According to Abdullah Ramazanoglu:
> Since this is a quick feedback mail, I didn't subscribe to any
> list. Sorry if you feel your inbox is being littered.
> Also, instead of sending a mail to each party, I'm sending a
> composite mail to all parties at once, so some parts may not be
> of interest to you.
No problem. We appreciate this sort of feedback, even though in
this case these are mostly known problems. Most of these are fixed
in the htdig-3.2.0b4-081901.tar.gz snapshot. Please give it a try.
(The 082601 snapshot seems to be truncated, so use the next to last one,
which is the most recent working version.)
You can still post to htdig-general, or even htdig-dev (subject to
moderator approval), even if you're not subscribed, so that's still the
best place to post htdig-related correspondance.
> htdig= 3.2.0b3
>
> cygwin= 1.3.2-1 and latest set as of Aug 2001 especially,
> gcc-2.95.3-5 and peer libs, flex-2.5.4-1, autoconf-2.52-1,
> make-3.79.1-4
...
> Modifications to htdig :
>
> ==============================================================
>
> 1. htsearch/Display.cc :
> Add,
> #include <limits.h>
> #include <float.h>
>
> Notes: This step is necessary for both RedHat and Cygwin
> environments. RedHat solves (in their src.rpm) this by just
> adding "#include <values.h>" but values.h is being phased out,
> (limits.h and float.h recommended instead) and does not compile
> in cygwin environment. (Cygwin has a "dummy" values.h) The
> following steps are necessary only for Cygwin environment.
Fixed in the recent snapshot. At least, float.h is added. Is limits.h
actually needed on Cygwin?
> ==============================================================
>
> 2. htfuzzy/EndingsDB.cc :
>
> Comment-out the following :
> // This is an attempt to get around compatibility problems
> // with the included regex
> //
> #ifdef HAVE_BROKEN_REGEX
> //
> #include <regex.h>
> //
> #else
> //
> #include "regex.h"
> //
> #endif
>
> And add this :
> #include "../htlib/regex.h"
>
> ==============================================================
>
> 3. htfuzzy/Fuzzy.cc :
>
> Instead of:
> #include "Regex.h"
>
> Use:
> #include "../htlib/Regex.h"
> #include "Regex.h"
>
> ==============================================================
>
> 4. htfuzzy/Regex.cc :
>
> Instead of:
> #include "Regex.h"
>
> Use:
> #include "../htlib/Regex.h"
> #include "Regex.h"
These 3 problems should be fixed in the recent snapshot. The cause
was the ambiguity between regex.h and Regex.h on a case-insensitive
file system. The latter header has been renamed to Regexp.h.
> ==============================================================
>
> 5. htdb/htdb_load.cc & htdb/htdb_dump.cc & htdb/htdb_stat.cc :
>
> In each of these 3 programs, add the following to the end of
> "include" block:
>
> #ifdef HAVE_GETOPT_H
> #include <getopt.h>
> #endif
Fixed in the recent snapshot.
> ==============================================================
> ==============================================================
>
> If you're not interested in htdig please be informed that what
> follows on is relevant to htdig only.
>
> And some bugs in htdig:
>
> 1- Sort-by-title does not sort. It is not related to locale
> settings, because my locale (tr_TR) works properly, and it does
> not sort in C or en_US locale as well.
Fixed in the recent snapshot.
> ==============================================================
>
> 2- In templates, $(EXCERPT) HTML escaping works in a different
> (and unwanted to me) way: It HTML escapes when not being asked
> to. And it HTML-escapes the HTML code itself when asked to. Example:
>
> $(TITLE) correctly produces: "This is � umlaut & � cedilla"
> $&(TITLE) correctly produces: "This is ü umlaut &
> ç cedilla"
> which are what we would expect.
>
> But,
> $(EXCERPT) produces: "This is ü umlaut & ç cedilla"
> and,
> $&(EXCERPT) produces : "This <b>is</b> ü umlaut
> & ç cedilla"
>
> Which are unexpected. As it is, $(EXCERPT) is behaving as if
> $&(EXCERPT), and $&(EXCERPT) is html escaping the html code itself.
>
> I toyed around Display.cc to hard-disable SGML encoding totally,
> without luck so far. (It mangles up some local characters. They
> are defined in my locale, and gets html-escaped allright, but
> not every browser understand html-escapes for local characters,
> so shows garbage) Disabling html-escaping in all other fields
> {i.e. use $(TITLE) instead of $&(TITLE)} solved the problem, but
> disabling it in EXCERPT is not possible apparently. Any
> suggestions are welcome)
The EXCERPT template variable can't be used as $&(EXCERPT) because
it may contain HTML tags in it, which you don't want encoded as,
e.g. <strong>. So, you must use it as $(EXCERPT), but to prevent
any unencoded "<", "&", etc. from being directly inserted in with the
HTML, the plain text parts of the excerpt are SGML encoded.
The mangling of accented characters in non-ISO-8859-1 locales,
by htsearch's SGML encoding, is a known problem, but we don't have a
solution as yet. The only fix right now is to hack the HtSGMLCodec class
not to convert any locale-specific accented characters (those above 127),
but to still convert the plain ASCII characters that need it. I think
we need to add a configuration attribute to control whether this is done.
--
Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]>
Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil
Dept. Physiology, U. of Manitoba Phone: (204)789-3766
Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930
_______________________________________________
htdig-general mailing list <[EMAIL PROTECTED]>
To unsubscribe, send a message to <[EMAIL PROTECTED]> with a
subject of unsubscribe
FAQ: http://htdig.sourceforge.net/FAQ.html
