Closed Bug 384457 Opened 18 years ago Closed 17 years ago

Firefox productization for Serbian (sr)

Categories

(Mozilla Localizations :: sr / Serbian, defect)

Product:
Mozilla Localizations
Component:
sr / Serbian
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: Pike, Assigned: filmil)

References

Details

Attachments

(2 files, 3 obsolete files)

Inverted previous patch (the directories were given in the inverse order, this patch corrects this problem)
5.30 KB, patch
Gavin
: review+
Details | Diff | Splinter Review
Removal, against trunk
565 bytes, patch
Pike
: review+
Details | Diff | Splinter Review
We need to find a good set of search engines for Serbia, as well as a news feed and rss readers. Filmil, any suggestions on those?
No longer blocks: land-sr
I considered a few. Can you point me to criteria (thus in writing, if possible) the search engines must fulfill?
Hi Filmil we don't currently have anything written down to describe criteria for search engine's (so thanks, I will add that to documentation I'm working on as something we should do!). In the mean time we usually provide up to 6 search engine's for the local version. The first few should cover search generally e.g., a local version of Google or Yahoo. What I mean by local is a search engine that searches Serbian pages particularly well and is liked by Serbian speaking users. Then we provide search engines that search books/shopping/auction, it could be Amazon, eBay or a Serbian version. Wikipedia has a Serbian version and we would normally include that as well. Here is a link to a page with the typical search engine's we use (which hopefully isn't too much information). http://wiki.mozilla.org/Firefox3/L10n_Requirements does that help?
Yes, this is fine so far.
Status: NEW → ASSIGNED
Blocks: land-sr, fx20-sr
Here are some engines specific for Serbian. The dictionaries and search engines I can bundle within the browser/searchplugins directory. What should I do with the news aggregators and RSS feeds? Thanks, f * Serbian search engines http://www.krstarica.co.yu http://www.pogodak.co.yu * Serbian dictionaries http://www.vokabular.org http://www.metak.com * Serbian Wikipedia http://sr.wikipedia.org * Serbian news www.naslovi.net
Assignee: nobody → filmil
Status: ASSIGNED → NEW
I added search and dictionary search engines to the CVS. Where do I put the news feed and rss readers, and what of them should I put? Their home pages, or their generated RSS feeds, or something else? And where. In short, what to do with them? f
Status: NEW → ASSIGNED
Hi the news source, if that's the best Serbian news source and for your users a better experience than the BBC than you use their generated RSS feeds, however, we need to get permission from them first. (In the past we have crashed servers for providers who weren't expecting the additional traffic we generate). I can help you do this if you wish, unless you already have a contact there? on search engines and dictionaries - did you add all of them? usually one dictionary is sufficient versus two. I can't really tell if one is better than the other, can you? and can you give me the current list (in order) of the search engine's you've ended up with. thanks mic
The engines given in https://bugzilla.mozilla.org/show_bug.cgi?id=384457#c5 are specific for Serbian language. This is not to say that Serbian users do not use Google, or BBC or any other worldwide website. Serbian sites have some advantages here and there. E.g. pogodak searches both Cyrillic and Latin scripts. Also, the dictionaries are not straightforward. metak is an serbian<->english dictionary. Vokabular is a dictionary of words with foreign origin. Thus metak and Vokabular are not interchangeable and you cannot leave one out on the account that it is 'better' than the other because their functions are orthogonal. Further, there are no 'amazon' or 'eBay' for Serbian market, AFAIK. But does this mean I should remove them from the list? It does not seem so to me. All in all, so far I see no clear criterion by which to choose preferred search engines. Can you give me at least rough guidelines so that I can make a choice? The priority list can be found in the CVS, file list.txt in l10n/sr/browser/searchplugins directory.
