-----Original Message----- From: openguides-dev-bounces@openguides.org [mailto:openguides-dev-bounces@openguides.org]On Behalf Of Earle Martin Sent: 27 January 2004 10:37 To: openguides-dev@openguides.org Subject: [OpenGuides-Dev] OpenGuides on Windows?
I've received this message from an interested party:
I am a chinese perl lover and I love this open guides and want to have a test on my computer. But can't install. My system is windows xp, and does it can't be installed on this system? If can, would you please tell me some tips for installation?
Issues of using a Windows XP box for web hosting aside, is running OG on Windows a realistic possibility? I know there are things like ActivePerl, but I'm not familiar with them at all.
Search::InvertedIndex is going to be the problem. Fails tests on Win32 big time!
I did have a go at putting OG and CGI::Wiki::Kwiki up on my Win98 box, but gave up when finding it couldn't handle S::II
On Tue, Jan 27, 2004 at 05:42:30AM -0500, IvorW wrote:
I did have a go at putting OG and CGI::Wiki::Kwiki up on my Win98 box, but gave up when finding it couldn't handle S::II
Hmm. We really do have issues with that one, don't we.
I just got this reply from my correspondent:
Thank you for your reply. For these time I am just looking the CGI::Wiki module code. Because the search function is based on other full text inverted search modules, and those are for english or other european user, so I must modify some code for suit for chinese :( I am stuck on this project and want do an open guides on china city. And I hope to get you help and get more detail material. Thank you.
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
Search::InvertedIndex is going to be the problem. Fails tests on Win32 big time!
Is there any chance you could try to do something about this? Jerry (its author) doesn't have access to a Windows box and neither do I. It looks like you are in an excellent position to really help us out. If you could get even one extra test passing and send him a patch, that would be a start.
Kake
----- Original Message ----- From: "Kate L Pugh" kake@earth.li To: "Discussion of development on the OpenGuides software" openguides-dev@openguides.org Sent: 27 January 2004 18:38 Subject: Re: [OpenGuides-Dev] OpenGuides on Windows?
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
Search::InvertedIndex is going to be the problem. Fails tests on Win32 big time!
Is there any chance you could try to do something about this? Jerry (its author) doesn't have access to a Windows box and neither do I.
Congratulations! Linus would be proud of you :).
It looks like you are in an excellent position to really help us out. If you could get even one extra test passing and send him a patch, that would be a start.
I did have a go, but the test script supplied: t/base.t is arcane to say the least.
When I try to run it through the debugger (make testdb), it hits an evil eval and spews tons of crap while dieing. Not very helpful.
It's failing all the tests. It's barffing in test1 at line 2108:
my $inv_map = Search::InvertedIndex->new({ -database => $database });
but I haven't got to the bottom of where the problem is.
Maybe it might be possible to rewrite t/base.t in a more friendly fashion, but I don't know if the issues are with the test, with the module, or the environment :(.
In my opinion, t/base.t needs replacing with a more comprehensive and comprehensible set of tests. Unfortunately, I don't feel I understand enough of the module and how it works, to take this on. Maybe, given time, particularly, if I want to do a search index module of my own, I might tackle this.
Sorry not to be of more help.
Ivor.
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
Search::InvertedIndex is going to be the problem. Fails tests on Win32 big time!
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
I did have a go, but the test script supplied: t/base.t is arcane to say the least.
Damn, damn, damn, I *know* I rewrote at least half of the tests to be more modular, and now I can't find the CVS repository I left them in! I did find one, which I mailed to Jerry as an example. It's attached - see if you can get that passing.
Kake
At this point I'd like to introduce Wei Litao to the list, my correspondent from China. Welcome aboard!
----- Original Message ----- From: "Kate L Pugh" kake@earth.li To: "Discussion of development on the OpenGuides software" openguides-dev@openguides.org Sent: 28 January 2004 17:57 Subject: Re: [OpenGuides-Dev] OpenGuides on Windows?
