[Announce] Erminig-NG: two-way synchronization Google Calendar <-> Fremantle Calendar - Page 13 - maemo.org

Active Topics

The N900 is now 15 years old! (9)
to Nokia N900 by Arno_11 - 4 hrs, 4 mins ago
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 16 hrs ago
more...

Page 13 of 63

Thread Tools

lorelei	2010-02-04 , 17:38
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#121

Originally Posted by Quanzel

What do you think about autosync and running in the background? Will it be possible to set a syncinterval?

Traditionally (= in Erminig for Diablo) there was the possibility of running Erminig in text-mode and some users defined periodical runs with cron. This is not really user-friendly, so yes the possibility to auto-sync/proper backgrounding is a good idea.

This is something that I can implement for sure: now the question is when

Let's say that I will keep this idea in mind for the future developments, once I have stable release rolling.

Quote & Reply |

The Following 2 Users Say Thank You to lorelei For This Useful Post:
Quanzel, Soder

Palleman	2010-02-04 , 22:47
Posts: 123 \| Thanked: 33 times \| Joined on Jan 2010 @ Stockholm	#122

Originally Posted by lorelei

Code:

                print "updated_min=%s" % (query.updated_min)
                print "updated_max=%s" % (query.updated_max)

Min 2010-02-02T11:37TZ and max 2010-02-04T22:36:36.64TZ

The second one includes seconds and 1/100 of seconds. Never seen that before, but maybe it's normal.

The error messages are the same, but the first two errors have a line number of one less (525 and 504). Guess you removed a line between the versions.

Originally Posted by lorelei

PS: just to make sure: did you remove the symlink to iso8601.py you created?

Yup.

Is it possible to run screen on N900? If so, and if you want to debug this, I could let you in through ssh and watch over you with screen -x.

Bonus URL.

/P

Quote & Reply |

The Following User Says Thank You to Palleman For This Useful Post:
lorelei

lorelei	2010-02-04 , 23:17
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#123

Originally Posted by Palleman

Min 2010-02-02T11:37TZ and max 2010-02-04T22:36:36.64TZ

The second one includes seconds and 1/100 of seconds. Never seen that before, but maybe it's normal.

Interesting. I have this:

updated_min=2010-02-03T22:47:08Z
updated_max=2010-02-04T22:56:54.78Z

notice how I don't have a trailing TZ but only a Z. I now understand why Google is barfing on the dates it receives from you (TZ is not valid).

Now the question is why does Erminig generates stuff with TZ.

Please put the file attached to this post in the same directory as erminig-ng, and run it like this (I've just extracted a fragment of the code that manages the dates):

Code:

python2.5 testdate.txt 126511062

the number above is the UNIX timestamp you posted previously.
I get this output, and yours should be exactly the same:

Code:

['__builtins__', '__date_re', '__datetime_re', '__datetime_rx', '__doc__', '__extract_date', '__extract_time', '__extract_tzd', '__file__', '__find_julian', '__name__', '__time_re', '__tzd_re', '__tzd_rx', '__version__', 'ctime', 'parse', 'parse_timezone', 'time', 'tostring']
 *** 2010-02-02T11:37Z 
1265110620 -> 2010-02-02T11:37:00.000Z

Bonus URL.

thanks

Attached Files

testdate.txt (390 Bytes, 122 views)

Quote & Reply |

The Following User Says Thank You to lorelei For This Useful Post:
Palleman

Palleman	2010-02-05 , 00:32
Posts: 123 \| Thanked: 33 times \| Joined on Jan 2010 @ Stockholm	#124

Originally Posted by lorelei

notice how I don't have a trailing TZ but only a Z. I now understand why Google is barfing on the dates it receives from you (TZ is not valid).

Mittåt! My mistake. Disregard the T, I was mislead by the 7Z on the first line. There are only Z's.

Should I still run the time converter?

/P

Quote & Reply |

The Following User Says Thank You to Palleman For This Useful Post:
lorelei

lorelei	2010-02-05 , 00:37
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#125

Originally Posted by Palleman

Mittåt! My mistake. Disregard the T, I was mislead by the 7Z on the first line. There are only Z's.

Should I still run the time converter?

/P

yes please: there are some other information in the output that may be of interest.

thanks

Quote & Reply |

The Following User Says Thank You to lorelei For This Useful Post:
Palleman

Palleman	2010-02-05 , 01:08
Posts: 123 \| Thanked: 33 times \| Joined on Jan 2010 @ Stockholm	#126

Originally Posted by lorelei

Code:

python2.5 testdate.txt 126511062

Code:

['__builtins__', '__date_re', '__datetime_re', '__datetime_rx', '__doc__', '__extract_date', '__extract_time', '__extract_tzd', '__file__', '__find_julian', '__name__', '__time_re', '__tzd_re', '__tzd_rx', '__version__', 'ctime', 'parse', 'parse_timezone', 'time', 'tostring']
126511062 -> 1974-01-04T05:57:42.000Z

Originally Posted by lorelei

the number above is the UNIX timestamp you posted previously.