For the technical side of the process, please attach patches for changes to search engines and RSS readers etc.. There are some legal issues to resolve on the Mozilla side enabling us to actually incorporate other companies trademarks and services. Request review from both me and Mic, Mic will request the required information and agreements, and I'll be doing a technical review. I'm pretty sure that the current set gets an r-, and there are searches that aren't even remotely explained in this bug. Like, Google-[es], searching on elitesecurity.org? The current list has 13 entries, that should be just confusing for new users. One engine per category should do it, following the "less is more" principle. The categories should more or less correspond to those in the en-US build, unless we find out in this bug that we want to change that. On to the actual list: Regarding dictionaries, sr.wikipedia.org has 50k entries. Sure, that's not top of the world, but it's not totally bad. I'd suggest to just not ship metak and Vokabular at all. Regarding ebay, if we don't have an alternative, let's just stick with ebay. For amazon, we might consider using amazon.co.uk instead of amazon.com, that might at least perform better in terms of shipping. I have no idea what elitesecurity.org or naslovi.net would be. Before we do more changes to the sources, let's finalize a list of plugins here in the bug, and then get a patch to implement that, that helps us to keep the products decisions apart from the technical implementation. Sorry for being stubborn here, but this is Mozilla exposing both trademarks and resources of other folks, and we don't want them to get grumpy on us. Been there, done that, better not go there again.
(In reply to comment #9) > Sorry for being stubborn here, but this is Mozilla exposing both trademarks and > resources of other folks, and we don't want them to get grumpy on us. Been > there, done that, better not go there again. No problem at all, I am grateful for the explanations. When I wanted to make a list of search plugins, I realized I had no criterion to rank the engines I added. Further, given that the international websites are far more advanced than Serbian ones, on the grounds of technical perfection, the international sites would win. However, it would be nice to have at least one local site around. So these two goals conflict, hence my questions.
Yeah, we're having this problem in many regions. I'm not sure if we have a way to make it easier for localized services to get better, or even if that's necessarily a goal. But for inclusion in Firefox, we should go by current user value, not by potential, IMHO. If another service catches up or takes over, let's change it then. We're probably (hopefully) having a market impact either way, not sure if there's a way to get half a chicken and half an egg without getting a mess. (Let's hope that my cutting the chicken-egg-problem in half makes sense to anybody.)
Point taken. I removed the custom search engines for Serbian, leaving the original ones in. This takes care of the issue with the search engines. So far I am content that at least Google will offer its localized variant. Consider the search plugins issue solved.
Mic, this needs your insight. http://bonsai-l10n.mozilla.org/cvsblame.cgi?file=l10n/sr/browser/searchplugins/list.txt&rev=1.5&cvsroot=/l10n is the longest list of plugins that were checked in at one point, which is likely the answer to your question in comment 7.
Whiteboard: needs-mic
Ping?
No longer blocks: fx20-sr
Blocks: fx20-sr
Hi my apologies, somehow this fell off my radar :( Please give me your comments based on the following: using the list provided in comment#13 by pike not ok to use: naslovi-net as a search plug in. it's a repeat of google as they use google to power their search so there's no additional value I can see. If you want to use them as the feed over the BBC (see comment #7) then that's ok so long as we have their permission. I have pinged them for permission now. Please confirm that you do suggest they would be a better news provider than BBC. not ok to use: google.es I just don't understand the value of this considering it's a Serbian build. If a user speaks both Serbian and Spanish they can always add the plug in themselves. Unless Spanish is spoken by a majority of the population I don't see any other reason to add it shopping - is there no local shopping engine for a better user experience than USA amazon or ebay? i've done some research here and haven't found any. if there are no other options than I would suggest as Axel does in comment # to use amazon.uk and for ebay choose one of their european versions from http://www.ebay.eu/ for google - why aren't we using: http://www.google.com/intl/sr/ this seems like the serbian version of Google to me. for yahoo - i can't find a serbian version but is there another version of Yahoo besides yahoo.co that you feel is more relevant for Serbian users? e.g., yahoo.co.uk? or another one? you should choose between one of the two local search engine's you listed, assuming they do a better job of recognizing serbian language in searches and pulling up sebrian sites than google i'd accept including one of them. the one to choose is based on a combination of coverage, relevance and popularity. coverage meaning the most number of results as in it didn't miss anything, relevance most relevant results and popularity - which site is most popular among Serbian users: http://www.krstarica.co.yu http://www.pogodak.co.yu for dictionaries - i would leave off both of these unless you felt from a user experience perspective that users need/want this is a default option. I don't think we do this anywhere else (Axel can you confirm if that's true) where we ship with a set of dictionaries in the default search plug in list. the order of search engine's are incorrect it should be: google yahoo then alphabetical order of the balance of the engine's included so I think the final list would be: http://www.google.com/intl/sr/ yahoo.com I'd suggest as the 3d or even 2nd option you include one of the local Serbian search engine's. We've done this in Slovenia and it seems to work well for users. Especially considering Yahoo doesn't appear to offer anything in Serbian or for Serbian users specifically. amazon.co.uk answers.com creativecommons ebay (from europe vs USA) wikipedia.sr