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
Search::InvertedIndex is going to be the problem. Fails tests on Win32 big time!
On Tue 27 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
I did have a go, but the test script supplied: t/base.t is arcane to say the least.
Damn, damn, damn, I *know* I rewrote at least half of the tests to be more modular, and now I can't find the CVS repository I left them in! I did find one, which I mailed to Jerry as an example. It's attached
- see if you can get that passing.
Still blows up. Here is the offending line in the debug:
DB<1> n Search::InvertedIndex::DB::DB_File_SplitHash::_open_multi_map(blib\lib/Search/In vertedIndex/DB/DB_File_SplitHash.pm:677): 677: CORE::open ($fh, "+<&=$fd") or croak (__PACKAGE__ . "::_open_multi_m ap() - unable to open file descriptor for locking: $!"); DB<1> p $fd -1 DB<2>
Unfortunately when I step, the information scrolls off, so I can't see what $! contained :(. (windows)-- #No scroll bar
'>' Redirects don't seem to work two well in perl5db.
Ho hum...
On Fri 30 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
677: CORE::open ($fh, "+<&=$fd") or croak (__PACKAGE__ . "::_open_multi_m ap() - unable to open file descriptor for locking: $!");
What does it print when you run it outside the debugger?
Kake
----- Original Message ----- From: "Kate L Pugh" kake@earth.li To: "Discussion of development on the OpenGuides software" openguides-dev@openguides.org Sent: 31 January 2004 01:21 Subject: Re: [OpenGuides-Dev] OpenGuides on Windows?
On Fri 30 Jan 2004, IvorW ivorw-openguides@xemaps.com wrote:
677: CORE::open ($fh, "+<&=$fd") or croak (__PACKAGE__ . "::_open_multi_m ap() - unable to open file descriptor for locking: $!");
What does it print when you run it outside the debugger?
Took some persuasion and a cygwin shell:
Search::InvertedIndex::DB::DB_File_SplitHash::_open_multi_map() - unable to open file descriptor for locking: Invalid argument at blib/lib/Search/InvertedIndex.pm line 282 (in cleanup) Class::NamedParms::get() - Attempted to retrieve an undeclared or unitialized named parameter: '-hash' Class::NamedParms::get('Search::InvertedIndex::DB::DB_File_SplitHash=HASH(0x177529c)', '-hash') called at blib/lib/Search/InvertedIndex/DB/DB_File_SplitHash.pm line 373 Search::InvertedIndex::DB::DB_File_SplitHash::close('Search::InvertedIndex::DB::DB_File_SplitHash=HASH(0x177529c)') called at blib/lib/Search/InvertedIndex.pm line 1346 Search::InvertedIndex::close('Search::InvertedIndex=HASH(0x24d56d0)') called at blib/lib/Search/InvertedIndex.pm line 1367 Search::InvertedIndex::DESTROY('Search::InvertedIndex=HASH(0x24d56d0)') called at t/010_update.t line 0 eval {...} called at t/010_update.t line 0 (in cleanup) Class::NamedParms::get() - Attempted to retrieve an undeclared or unitialized named parameter: '-hash' Class::NamedParms::get('Search::InvertedIndex::DB::DB_File_SplitHash=HASH(0x177529c)', '-hash') called at blib/lib/Search/InvertedIndex/DB/DB_File_SplitHash.pm line 373 Search::InvertedIndex::DB::DB_File_SplitHash::close('Search::InvertedIndex::DB::DB_File_SplitHash=HASH(0x177529c)') called at blib/lib/Search/InvertedIndex/DB/DB_File_SplitHash.pm line 395 Search::InvertedIndex::DB::DB_File_SplitHash::DESTROY('Search::InvertedIndex::DB::DB_File_SplitHash=HASH(0x177529c)') called at t/010_update.t line 0 eval {...} called at t/010_update.t line 0 # Looks like your test died before it could output anything.
Hope this helps,
Ivor.
openguides-dev@lists.openguides.org
-
Earle Martin -
IvorW -
Kate L Pugh