No, it's not.

Originally Posted by lorelei

Code:

['__builtins__', '__date_re', '__datetime_re', '__datetime_rx', '__doc__', '__extract_date', '__extract_time', '__extract_tzd', '__file__', '__find_julian', '__name__', '__time_re', '__tzd_re', '__tzd_rx', '__version__', 'ctime', 'parse', 'parse_timezone', 'time', 'tostring']
 *** 2010-02-02T11:37Z 
1265110620 -> 2010-02-02T11:37:00.000Z

But that one is!

I get the exact same output, except that the line beginning with three stars is not there:

Code:

['__builtins__', '__date_re', '__datetime_re', '__datetime_rx', '__doc__', '__extract_date', '__extract_time', '__extract_tzd', '__file__', '__find_julian', '__name__', '__time_re', '__tzd_re', '__tzd_rx', '__version__', 'ctime', 'parse', 'parse_timezone', 'time', 'tostring']
1265110620 -> 2010-02-02T11:37:00.000Z

Quote & Reply |

The Following User Says Thank You to Palleman For This Useful Post:
lorelei

lorelei	2010-02-05 , 09:02
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#127

Originally Posted by Palleman

No, it's not.

correct. I missed the trailing 0 in my copy-paste

But that one is!

I get the exact same output, except that the line beginning with three stars is not there:

Again my fault (note to self: I shouldn't do debugging when I'm tired).

In fact I've been able to replicate the problem this morning, and I also have the right solution in mind. Please expect an updated by the end of the day (i.e. as soon as I get back to home

)

The problem is that in some cases the iso8601 module doesn't return the string date in the a format that Google likes, and your timestamp in particular is one of those that make it spit out bad values. Originally there's already code to manage those cases (it's part of the snippet of code I made you try a few posts ago), but for some reason that I can't come up with, the part of the code that generates the date boundaries for Google doesn't make use of the "workaround code" (albeit being used elsewhere in the code...I gather that the buggy code is probably the very first part of Erminig written, when I did not hit date problems at that moment).

thank you very much for your patience!

Quote & Reply |

The Following User Says Thank You to lorelei For This Useful Post:
Palleman

lorelei	2010-02-05 , 18:16
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#128

A new version (with an easy-to-use GUI) is now available at
this post.

The version that was available here has been removed, and only the text of this post will remain for historical purposes.

************************************************** ******************

Here is an updated version of Erminig-NG (namely version 0.1.2). It fixes the bug Palleman was having in respect of dates incorrectly formatted.

This is the only bug fix in this new release. This is not a mandatory update, but under specific conditions the bug can be triggered, so I advise to make the update if you have the occasion to.

There's a chance in 60 (very exactly!) that you will get this bug: in fact it is triggered when the timestamp of the last update falls exactly at a whole minute (for instance 12:30:00).

Last edited by lorelei; 2010-03-09 at 08:35. Reason: New version available...

Quote & Reply |

The Following 3 Users Say Thank You to lorelei For This Useful Post:
BrettQ, cibyr, Palleman

shoko	2010-02-05 , 20:01
Posts: 1 \| Thanked: 0 times \| Joined on Feb 2010	#129

Just a quick note: When the sync-calenders option is executed as root (standard for anyone using ssh server) the calendars, the operation runs without a problem (calendars GUI prompts and all) but the calendars aren't created under 'user'. Removing ~/.erminig.db and resyncing as 'user' fixes the problem.

Quote & Reply |

lorelei	2010-02-05 , 21:46
Posts: 479 \| Thanked: 641 times \| Joined on Dec 2007 @ Switzerland	#130

Originally Posted by shoko

Just a quick note: When the sync-calenders option is executed as root (standard for anyone using ssh server) the calendars, the operation runs without a problem (calendars GUI prompts and all) but the calendars aren't created under 'user'. Removing ~/.erminig.db and resyncing as 'user' fixes the problem.

That's true. The path to the erminig.db database is hardcoded as /home/user. I will change that to be $HOME. This shouldn't be an issue with the later releases, where the GUI will be available and normally not run as root.

(fixed in my internal release, will be available in the next public release)

Last edited by lorelei; 2010-02-06 at 00:42.