Hi Filip with regards to the RSS readers. we usually ship with the defaults my yahoo bloglines igoogle as bloglines doesn't ship in your language but Netvibes does, it would be great to add that I think. What do you think? thanks mic PS I hope the information in comment #15 provided more clarity on what we usually do and why. each time I explain it I hope it gets better :)
(In reply to comment #16) > Hi Filip > with regards to the RSS readers. we usually ship with the defaults > my yahoo > bloglines > igoogle > > as bloglines doesn't ship in your language but Netvibes does, it would be great > to add that I think. What do you think? I will have to check what it is about. I don't use these services so will need some time to investigate.
(In reply to comment #15) > not ok to use: naslovi-net as a search plug in. it's a repeat of google as Ok. > not ok to use: google.es I just don't understand the value of this considering > it's a Serbian build. If a user speaks both Serbian and Spanish they can It is not google.es, but rather a google-powered search of Elitesecurity.org, a very popular IT-oriented web site. > shopping - is there no local shopping engine for a better user experience than > USA amazon or ebay? i've done some research here and haven't found any. if None that I know of. Mostly due to the fact that online commerce is only just starting. Situation may have changed, I have to check. > for google - why aren't we using: http://www.google.com/intl/sr/ > this seems like the serbian version of Google to me. What I understood from the recommendations, we should use the generic google URL and let Google figure out which UI it will present to the user. > for yahoo - i can't find a serbian version but is there another version of > Yahoo besides yahoo.co that you feel is more relevant for Serbian users? e.g., > yahoo.co.uk? or another one? Yahoo has been notorious for poor support for the Serbian cyrillic script. For this reason it seems that there is no support from Yahoo. > pulling up sebrian sites than google i'd accept including one of them. the one > to choose is based on a combination of coverage, relevance and popularity. > coverage meaning the most number of results as in it didn't miss anything, > relevance most relevant results and popularity - which site is most popular I would go for http://www.pogodak.co.yu, as it searches simultaneously in cyrillic an latin scripts, and supports declensions. These are typically not found in search engines. Although Google has done something on that field, they still have a way to go, but I suppose no immediate interest in doing so. This is where Pogodak scores high in the local market. (nb: I am not affiliated with Pogodak in any way) > for dictionaries - i would leave off both of these unless you felt from a user > experience perspective that users need/want this is a default option. I don't > think we do this anywhere else (Axel can you confirm if that's true) where we > ship with a set of dictionaries in the default search plug in list. I'd rather have a nice web site from where the add-ons such as the dictionary could be easily downloaded and installed. The only reason I would have in favor of including the dictionaries is that, for some reason, Serbian users default to assuming that localizations are not available --- and would thus not go through the trouble of looking the Serbian dictionary up. Otherwise, there is no reason to bloat the install with dictionaries, and I'd rather solve the "not looking for" issue by educating the users. > > the order of search engine's are incorrect it should be: > google > yahoo > then alphabetical order of the balance of the engine's included so I think the > final list would be: > http://www.google.com/intl/sr/ > yahoo.com > I'd suggest as the 3d or even 2nd option you include one of the local Serbian > search engine's. We've done this in Slovenia and it seems to work well for > users. Especially considering Yahoo doesn't appear to offer anything in Serbian > or for Serbian users specifically. > amazon.co.uk > answers.com > creativecommons > ebay (from europe vs USA) > wikipedia.sr > I agree with this list, and would suggest Pogodak as the local search engine. Let me know what to do next, please.
perfect, and Pogodak is good. we should get permission from them as well. i will email them now. also, given what you say about Yahoo. I'd suggest that we use Pogodak in the 2nd position and list Yahoo 3d. On netvibes, a few other localizations have included Netvibes. you can ping the L10n group and get some additional information. Pike may have some ideas as well. For http://www.netvibes.com/ look at settings in the top right corner and you should find Serbian there. Confirming in this bug we received permission from Miloš Mileusnić of Naslovi to include their RSS feed. best, mic PS did the information I provided help explain what we look for in search engine's better?
