[Demexp-dev] [FIX] A classification bug that crashes the client

2007-03-26 Thread David MENTRE

Hello,

A recent issue in the classification (two tags having the same
classification level, "faa:") has triggered a bug that crashes the
client. The following patch makes the client more tolerant to such
issues. It will be integrated in next release.

In the meantime, Félix or Fred, could you change the classification so
that tag with id 89 (I think it is "faa: La Poste") has a different
classification level that another tag at the same level ("faa:", I
think it is something related to "Impôts" or "fiscalité")?

Best wishes,
netless d.bugging
changeset:   489:368dc1bbf5f55d4c6d2168fba37f0f1dc20747c9
tag: tip
user:David MENTRE <[EMAIL PROTECTED]>
date:Sun Mar 25 16:07:14 2007 +0200
files:   gtk2-clnt/tags.ml.nw srv/voting.ml.nw
description:
[bug fix][gtk2 client] Don't break in case of duplicated label in tag's hierarchy

In case the same level appears twice (e.g. tags "fa: A" and "fa: B"),
hierarchy building was raising an exception. We now handle correctly
this case, putting a warning on stderr.


diff -r 1560469c9ee79c48c88c39fcf57eee43582aa07f -r 368dc1bbf5f55d4c6d2168fba37f0f1dc20747c9 gtk2-clnt/tags.ml.nw
--- a/gtk2-clnt/tags.ml.nw	Fri Oct 27 22:13:11 2006 +0200
+++ b/gtk2-clnt/tags.ml.nw	Sun Mar 25 16:07:14 2007 +0200
@@ -94,7 +94,12 @@ let rec insert_into_hierarchy hierarchy_
 insert_into_hierarchy node_children (lvl_tail, id, label) in
   [Tag_node(node_lvl, new_node_children, node_id, node_label)]
 @ other_level
-  | _ -> assert(false)
+  | Tag_node(node_lvl, node_children, node_id, node_label) :: tl ->
+  Printf.eprintf "WARNING: duplicate hierachy for tag:%d\n" id;
+  let new_node_children = 
+insert_into_hierarchy node_children (lvl_tail, id, label) in
+  [Tag_node(node_lvl, new_node_children, node_id, node_label)]
+  @ tl @ other_level
 @ 
 
 Function [[tree_of_tag_hash]] converts a hash-table of tags indexed by
diff -r 1560469c9ee79c48c88c39fcf57eee43582aa07f -r 368dc1bbf5f55d4c6d2168fba37f0f1dc20747c9 srv/voting.ml.nw
--- a/srv/voting.ml.nw	Fri Oct 27 22:13:11 2006 +0200
+++ b/srv/voting.ml.nw	Sun Mar 25 16:07:14 2007 +0200
@@ -115,7 +115,7 @@ The simplest case to consider is when a 
 The simplest case to consider is when a response wins over all responses
 in a one-to-one match. The [[unambiguous_winner]] code returns, for a
 given pairwise matrix [[mat]], the number of this winner if it exists,
-with a [[majority]], or raise [[Not_found]] otherwise.
+with a [[majority]], or raises [[Not_found]] otherwise.
 
 \assumption{The matrix [[mat]] given as argument is square. 
   [[majority]]$\geq 1$}


___
Demexp-dev mailing list
Demexp-dev@nongnu.org
http://lists.nongnu.org/mailman/listinfo/demexp-dev


Re: [Demexp-dev] [FIX] A classification bug that crashes the client

2007-03-26 Thread Frederic Lehobey
Hi,

On Mon, Mar 26, 2007 at 10:00:06AM +0200, David MENTRE wrote:

> In the meantime, Félix or Fred, could you change the classification so
> that tag with id 89 (I think it is "faa: La Poste") has a different
> classification level that another tag at the same level ("faa:", I
> think it is something related to "Impôts" or "fiscalité")?

Done. 'fa: La Poste' changed for 'fc: La Poste'.

> netless d.bugging

I am impressed.  :-)

Experimentally,
Frédéric


___
Demexp-dev mailing list
Demexp-dev@nongnu.org
http://lists.nongnu.org/mailman/listinfo/demexp-dev


Re: [Demexp-dev] [FIX] A classification bug that crashes the client

2007-03-26 Thread David MENTRE

Hi Fred,

2007/3/26, Frederic Lehobey <[EMAIL PROTECTED]>:

Done. 'fa: La Poste' changed for 'fc: La Poste'.


I'm not sure it fixed the issue (and that I found the root cause of
the issue). I'm still unable to launch a new client. :-(

Can somebody else test with a new client?

Best wishes,
d.


___
Demexp-dev mailing list
Demexp-dev@nongnu.org
http://lists.nongnu.org/mailman/listinfo/demexp-dev


Re: [Demexp-dev] [FIX] A classification bug that crashes the client

2007-03-26 Thread David MENTRE

Hi,

2007/3/26, David MENTRE <[EMAIL PROTECTED]>:

I'm not sure it fixed the issue (and that I found the root cause of
the issue). I'm still unable to launch a new client. :-(


Ok, in fact the cache is not properly updated (why??) and the client
still triggers the bug. To avoid the issue, delete file
"$HOME/.demexp/cache-official.demexp.org-5".

Now if only I could understand why the cache is not properly updated... :-(

Best wishes,
d.


___
Demexp-dev mailing list
Demexp-dev@nongnu.org
http://lists.nongnu.org/mailman/listinfo/demexp-dev