KDE help: Search uses different language prefix

Bug #115644 reported by Cictani
6
Affects Status Importance Assigned to Milestone
KDE Base
Fix Released
Medium
KDE PIM
Invalid
Undecided
Unassigned
kdebase-runtime (Ubuntu)
Triaged
Undecided
Unassigned

Bug Description

Hi,

If I search with KDE help I get the following output:

khc_htsearch.pl --docbook --indexdir=/home/andreasw/.kde/share/apps/khelpcenter/index/ --config=kde_application_manuals --words=dddd --method=and --maxnum=5 --lang=en

As you can see it uses en as prefix but it should use de (because my kde is localized for german). This bug also occurs if you create a new search index. It will only index documentation in the "en" directory.

Best regards

Cictani

Revision history for this message
In , Koolniczka (koolniczka) wrote :

Version: (using KDE KDE 3.5.1)
Installed from: Slackware Packages
Compiler: 3.3.6
OS: Linux

There are similar bug #103266 and #102912 but i wish to report it for the 3.5.1.

Anyway the problem is widespread in many distributions and i experience it on Slack 10.2 with KDE3.5.1.

The kde_application_manuals index building routine is utterly broken. It does not work either with
htdig 3.1.6 -- fails in the htmerge step
nor with
htdig 3.2.0b6. -- routine finishes with the errors described above and no usable index is built.

The first problem is that the htdig PATHs hardcoded (sic!) in the khc_docboodig.pl wrapper (called from khelpcenter) are "distro specific" at best (not to say absolutely bogus) and the script doesnt even honor sytem wide paths. Hardcoded assorted locales don't do much good either.

Other problem is that htdig local URL substitution appears to be broken or the use of it in this wrapper isn't appropriate (for either versions), the htdig 3.2.0b6 doesn't return any errors upstream, finishes correctly but increased verbosity revaeals errors such as previously described here

http://bugs.kde.org/attachment.cgi?id=11431&action=view

and the subsequent search doesn't return any results (although the index is reportedly built).

Maybe finaly fixing this (reports since 2003?) would be worth a the effort because it simply "Doesn't Work" (tm)

Revision history for this message
In , Dofjmwxtke (dofjmwxtke) wrote :

Yes, yes,

Please do fix this. So very annoying to get that message:

Creating index for 'kde_application_manuals'
htdig failed

Truly, us "stupid" users might adopt Linux more readily if things like the "Documentation" system actually worked out of the box and we didn't have to spend an hour plus tracking down information on a bug that's been know to exist for the last three years. . .

Mandriva user

Revision history for this message
In , Wapcaplet99 (wapcaplet99) wrote :

Agreed, this has been broken on Gentoo also, for as long as I can remember.

Revision history for this message
In , DominikBodi (dominik-bodi) wrote :

*** This bug has been confirmed by popular vote. ***

Revision history for this message
In , Metrobotte (metrobotte) wrote :

Kde 3.5.5 also suffer this bug!
:(

Revision history for this message
In , Metrobotte (metrobotte) wrote :

This bug is solved on slackware
you need to edit
2 files /opt/kde/bin/khc_htdig.pl
and /opt/kde/bin/khc_docbookdig.pl
and edit the variables,should be like this

my $htdigdata = "/opt/www/htdig/common/";
my $htdigbin = "/opt/www/htdig/bin";
my $kdeprefix = "/opt/kde";

Revision history for this message
In , Metrobotte (metrobotte) wrote :

I forgot..
and edit the file /opt/kde/bin/khc_docbookdig.pl

my $htsearchpath="/opt/www/htdig/cgi-bin/htsearch";

Htdig MUST BE the latest 3.20b6

Revision history for this message
In , Me-ngeefk4xayt3t4u9watah405veq (me-ngeefk4xayt3t4u9watah405veq) wrote :

*** Bug 112518 has been marked as a duplicate of this bug. ***

Revision history for this message
In , Me-ngeefk4xayt3t4u9watah405veq (me-ngeefk4xayt3t4u9watah405veq) wrote :

