Closed Bug 485251 Opened 15 years ago Closed 15 years ago

"On <date> <sender> wrote:"-string is not a correct sentence in swedish l10n

Categories

(Mozilla Localizations :: sv-SE / Swedish, defect)

x86
macOS
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: hwaara, Assigned: u60234)

Details

Using Thunderbird, the latest beta, for Linux.

The reply-string that is prepended the message is not grammatically OK in swedish. It seems to be word-for-word translated from the english version.

Right now, it's (typing from memory; I saw it at work and am home now):

"Den 25 mars 2009, Karolina skrev:"

That's not grammatically right and just weird.

But it should be for example one of these that are more OK, but choosing one is up to the localizer I guess:

1) "Karolina skrev den 25 mars 2009:"
2) "Den 25 mars 2009 skrev Karolina:"
Or even:

3) "Karolina, den 25 mars 2009:"
So is this a bug in the swedish localisation (if so it should be moved to the Mozilla Localizations product) or that Thunderbird's code base doesn't allow this to be localised correctly?
Midaired, but I'll submit anyway :)

The string comes from two .properties, mailnews.reply_header_authorwrote and mailnews.reply_header_ondate, with the name and date substituted in at the spot of the %s, and which are put together according to mailnews.reply_header_type as nothing (0), authorwrote (1, the SM default), ondate+authorwrote (2, the Tb default), or authorwrote+ondate (3).

Are you saying that it's not possible to create strings which would be correct from a combination of those (looks to me like 1 is "switch reply_header_type to 3 for sv-SE", 2 is "move the %s in reply_header_authorwrote" unless not having a comma is grammatically required, and 3 is the probably-bad-idea "change reply_header_authorwrote to just '%s'"), and that we need to change something in nsMsgCompose.cpp to allow for further per-locale customization? If so, what is that something? Move the comma into a localized string (probably two, one for whether ondate+authorwrote needs punctuation and another for authorwrote+ondate)?

Or did you mean to file this against the sv-SE localization, and you just forgot where the Mozilla Localizations product is? ;)
One part of this is surely a bug for the l10n component. Since I haven't looked at the code, I don't know if part of the bug is also to make this string more localizable.
Assignee: nobody → hasse
Component: Message Compose Window → sv / Swedish
Product: Thunderbird → Mozilla Localizations
QA Contact: message-compose → slicedlime
We can fix this string to (1) in l10n, with some reservation for the date. I don't see the long date format "29 mars 2009" when testing, only the short format "2009-03-29". So this is what we get:

"Karolina skrev 2009-03-29 08:25:"

I do wonder why only two of the four prefs we have to change to localize this string are localized prefs. The other two, "mailnews.reply_header_separator" and "mailnews.reply_header_type" are not. We can fix this bug by adding them to all-l10n.js, but it does seem overly complicated.

http://hg.mozilla.org/l10n-central/sv-SE/rev/dce384593c7a
http://hg.mozilla.org/releases/l10n-mozilla-1.9.1/sv-SE/rev/a81a23ff23f7
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → FIXED
Because they came from the source of many nightmares, bug 70842 and bug 120870.

The intention there was that nothing would ever ship with anything but reply_header_type as 1, and thus l10n of what shipped could be completely done by changing reply_header_authorwrote. The other things were only intended as backend for the hypothetical UI that bug 107884 would hypothetically add (except reply_header_colon - apparently they didn't think any locale would want to change it), and any l10n problems with assembling the parts would then be the problem of whoever had to implement that UI.

Then Tb switched to reply_header_type 2 in bug 438375, where I failed to grasp that even though we appear to have supported three different localized forms for years, practically we've supported just one, the only one most people ever saw.

So, other than making reply_header_separator a localized pref, what can I do to make it easier for you to survive the fact that even if you change reply_header_type, people can change it to something else, and then blame you when the bullet hits their foot?

Now you're at:

1. Karolina skrev:
2. 2009-03-29 08:25 Karolina skrev:
3. Karolina skrev 2009-03-29 08:25:

Are those what you would choose, if you could? If not, what do you need to be able to get all three right? Do 2 and 3 need to be separate localized strings rather than assembling the same parts in different orders, so you can say (for example) "Den %d %s skrev" but "%s skrev %d"?
I would like 2 to be:

2009-03-29 08:25 skrev Karolina

So we would really need to change reply_header_authorwrote if reply_header_type is 2. As long as the date is never in the long format, "29 mars 2009", there is no need to start with a "Den" article.

But I imagine there are a large number of ways this can fail in most languages if users change reply_header_type. So we will either have to live with that, or have separate strings for the 3 possible header types.
Hasse:

Slightly offtopic, but isn't it more common for time to be written e.g. "08.25" in swedish? It would also look better when the colon is appended at the end of the reply string ("08:25:").
Yes, a dot might look better here. But I am afraid the date and time formats are taken from the system preferences, so I don't think there is much we can do about that.
You need to log in before you can comment on or make changes to this bug.