v5.3. Is there some reason it should be unable to tell
> that a directory is a directory?
If directory.c line 104 reads:
if ( is_directory( path ) && recurse_subdirectories )
then the answer is "no" and it's not obvious to me why it
doesn't work on your system alone.
- Paul
On Wed, 25 Aug 1999, Wilkinson, Jeffrey A. wrote:
> index ../..
> or index ../../*.*
> or index ./../*.* -r
> it will run ok without ExcludeClass, but will crash with it in swish++.conf
>
> if I run from there and use
> index ../../*.htm
> or index ../../*.html
> it will run fine, even with ExcludeClass.
>
> At the level of the top directory (public_html) there are 9 htm files, 1
> html file and a robots.txt file. I have IncludeExtension set to only htm
> and html files.
And how many subdirectories? What are their names?
> Either way I do this, it still also refuses to recurse subdirs.
In what you gave above, only:
index ../..
is guaranteed to index subdirectories if any exist. If you
specify only files on the command line such as you did when you
did:
index ../../*.htm
then NATURALLY it won't recurse subdirectories since you never
told it to index any directories (which are the only things
that can have subdirectories) in the first place.
The other case of:
index ../../*.*
will recurse subdirectories ONLY for directories that contain a
dot in their names. If you have no such directories, then,
again, NATURALLY it won't recurse since you didn't tell it to
index any directories.
- Paul
P.S.: IMHO, it's an odd idea to intall SWISH++ under your
document root.
On Wed, 25 Aug 1999, Wilkinson, Jeffrey A. wrote:
> On further testing to narrow things down to one html file that causes it to
> crash on ExcludeClass, I'm finding something interesting.
And, pray tell, which file would that be of the ones you sent?
That's the only one I care about.
- Paul
On Wed, 25 Aug 1999, Wilkinson, Jeffrey A. wrote:
> On further testing to narrow things down to one html file that causes it to
> crash on ExcludeClass, I'm finding something interesting.
Another thing: recompile with -g on and tell me WHERE it
crashes using a debugger and the 'where' command.
FYI: I just indexed your HTML files and it didn't crash.
- Paul
P.S.: Your .HTM files are not being indexed: extensions are
case-sensitive. This is documented.