Re: [webkit-changes] cvs commit: JavaScriptCore/kjs date_object.cpp
I'm guessing that this fixes the case of adding a value to a date and crossing the DST boundary. At least prior to this patch it would be off by an hour. I posted a possible fix which seems to fix it for me and has been applied to KJS. It's portable, and this is not. Any reason why the portable one wasn't used, and is it possible to commit it also? (assuming that this patch does what I think..) On Monday 03 October 2005 18:44, Timothy wrote:
* kjs/date_object.cpp: (timetUsingCF): Fix one of the date tests my making the CF version of mktime have the same quirk about the DST field as the real mktime. * tests/mozilla/expected.html: Updated for newly fixed test.
- + CFAbsoluteTime absoluteTime = CFGregorianDateGetAbsoluteTime(date, timeZone); + + if (tm->tm_isdst >= 0) { + if (CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && !tm->tm_isdst) + absoluteTime += 3600; + else if (!CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && tm->tm_isdst) + absoluteTime -= 3600; + } + CFTimeInterval interval = absoluteTime +
-- George Staikos KDE Developer http://www.kde.org/ Staikos Computing Services Inc. http://www.staikos.net/
On Oct 3, 2005, at 9:14 PM, George Staikos wrote:
I'm guessing that this fixes the case of adding a value to a date and crossing the DST boundary. At least prior to this patch it would be off by an hour. I posted a possible fix which seems to fix it for me and has been applied to KJS. It's portable, and this is not. Any reason why the portable one wasn't used, and is it possible to commit it also? (assuming that this patch does what I think..)
This is a merge to a branch of a fix from way back (well, 8/11, almost two months ago). It has been on TOT a long time. The mktime man page implies that if the tm_isdst field is set and nonnegative, it overrides the actual DST for that time. This patch attempts to emulate this quirk for the CoreFoundation time function case. The higher-level code that calls mktime counts on this quirk. Now, maybe that code was always wrong - I dunno. I'm not sure which other fix of yours you have in mind. Regards, Maciej
On Monday 03 October 2005 18:44, Timothy wrote:
* kjs/date_object.cpp: (timetUsingCF): Fix one of the date tests my making the CF version of mktime have the same quirk about the DST field as the real mktime. * tests/mozilla/expected.html: Updated for newly fixed test.
- + CFAbsoluteTime absoluteTime = CFGregorianDateGetAbsoluteTime (date, timeZone); + + if (tm->tm_isdst >= 0) { + if (CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && !tm->tm_isdst) + absoluteTime += 3600; + else if (!CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && tm->tm_isdst) + absoluteTime -= 3600; + } + CFTimeInterval interval = absoluteTime +
-- George Staikos KDE Developer http://www.kde.org/ Staikos Computing Services Inc. http://www.staikos.net/ _______________________________________________ webkit-dev mailing list webkit-dev@opendarwin.org http://www.opendarwin.org/mailman/listinfo/webkit-dev
Sounds like fine comments to have in the code, not just the changelog. Dave On Oct 3, 2005, at 10:24 PM, Maciej Stachowiak wrote:
On Oct 3, 2005, at 9:14 PM, George Staikos wrote:
I'm guessing that this fixes the case of adding a value to a date and crossing the DST boundary. At least prior to this patch it would be off by an hour. I posted a possible fix which seems to fix it for me and has been applied to KJS. It's portable, and this is not. Any reason why the portable one wasn't used, and is it possible to commit it also? (assuming that this patch does what I think..)
This is a merge to a branch of a fix from way back (well, 8/11, almost two months ago). It has been on TOT a long time. The mktime man page implies that if the tm_isdst field is set and nonnegative, it overrides the actual DST for that time. This patch attempts to emulate this quirk for the CoreFoundation time function case. The higher-level code that calls mktime counts on this quirk.
Now, maybe that code was always wrong - I dunno. I'm not sure which other fix of yours you have in mind.
Regards, Maciej
On Monday 03 October 2005 18:44, Timothy wrote:
* kjs/date_object.cpp: (timetUsingCF): Fix one of the date tests my making the CF version of mktime have the same quirk about the DST field as the real mktime. * tests/mozilla/expected.html: Updated for newly fixed test.
- + CFAbsoluteTime absoluteTime = CFGregorianDateGetAbsoluteTime(date, timeZone); + + if (tm->tm_isdst >= 0) { + if (CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && !tm->tm_isdst) + absoluteTime += 3600; + else if (!CFTimeZoneIsDaylightSavingTime(timeZone, absoluteTime) && tm->tm_isdst) + absoluteTime -= 3600; + } + CFTimeInterval interval = absoluteTime +
-- George Staikos KDE Developer http://www.kde.org/ Staikos Computing Services Inc. http://www.staikos.net/ _______________________________________________ webkit-dev mailing list webkit-dev@opendarwin.org http://www.opendarwin.org/mailman/listinfo/webkit-dev
_______________________________________________ webkit-dev mailing list webkit-dev@opendarwin.org http://www.opendarwin.org/mailman/listinfo/webkit-dev
On Tuesday 04 October 2005 13:16, David Harrison wrote:
Sounds like fine comments to have in the code, not just the changelog.
I understand what happened here - it was just a coincidence that this patch went in at the same time. Apparently my patch is still pending and mostly unrelated. -- George Staikos KDE Developer http://www.kde.org/ Staikos Computing Services Inc. http://www.staikos.net/
participants (3)
-
David Harrison
-
George Staikos
-
Maciej Stachowiak