*** Bug 136043 has been marked as a duplicate of this bug. ***

Revision history for this message
In , Me-ngeefk4xayt3t4u9watah405veq (me-ngeefk4xayt3t4u9watah405veq) wrote :

*** Bug 125192 has been marked as a duplicate of this bug. ***

Revision history for this message
Cictani (bla-bla-deactivatedaccount) wrote :

Not only the search command but already the index find command uses en as language instead of de:

FINDCMD: find /usr/share/doc/kde/HTML/en/ -name index.docbook

Revision history for this message
Thomas David (nemphis) wrote :

Same problem here. I didn't find any config files or options to change this attitude.

Revision history for this message
Emanuel Goscinski (emu--deactivatedaccount) wrote :

I can confirm this to.
This makes the search function worthless, for german users.

Kubuntu 7.04
Kde 3.5.6

Revision history for this message
Robert Frießleben (robertfriessleben) wrote :

same here
$ locale
LANG=de_DE.UTF-8
LANGUAGE=de_DE:de:en_GB:en
LC_CTYPE="de_DE.UTF-8"
LC_NUMERIC="de_DE.UTF-8"
LC_TIME="de_DE.UTF-8"
LC_COLLATE="de_DE.UTF-8"
LC_MONETARY="de_DE.UTF-8"
LC_MESSAGES="de_DE.UTF-8"
LC_PAPER="de_DE.UTF-8"
LC_NAME="de_DE.UTF-8"
LC_ADDRESS="de_DE.UTF-8"
LC_TELEPHONE="de_DE.UTF-8"
LC_MEASUREMENT="de_DE.UTF-8"
LC_IDENTIFICATION="de_DE.UTF-8"
LC_ALL=

$ kde-config --version
Qt: 3.3.7
KDE: 3.5.7
kde-config: 1.0

$ lsb_release -d
Description: Ubuntu 7.04

Revision history for this message
Robert Frießleben (robertfriessleben) wrote :

I think, the Problem for creating index is in
/usr/bin/khc_docbookdig.pl on line 59
---8<-----------------------------
if ( !$lang ) { $lang = "en"; }
----------------------------->8---
because --lang will not be set when calling khc_docbookdig.pl for creating a searchindex.
This could be solved by retrieving the userlanguage from %ENV.

But in Search there is another Problem, because --lang is set to en, so even if you indexed the right HTML files (in my case "de") khc_docbookdig.pl would not search in it.

Revision history for this message
Emanuel Goscinski (emu--deactivatedaccount) wrote :

Even there for KDE 3.5.7 on fresh and updated Gutsy Gibbon

Revision history for this message
In , Emanuel Goscinski (emu--deactivatedaccount) wrote :

This exists further in 3.5.7 on Kubuntu Gutsy Gibbon.
This is also the reason, why indexing fails when having a different localization then english.

For example: Error-Message for german localization:

INDEXDIR: /home/emu/.kde/share/apps/khelpcenter/index/
FINDCMD: find /usr/share/doc/kde/HTML/en/ -name index.docbook
Creating index for 'kde_application_manuals'
htdig failed

Changed in kdebase:
status: Unknown → Confirmed
Changed in kdepim:
status: Unknown → Confirmed
Changed in kdebase:
status: New → Confirmed
Revision history for this message
Jonathan Thomas (echidnaman) wrote :

Still an issue in Intrepid?

Changed in kdebase:
status: Confirmed → Incomplete
Changed in kdepim:
importance: Unknown → Undecided
status: Confirmed → New
status: New → Invalid
Revision history for this message
Christian Prim (christian-prim) wrote :

Yes!

Same problem in intrepid.

Changed in kdebase-runtime:
status: Incomplete → Triaged
Revision history for this message
In , Myriam Schweingruber (myriam) wrote :

Is this bug still relevant for KDE4? No working help system would be really bad...