> Confirming in this bug we received permission from Miloš Mileusnić of Naslovi > to include their RSS feed. Yeap, it is confirmed. ;) Naslovi.net is kind of "Google News" for Serbia, so you can also include Naslovi.net news search in search box if you like. About search engines, instead of http://www.google.com/intl/sr/ you can use www.google.co.yu which is now owned by Google. I believe that Google and Pogodak would be the best solution for web search.
Whiteboard: needs-mic → needs-Pike
Ping?
I guess I don't have comments on the setup here, other than I'd like to see a single wrap-up comment by Mic on what's actually supposed to be implemented. Then we should have a patch implementing that, with a review request on me again. I don't think we should have a serbian-specific google plugin, though. status-whiteboard-ping-pong :-)
Whiteboard: needs-Pike → needs-mic
hi guys on my radar. will check back through email for approvals and summarize in the bug in a few hours thanks for the ping :) mic
Hi Mic. I suspect that the few ours in question have expired. :) Can I have any news about what needs to be done further?
Hi we now have criteria for search which you can reference: http://wiki.mozilla.org/Firefox_web_services_guidelines ping me on IRC or IM to discuss can you summarize the current search engine's you want to include from above, my guess is: google popdak ?? which others wikipedia and the RSS feed is: Naslovi.net for RSS readers not sure if any are in Serbian, do you know? thanks
Well, I have also suggested Naslovi.net as a search option as well, since it is news aggregation site and news search engine similar to Google News. Unfortunately, news plays important parts of our lives here in Serbia, so it could be categorized as "specific interest search" according to Firefox web services guidelines: > Where user are using popular search engine's to search > more specific areas of interest that are very popular in your language or > locale. There should be no more than two. But that is just a suggestion, I don't want to be importunate since Naslovi.net is already counted as RSS feed source. :) Miloš, Naslovi.net
filmil i've got back over this bug a few times and cross referenced to the link i provided in comment #25 for web services, so here is what I have as final, please confirm this search plug in list and implement in a new patch Note: for the two dictionaries, please make a choice on including one or both or none Google Pogodak (I received permission from Bojan Rendulic on Oct 15) eBay.co.uk Amazon.co.uk http://www.vokabular.org http://www.metak.com Wikipedia-sr
for RSS feed please use Naslovi.net
for RSS readers from comment #17. did you find any information out about what is being used in Serbian? I will try and get some information as well.
one more thought on RSS readers, if we can not find any in Serbian would using another in Romanian, Croatian, Slovakian or Hungarian work? I believe NetVibes is translated into at least one of these languages and I understand from Wikipedia that these other languages are spoken in Serbia.
Mic, thanks for your input. My recommendation is to use english as fallback. Concerning the fallback languages, it is my opinion that none of those you suggested would work well. Romanian, Slovakian and Hungarian are spoken in limited bi/multilingual areas close to the borders with Hungary and Romania. The majority of the population does not understand any of these languages, so falling back to them does not make sense. Concerning Croatian language as a fallback, it is probably wise not to do it. Croatian language is rather similar to Serbian, and in most cases the speakers of the two languages can communicate with each other without misunderstanding. However, due to recent events, such a decision would in some cases be greeted with disdain.
