Closed Bug 368212 Opened 18 years ago Closed 18 years ago

sl: search plugin update

Categories

(Mozilla Localizations :: sl / Slovene, defect)

x86
Linux
defect
Not set
normal

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: aleksr_si, Assigned: aleksr_si)

Details

(Keywords: verified1.8.1.2)

Attachments

(1 file, 1 obsolete file)

The Najdi.si website has changed its encoding from ISO-8859-2 to UTF-8; could we change the search plug-in to reflect in the next series of updates to Firefox 2? Or do we have to wait until the next 'big' version?

It would be really great if we could do this now as it seems the Najdi.si owners are willing to advertise Firefox for free for us, but it would probably be more difficult to get them to give the final 'yes' if their search plug-in wasn't working properly.

Also the owners would like to add a referrer tag to the plug-in so that they can track how many people are using this search plug-in. Can we do that?
Attached patch Patch (obsolete) — Splinter Review
This would be the patch. I hope the 'mode="results"' parameter is right - that's what they gave me from Najdi.si, but their search plug-in is a Sherlock one and I'm not sure if this is the right way to go about in this format as well.
Attachment #252800 - Flags: approval-l10n?
CC'ing Mic, she should make the call about the parameter. I'd like to see the actual intended result in the query string, as there is no such thing as a mode in MozSearch. Details on that are on http://developer.mozilla.org/en/docs/Creating_MozSearch_plugins.

Changing the encoding seems like a good thing to do, could you request approval1.8.1.2 instead of approval-l10n? The latter isn't really used anymore.
Comment on attachment 252800 [details] [diff] [review]
Patch

Thanks Axel.

I'm not sure what the mode parameter is meant to do. I tried using the Sherlock plug-in which I was given, and it doesn't seem to come up in the URL at all as a GET attribute. So I imagine it's either some directive for Firefox/Mozilla, or else it may be that the author messed something up.

Shall I send Najdi.si an email and ask what they mean by the parameter?

(I suppose the 'approval1.8.1.2' flag should probably wait until this is cleared up ...)
Attachment #252800 - Flags: approval-l10n? → approval1.8.1.2?
(In reply to comment #3)
> I'm not sure what the mode parameter is meant to do. I tried using the Sherlock
> plug-in which I was given, and it doesn't seem to come up in the URL at all as
> a GET attribute. So I imagine it's either some directive for Firefox/Mozilla,
> or else it may be that the author messed something up.

As far as I can tell, the 1.5-and-lower search service paid no attention to the mode attribute on <input>s other than to ignore those that contain the exact string "mode=browser" (http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/xpfe/components/search/src/nsInternetSearchService.cpp&rev=1.252#4630).

The new search service certainly doesn't pay any attention to it, so you should just remove it.
Attached patch Patch #2Splinter Review
OK, excellent. Here's the new patch with the 'mode' parameter removed. :)
Attachment #252800 - Attachment is obsolete: true
Attachment #252925 - Flags: approval1.8.1.2?
Attachment #252800 - Flags: approval1.8.1.2?
hey
from what i understand of the comments above, the only change now to the nadji plug in reflects their change in encoding from ISO-8859-2 to UTF-8 and no longer has a parameter added? 
i apologize but i have a limited understanding if this change impacts the user experience, my sense from the research I've done on wikipedia is there is no change, it's a back end code compliance type of move. if that's correct (no impact to user experience and no parameter required) and the team is ok on adding this as a patch then BD/me would give approval for the change
thanks
mic
Hi Mic,
there is still a change in search parameters, in particular

 sourceid=mozilla-search

becomes

 foxsbar=ff

in the query string. The changes discussed before are merely about the technical implementation of that change.
Attachment #252925 - Flags: review?(l10n)
Comment on attachment 252925 [details] [diff] [review]
Patch #2

This is on Mic's desk, as can be seen in the comments.
Attachment #252925 - Flags: review?(l10n) → review?(mic)
aleks,
can you explain to me the rationale behind the change in search parameters? i just need to understand the purpose and if there is any functional impact to the user (my guess on the latter question is none).
thanks
mic
There shouldn't be any functional impact to the user.

The change in this parameter (sourceid vs foxsbar) was requested by the Najdi.si site owners because they'd like to monitor how many people use the Firefox search plugin (they use the same parameter in their own downloadable plugin for Firefox - which is available for anyone using the en-US version or Seamonkey, for example). 

I don't know why they can't simply use 'sourceid' - but it is true that 'sourceid' seems to be the default value (it's used in Amazondotcom and Creativecommons search plugins in en-US, for example, and it used to be used in the Google plugin as well, which is I think where it comes from). So they've probably developed their own system to follow this, and sourceid is simply obsolete in this plugin.

So to sum it up - it is only the background process that is affected with changing this parameter. There is no change for the user whatsoever.

Changing the encoding to UTF-8 will, on the other hand, improve user experience, because after the change on the server occurred, the letters č, š and ž in the Slovene alphabet are no longer searchable using this search plugin (as they get incorrectly encoded).
Aleks, got you. thanks very much for the explanation, you have BD approval to go ahead. 
Team, I will document this change on our internal wiki page and reference this bug as well.
best
mic
Comment on attachment 252925 [details] [diff] [review]
Patch #2

[sl] approval to update search plugin to match new site encoding, and update tracking code as reviewed by mic for Firefox 2.0.0.2.

Please use the fixed1.8.1.2 and verified1.8.1.2 keywords as appropriate.
Attachment #252925 - Flags: review?(mic)
Attachment #252925 - Flags: review+
Attachment #252925 - Flags: approval1.8.1.2?
Attachment #252925 - Flags: approval1.8.1.2+
Thanks everyone. I've now committed the changed file to CVS, so I'm marking the bug fixed.
Status: NEW → RESOLVED
Closed: 18 years ago
Keywords: fixed1.8.1.2
Resolution: --- → FIXED
Aleks, could you ask Najdi to pull a nightly from ftp and to do an end-to-end test to verify that the search key produces the results they expect? Make sure that the boxens cycled and the have the change, of course.
Yes, of course, I'll do that. Thanks.
The changed plugin works as expected now. I've got a reply from Najdi.si as well, and they're very satisfied with the change (they've tested the plugin as well and say it works as intended).

Hence I'm marking the bug as verified. (Do I need to remove 'fixed1.8.0.2' from the keywords list now?)

Again, thanks everyone for your input. :)
Status: RESOLVED → VERIFIED
Keywords: verified1.8.1.2
Yep, it should be either fixed or verified. Wasn't exactly sure myself, though, either ;-).
Keywords: fixed1.8.1.2
You need to log in before you can comment on or make changes to this bug.