Revision history for this message
In , David Faure (faure) wrote :

"htdig failed" means that htdig was not found. More precisely, it's currently looking for /usr/bin/htdig. I added a better error message now (cf bug 146010), but if htdig can come from other directories then the hardcoded /usr/bin/htdig is bad indeed.
The question is: for those of you who have it in e.g. /opt/www/htdig/bin, is that directory in the $PATH? And if not, how is KDE supposed to find it? :-)

Revision history for this message
In , David Faure (faure) wrote :

SVN commit 1025097 by dfaure:

Look for htsearch not only in /srv/www/cgi-bin (SUSE-specific?) but also in /usr/lib/cgi-bin
BUG: 126709
CCBUG: 122437

 M +2 -1 searchengine.cpp
 M +9 -9 searchengine.h
 M +6 -2 searchhandler.cpp
 M +1 -0 searchhandler.h
 M +3 -1 searchhandlers/docbook.desktop.cmake
 M +3 -1 searchhandlers/htdig.desktop.cmake
 M +1 -1 searchhandlers/khc_htsearch.pl

WebSVN link: http://websvn.kde.org/?view=rev&revision=1025097

Revision history for this message
In , David Faure (faure) wrote :

SVN commit 1025099 by dfaure:

Also look in /opt/www/htdig/bin (for Slackware, according to comment #5)
CCBUG: 122437

 M +1 -1 docbook.desktop.cmake
 M +1 -1 htdig.desktop.cmake

WebSVN link: http://websvn.kde.org/?view=rev&revision=1025099

Revision history for this message
In , David Faure (faure) wrote :

*** Bug 146579 has been marked as a duplicate of this bug. ***

Revision history for this message
In , Ilgaarsoft (ilgaarsoft) wrote :

I got the same problem with OpenSUSE 11.3, I get this output when I try to search for something.

op-Level Documentation
Application Manuals
openSUSE manual: Security Guide Documentation (en)
 Htdig error:
Unable to read word database file '/home/me/.kde4/share/apps/khelpcenter/index//opensuse-manuals_en.words.db'
Did you run htdig?

Revision history for this message
In , Digulla-hepe (digulla-hepe) wrote :

On Ubunutu 10.10 with KDE 4.5.4, I get no error after running htdig but it also doesn't return any results. When I run htdig, it seems to index something:

htdig: 44 servers seen:
htdig: amarok:80 1 document
htdig: ark:80 1 document
htdig: dolphin:80 1 document
htdig: dragonplayer:80 1 document
htdig: gwenview:80 1 document
htdig: kate:80 1 document
...

but that barely scratches the number help documents I have installed. Since "amarok" was indexed, I searched for two words that I copied from the help document: "music player"

No results. It seems that searching the help is currently broken.

Changed in kdebase:
importance: Unknown → Medium
Changed in kde-baseapps:
status: Confirmed → Unknown
Revision history for this message
In , Pino Toscano (pinotree) wrote :

Git commit 583ed246c75ce2c686957b6e1dc34aaed31206b4 by Pino Toscano.
Committed on 12/03/2016 at 23:01.
Pushed by pino into branch 'master'.

Add Xapian support for indexing and searching

Start using Xapian for indexing and searching the documentation: it
works much better than htp://dig, it provides a C++ API, and it is well
maintained.

Introduce two helper tools to create/update an index, and search on it,
starting from the cache.bz2 files generated by meinproc: this way it is
possible to index the whole text, with the entities already expanded
(something which would not be there when parsing the docbook files
directly). Make use of libxml2 to parse the HTML documents, and get all
the text in them.