(In reply to comment #29) > for RSS readers from comment #17. did you find any information out about what > is being used in Serbian? No RSS readers qualify. f
thanks for information per comment #31 so for RSS Readers, we will use en-US defaults can you please confirm the search plug in list from comment 27
(In reply to comment #27) > http://www.vokabular.org > http://www.metak.com I prefer Vokabular. Other listed sites are fine. f
great. so, can you create a new patch for search plug-ins with the final list: Google Pogodak eBay.co.uk Amazon.co.uk http://www.vokabular.org Wikipedia-sr from previous comments, we'll use default RSS Readers from en-US (so no action required there) and, finally, for Live Bookmark/RSS Feed we use naslovi.net once you have that done, you're fine on my end and we turn it over to Axel for final technical review and then we're ready to complete the build :)
1) Where do I get the descriptions of eBay.co.uk and Amazon.co.uk? From en-GB? Do I need to copy the XML files from there or should I just put the labels in list.txt 2) Where do I update the Live Bookmark/RSS Feed? Thanks, f
1) This only works for en-US plugins, so if you need the British ones, you should copy them from en-GB and update list.txt 2) In bookmarks.html you already have the http://sr.fxfeeds.mozilla.com/sr/firefox/headlines.xml URL as the live bookmark address; you need to file a bug in the Websites component to get that URL redirected to the real feed.
filip, i heard from a friend of the community in Serbia (Danilo Segan), he suggested these web based RSS Readers. Let me know what you think: web-based RSS readers, I noticed that netvibes is indeed being translated to Serbian, but in Latin script (and looking at the Serbian translators' forum, they want to do it in Cyrillic as well): http://translate.netvibes.com/forum/viewforum.php?id=35 A completely Serbian RSS reader is http://www.blogodak.com/.
(In reply to comment #38) > filip, i heard from a friend of the community in Serbia (Danilo Segan), he > suggested these web based RSS Readers. Let me know what you think: This is good news. I know Danilo online (from his i18n/l10n contributions and other places) and met him briefly on occasion. His experience is definitely valuable for Serbian l10n. > web-based RSS readers, I noticed that netvibes is indeed being translated to > Serbian, but in Latin script (and looking at the Serbian translators' > forum, they want to do it in Cyrillic as well): > http://translate.netvibes.com/forum/viewforum.php?id=35 > > A completely Serbian RSS reader is http://www.blogodak.com/. Unsure if Blogodak qualifies, hence comment #32. Blogodak offers different functionality compared to, say, Google Reader. Blogodak has a fixed selection of RSS feeds, and each user can select a subset to read online; whereas Reader allows an user to insert and use own set of RSS feeds . It is AFAIK not possible to insert own URL in Blogodak, though it is possible to register an URL, and have it included in the global list upon approval. In case this behavior qualifies as RSS reader, I've no special complaint against its inclusion.
I filed a bug per comment #37. Committed the search plugin. What now?
See bug 409963 (per comment #40)
Filip, changes to the search plugins should always be done with patches in bugzilla first. I'll review the status quo in the tree when I get back to a real network connection next year. The review is good from a technical point of view, and the patch in the bug gives us a good documentation of what was changed and why. In particular when the check-in comment references the bug.
Okay, I'll remedy the lack of the patch by producing a cvs diff and attaching it here. Once this is done, the effect *should* be the same as if I sent a patch in first, only that the patch is actually committed too. If you absolutely require the patch to come in first, I'll remove the contents of searchplugins after the patch has been produced here.
The process here is to reduce the risk of errors. I'll look at the actual state in cvs as soon as I get back to my office, and then we'll see. There's no need for further noise in the cvs history of the affected files before that.
Ping?
Sorry for the lag. We'll need a bunch of changes to the current state of the tree. list.txt, amazon-en-GB.xml, and eBay-en-GB.xml are fine. pogodakyu.xml isn't an opensearch plugin at all, and vokabular.xml has a bunch of stuff that isn't either. Please consult http://developer.mozilla.org/en/docs/Creating_MozSearch_plugins on how to write these plugins. For wikipedia-sr.xml, please take a look at http://mxr.mozilla.org/l10n/source/lv/browser/searchplugins/wikipedia-en.xml, up to the non-needed definition of the os namespace, that's the preferred format. Generally, you shouldn't need to use OpenSearch namespace, you should prefer <Param> elements over complex <Url>s, there's no update information, and you should have a <SearchForm> element.
Whiteboard: needs-mic → needs-patch
In general, I asked the people to provide search plugins for their stuff and that is what I got. I'll rework these.
Axel, what to do if no update information is available? I see that the English Wikipedia search (http://mxr.mozilla.org/l10n/source/lv/browser/searchplugins/wikipedia-en.xml?raw=1) doesn't use it.
Just don't provide any update information at all, because these plugins will be updated by the regular application updates, if required.
Should I use the moz: namespace for the SearchForm?
There shouldn't be any need for prefixed element names, just make them all be in the default namespace, http://www.mozilla.org/2006/browser/search/.
Ok. I updated the plugins, hopefully in accordance with the link you provided. Can you take a second look now in the CVS please?
Whiteboard: needs-patch → needs-pike
Ping?
Hi Filip, sorry for the lag. r- on these, based on the following comments: Pogodakyu: You copied over the search key foxsbar=ff from the sherlock plugin, http://www.pogodak.co.yu/privzeti/sherlocsearch/pogodakyu.src (view-source: that for better viewing pleasure). You used the Cyrrilic script name for the plugin, but search the latin host. Should we add a locale=sr__CYR to the plugin, or should we depend on users picking either one, and then keep their cookie to use that setting. I think that latin and cyrillic script are both used, so maybe having this as a user preference is actually good. Mic, can you check back with Pogodak on which keys to send? Technically, the xmlns:os="http://a9.com/-/spec/opensearch/1.1/" should be removed, here, and in the other plugins. Vokabular: The <Description> element seems to have an encoding issue. Please remove the parts that are commented out, and <IconUpdateUrl>http://www.vokabular.org/favicon.ico</IconUpdateUrl>, too. wikipedia-sr: Looks good, but if you use Википедија (Српски / Srpski) instead of <ShortName>Википедија (српски)</ShortName>, Firefox will know that it has the plugin already and not offer installing another one when visiting sr.wikipedia.org. Filip, when Mic get's back with the comments from Pogodak, could you create a patch and request review from me before checking in? We need to fix region.properties, too, I just found out. browser.search.defaultenginename should be just "Google", as browser.search.order.1 should be. Both of those refer to a string in google.xml in en-US. browser.search.order.2 should be removed. As for browser.contentHandlers.types, it might make sense to add netvibes, or replace bloglines with it. Mic?
Whiteboard: needs-pike → needs-mic
(In reply to comment #54) > Hi Filip, sorry for the lag. r- on these, based on the following comments: No problem. Thanks for finding the time. First and foremost, I will further refrain from committing anything into CVS here before making a patch here and running it by you. Thanks for bearing with me up to now. > Pogodakyu: > You copied over the search key foxsbar=ff from the sherlock plugin, > http://www.pogodak.co.yu/privzeti/sherlocsearch/pogodakyu.src (view-source: > that for better viewing pleasure). This is true. It has a very simple explanation. I requested a search plugin from Pogodak and got back a file more or less similar to what you point to. I simply converted this to OpenSearch as best as I could. Do you suggest omitting foxsbar parameter altogether? > You used the Cyrrilic script name for the plugin, but search the latin host. > Should we add a locale=sr__CYR to the plugin, or should we depend on users > picking either one, and then keep their cookie to use that setting. I think > that latin and cyrillic script are both used, so maybe having this as a user > preference is actually good. My opinion is that we should follow suit and use Cyrillic throughout. However, I can imagine that users know better. > Technically, the xmlns:os="http://a9.com/-/spec/opensearch/1.1/" should be > removed, here, and in the other plugins. Ok. > Vokabular: > The <Description> element seems to have an encoding issue. Please remove the > parts that are commented out, and > <IconUpdateUrl>http://www.vokabular.org/favicon.ico</IconUpdateUrl>, too. Ok. > > wikipedia-sr: > Looks good, but if you use Википедија (Српски / Srpski) instead > of <ShortName>Википедија (српски)</ShortName>, Firefox will > know that it has the plugin already and not offer installing another one when > visiting sr.wikipedia.org. Ok. > Filip, when Mic get's back with the comments from Pogodak, could you create a > patch and request review from me before checking in? Will do as you suggest, and as said at the beginning of this comment. > We need to fix region.properties, too, I just found out. > > browser.search.defaultenginename should be just "Google", as > browser.search.order.1 should be. Both of those refer to a string in google.xml > in en-US. browser.search.order.2 should be removed. ACK on these too. Please allow me some time to get to doing this.