Xapian and libxml2 are currently considered as mandatory dependencies:
they are portable, generally available everywhere, and not requiring
themselves too many extra dependencies.
Related: bug 126710, bug 93664, bug 103266, bug 115935, bug 125276, bug 152671, bug 158633, bug 209415, bug 256397, bug 266290

M +13 -0 CMakeLists.txt
M +35 -0 searchhandlers/CMakeLists.txt
A +134 -0 searchhandlers/cachereader.cpp [License: GPL (v2+)]
A +49 -0 searchhandlers/cachereader.h [License: GPL (v2+)]
A +107 -0 searchhandlers/htmltextdump.cpp [License: GPL (v2+)]
A +29 -0 searchhandlers/htmltextdump.h [License: GPL (v2+)]
A +5 -0 searchhandlers/xapian.desktop.cmake
A +126 -0 searchhandlers/xapiancommon.cpp [License: GPL (v2+)]
A +50 -0 searchhandlers/xapiancommon.h [License: GPL (v2+)]
A +269 -0 searchhandlers/xapianindexer.cpp [License: GPL (v2+)]
A +151 -0 searchhandlers/xapiansearch.cpp [License: GPL (v2+)]

http://commits.kde.org/khelpcenter/583ed246c75ce2c686957b6e1dc34aaed31206b4

Revision history for this message
In , Pino Toscano (pinotree) wrote :

Git commit 163133768d77ae16e36b31d1174d6c60d53cfdce by Pino Toscano.
Committed on 12/03/2016 at 23:01.
Pushed by pino into branch 'master'.

Remove the ht://dig support

ht://dig is basically unmantained for the past decade, and its
integration in khelpcenter has always been a sort of hack (execute
a CGI, and print its resulting HTML text).

Since Xapian is used as better solution, all of the htp://dig
supporting code can be removed completely.
Related: bug 125276, bug 152671, bug 158633, bug 209415, bug 244091, bug 255384, bug 256397, bug 277464, bug 103266, bug 112958

M +0 -2 CMakeLists.txt
D +0 -14 README.htdig
M +0 -14 doc/khelpcenter/index.docbook
M +0 -9 docmetainfo.cpp
M +0 -3 docmetainfo.h
D +0 -61 htmlsearch.cpp
D +0 -51 htmlsearch.h
D +0 -47 htmlsearch/CMakeLists.txt
D +0 -5 htmlsearch/Messages.sh
D +0 -23 htmlsearch/checked.xpm
D +0 -470 htmlsearch/htmlsearch.cpp
D +0 -254 htmlsearch/htmlsearch.desktop
D +0 -62 htmlsearch/htmlsearch.h
D +0 -25 htmlsearch/index.cpp
D +0 -371 htmlsearch/kcmhtmlsearch.cpp
D +0 -84 htmlsearch/kcmhtmlsearch.h
D +0 -52 htmlsearch/klangcombo.cpp
D +0 -47 htmlsearch/klangcombo.h
D +0 -246 htmlsearch/ktagcombobox.cpp
D +0 -92 htmlsearch/ktagcombobox.h
D +0 -6 htmlsearch/long.html
D +0 -4 htmlsearch/meinproc_wrapper
D +0 -21 htmlsearch/nomatch.html
D +0 -89 htmlsearch/progressdialog.cpp
D +0 -32 htmlsearch/progressdialog.h
D +0 -1 htmlsearch/short.html
D +- -- htmlsearch/star.png
D +- -- htmlsearch/star_blank.png
D +0 -19 htmlsearch/syntax.html
D +0 -22 htmlsearch/unchecked.xpm
D +0 -16 htmlsearch/wrapper.html
D +0 -156 htmlsearchconfig.cpp
D +0 -58 htmlsearchconfig.h
M +0 -1 kcmhelpcenter.cpp
M +0 -2 kcmhelpcenter.h
M +2 -14 searchhandlers/CMakeLists.txt
D +0 -8 searchhandlers/docbook.desktop.cmake
D +0 -8 searchhandlers/htdig.desktop.cmake
D +0 -2 searchhandlers/htdig_long.html
D +0 -222 searchhandlers/khc_docbookdig.pl.cmake
D +0 -148 searchhandlers/khc_htdig.pl.cmake
D +0 -117 searchhandlers/khc_htsearch.pl
M +1 -2 tests/CMakeLists.txt

http://commits.kde.org/khelpcenter/163133768d77ae16e36b31d1174d6c60d53cfdce

Changed in kde-baseapps:
status: Unknown → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.