Sounds good.
pinged bojan.rendulic@pogodak.co.yu for search keys, awaiting response. Also for our reference, from Oct 15 here is a copy of his email providing permission: ----- Original Message ----- From: "Bojan Rendulic" <bojan.rendulic@pogodak.co.yu> To: "Michal Berman" <mic@mozilla.com> Sent: Monday, October 15, 2007 2:17:30 AM (GMT-0500) America/New_York Subject: Re: Firefox in Serbian Hello Michal, Thank you for your interest. Pogodak firefox Add-on is in attachment. I could send to you Add-on for all our Search Engines in other countries: Pogodak.hr (Croatia), Pogodak.ba (Bosnia and Herzegovina) and Pogodok.com.mk (Macedonia). Hello We would like to include your search engine as part of our default list of search engine's for our Serbian version of Firefox. This means you will get more traffic and we'd like to make sure that you are able to handle and want the extra traffic. Please let us know if this is ok. Thanks mic Best wishes Bojan -- Bojan Rendulic direktor Pogodak d.o.o. www.pogodak.co.yu
(In reply to comment #54) > Filip, when Mic get's back with the comments from Pogodak, could you create a > patch and request review from me before checking in? Okay, I went through your suggestions and corrected the things that don't depend on Mic. Once his input is in, I will make the patch as agreed. f
Ok, I'm not waiting anymore. :) Here is the first patch.
here is the latest, does this help?: ----- Forwarded Message ----- From: "Bojan Rendulic" <bojan.rendulic@pogodak.co.yu> To: "Michal Berman" <mic@mozilla.com> Sent: Thursday, January 24, 2008 1:43:12 AM (GMT-0500) America/New_York Subject: Re: Firefox in Serbian Dear Mic, I check bug tracking thread. The quick answer is that we don't want set the "locale=sr__CYR" parameter and script should stay as it is. User of Pogodak can set language preferences via links "latinica, ћирилица, sve u latinic" in upper left corner. Decision is the save in user cookie. And then the chosen language will be also use when user will search via FF search engine. P.S 1 : I check FF search engine script on FF 3.0 Beta 1 and it works OK. P.S 2 : you cannot check it right now, because link is not visible for FF 3.0 users yet (bug). We will fix this in next couple of days. Bojan
Attachment #300436 - Attachment is obsolete: true
Ping?
Whiteboard: needs-mic → needs-gavin
Comment on attachment 300453 [details] [diff] [review] Updated patch according to feedback from Pogodak >diff -u ./pogodakyu.xml ../../../../pot/sr-RS/browser/searchplugins/pogodakyu.xml >-<SearchPlugin xmlns="http://www.mozilla.org/2006/browser/search/"> >+<SearchPlugin xmlns="http://www.mozilla.org/2006/browser/search/" >+ xmlns:os="http://a9.com/-/spec/opensearch/1.1/"> Why these changes? It shouldn't be necessary to add this namespace (this applies to all files). >diff -u ./vokabular.xml ../../../../pot/sr-RS/browser/searchplugins/vokabular.xml >- <Description>Речник страних речи и израза</Description> >+ <Description>Речник страних речи и израза –í–æ–∫–∞–±—É–ª–∞—Ä —Ä–µ—á–Ω–∏–∫</Description> Is this an intentional change? >+ <!-- Ignoring the UpdateUrl as long as the update url uses no parameters --> >+ <!-- <UpdateInterval>14</UpdateInterval> --> >+ <!-- <UpdateUrl>http://metak.sosko.org/Vokabular.src</UpdateUrl> --> >+ <IconUpdateUrl>http://www.vokabular.org/favicon.ico</IconUpdateUrl> We don't specify IconUpdateUrls for shipped engines, since we can't update them anyways. If the icon ever needs to change we can just ship a new search plugin file. Just remove all of these lines.
Attachment #300453 - Flags: review-
Whiteboard: needs-gavin
Oh. I see now. The diff should be *inverse* from what you see here (i.e. lines beginning with '+' should begin with '-' and vice-versa). I will update the patch to reflect this. f
Comment on attachment 302656 [details] [diff] [review] Inverted previous patch (the directories were given in the inverse order, this patch corrects this problem) Ah, this makes much more sense! :)
Attachment #302656 - Flags: review+
Great. What now?
Have you landed the patch on the l10n CVS trunk? I don't really know if there are any more changes left after that patch. I imagine we'll end up reviewing your localization again at some point and suggest new changes then, if needed. I'm not really familiar with the entire l10n approval process.
once we have all web services in (as they should be) and gavin's done his technical review of your build (i.e., tinderbox green) we should be ok to move to resolved/fixed and ready your build for launch in B4. would suggest you watch google L10n group for opt=in to B4, when there's that opportunity opt in and we'll do final review including QA test (if you have community available to review the build you've done now across all platforms you should aim for that). once all that clears, should be ok to ship your build
(In reply to comment #68) > Have you landed the patch on the l10n CVS trunk? I don't really know if there > are any more changes left after that patch. I imagine we'll end up reviewing > your localization again at some point and suggest new changes then, if needed. > I'm not really familiar with the entire l10n approval process. > I haven't landed the patch on the CVS, as I agreed with Axel to do just that -- not commit. We haven't agreed when it would be OK to commit though, so here goes --- May I now commit the patch to CVS? Thanks.
Yes, you can land changes to the l10n trunk now.
Done. Gavin I expect you will take it from here. Let me know please how it goes.
Ping?
Whiteboard: needs-gavin
I'm not sure what's needed from me here. If there are no more changes to make to the Serbian "productization" files, then we can just resolve the bug as fixed. Mic?
Whiteboard: needs-gavin
i'll go back through and make sure that all outstanding issues have been dealt with i.e., permissions, etc and then recommend to close (esp if technical is good to go)
Whiteboard: needs-mic
About the naslovi.net RSS feeds, I would like to inform you that we now provide feeds on Cyrillic script by using "cyrillic" suffix to existing feeds, so the URL for main RSS feed on Cyrillic is http://www.naslovi.net/rss/cyrillic You can use that one if you prefer Cyrillic for Firefox. It is also supported for main categories feeds like http://www.naslovi.net/rss/sport/cyrillic http://www.naslovi.net/rss/ekonomija/cyrillic etc. Cheers, Milos
Blocks: fx3-wsl10n-sr
i think we still need pogodak keys - axel or filip can you confirm. otherwise we're pretty close to resolved, per comment #75
Whiteboard: needs-mic
(In reply to comment #77) > i think we still need pogodak keys - axel or filip can you confirm. AFAIK, keys per comment #60 are good in the CVS as it is now. Confirm with Axel, /me says OK. f
The keys are coming from pogodak, see comment 55. We probably should point them to B5 and invite them to do some end-to-end testing on their side, then we can still see if they want a special key. I have one last change request, Filip, could you provide a patch to remove the browser.search.order.2=Yahoo line from region.properties? We're not shipping yahoo, no need to add it to the ordered plugins.
Here's the patch as requested. Removing the key, hopefully that's enough.
Comment on attachment 310542 [details] [diff] [review] Remove search order 2 from browser's region.properties In principle, that's the right patch, but the wrong tree. You gave me a patch for the 1.8 branch, I need one for the trunk. We'll figure out what to do with that for the branch once that's in a sane state.
Attachment #310542 - Flags: review-
Oops. Seems I have to do things twice these days. Still unclear to me why the current branch on that directory was not HEAD, though.
Attachment #310542 - Attachment is obsolete: true
Comment on attachment 310547 [details] [diff] [review] Removal, against trunk r=me on this one. Please check it in with a review comment referencing this bug, and my review. Something like "bug 384457, removing yahoo from search order, too, r=l10n@mozilla.com" would work. On a general note, when you attach patches for search engines or region.properties, please request review from me, by setting the review flag on the patch to '?' and entering my bugzilla id as requestee.
Attachment #310547 - Flags: review+
Committed.
OK to close now? Thanks.
Status: ASSIGNED → RESOLVED
Closed: 17 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: