[gdal-dev] New config option for date/time output
Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] New config option for date/time output
15.03.2016, 11:15, Ari Jolma kirjoitti: Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. It seems that the standard way of formatting time offset is +hhmm or -hhmm (without ':' between hours and minutes) (ISO 8601:2000). Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] New config option for date/time output
GMT uses "yyy-mm-ddT[hh:mm:ss] (Gregorian) or -Www-ddT[hh:mm:ss] (ISO)" http://gmt.soest.hawaii.edu/doc/latest/gmt.conf.html#calendar-time-parameters It would be nice to use the same. Joaquim Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] New config option for date/time output
15.03.2016, 14:08, Joaquim Luis kirjoitti: GMT uses "yyy-mm-ddT[hh:mm:ss] (Gregorian) or -Www-ddT[hh:mm:ss] (ISO)" http://gmt.soest.hawaii.edu/doc/latest/gmt.conf.html#calendar-time-parameters It would be nice to use the same. I'm not sure I understand you correctly but strftime can do this: OGR_DATETIME_FORMAT = "%Y-%m-%dT%H:%M:%S" gives for example 2016-01-10T08:25:12 and OGR_DATETIME_FORMAT = "%Y-%W-%dT%H:%M:%S" gives 2016-01-10T08:25:12 for the same time - 10.1.2016 is on week two (strftime: %W = Week number with the first Monday as the first day of week one (00-53)) I think that there needs to be separate options for date and datetime, i.e., OGR_DATE_FORMAT and OGR_DATETIME_FORMAT. Ari Joaquim Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] New config option for date/time output
On Tue, 15 Mar 2016 13:01:29 -, Ari Jolma wrote: 15.03.2016, 14:08, Joaquim Luis kirjoitti: GMT uses "yyy-mm-ddT[hh:mm:ss] (Gregorian) or -Www-ddT[hh:mm:ss] (ISO)" http://gmt.soest.hawaii.edu/doc/latest/gmt.conf.html#calendar-time-parameters It would be nice to use the same. I'm not sure I understand you correctly but strftime can do this: OGR_DATETIME_FORMAT = "%Y-%m-%dT%H:%M:%S" gives for example 2016-01-10T08:25:12 and OGR_DATETIME_FORMAT = "%Y-%W-%dT%H:%M:%S" gives 2016-01-10T08:25:12 for the same time - 10.1.2016 is on week two (strftime: %W = Week number with the first Monday as the first day of week one (00-53)) Sorry, it was me who misunderstood your question/suggestion. I thought that printing with the above format was not possible yet and that's why I pointed to what GMT does. But if GDAL can do it already, than sorry for the noise. Joaquim I think that there needs to be separate options for date and datetime, i.e., OGR_DATE_FORMAT and OGR_DATETIME_FORMAT. Ari Joaquim Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] New config option for date/time output
15.03.2016, 15:59, Joaquim Luis kirjoitti: On Tue, 15 Mar 2016 13:01:29 -, Ari Jolma wrote: 15.03.2016, 14:08, Joaquim Luis kirjoitti: GMT uses "yyy-mm-ddT[hh:mm:ss] (Gregorian) or -Www-ddT[hh:mm:ss] (ISO)" http://gmt.soest.hawaii.edu/doc/latest/gmt.conf.html#calendar-time-parameters It would be nice to use the same. I'm not sure I understand you correctly but strftime can do this: OGR_DATETIME_FORMAT = "%Y-%m-%dT%H:%M:%S" gives for example 2016-01-10T08:25:12 and OGR_DATETIME_FORMAT = "%Y-%W-%dT%H:%M:%S" gives 2016-01-10T08:25:12 for the same time - 10.1.2016 is on week two (strftime: %W = Week number with the first Monday as the first day of week one (00-53)) Sorry, it was me who misunderstood your question/suggestion. I thought that printing with the above format was not possible yet and that's why I pointed to what GMT does. But if GDAL can do it already, than sorry for the noise. It can't. That's why I'm suggesting this. The core of my suggestion is to start using the strftime function and the config options. I've actually implemented it already for testing, that's why I may sound like that GDAL can do it already. Cheers, Ari Joaquim I think that there needs to be separate options for date and datetime, i.e., OGR_DATE_FORMAT and OGR_DATETIME_FORMAT. Ari Joaquim Folks, The OGRFeature::GetFieldAsString returns date/time fields formatted in non-standard way. For example date is output as "year/month/day" as more usual "year-month-day". This causes problems in many programs. I suggest adding a new config option OGR_DATE_FORMAT, which, if set, would make the method return date/time formatted according to it. I suggest that we use the strftime and its format argument as the target for the new option. The implementation is complicated a bit when milliseconds and timezones are involved since they are not included in the struct tm used by strftime. I suggest that we assume seconds are formatted as :\d\d in the format and simply add milliseconds (as .\d\d\d) and timezone (as [+-]\d\d(:\d\d)), if they exists, after the second. Do you think this needs a RFC? I'm willing to implement the changes. Ari ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev ___ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
[gdal-dev] [Windows/ C#] Gdal dll conflicts
Hello,
I'm currently deploying a Gdal-based application on various Windows-PCs.
For some of them, there are integration problems, due to dll conflicts. As an
example:
1) Here is my app directory structure :
MyApp
|
|-- gdal
||
||- gdal201.dll
||- libeay32.dll
||- ssleay32.dll
||- ...
|
|-- myApp.exe
|-- othersDll.dll
2) On some of the PCs I use, libeay32.dll and ssleay32.dll are also
present on C:\Windows\SysWOW64 folder, in versions different from mines.
3) When the app loads gdal201.dll, dependencies are loaded first in
C:\Windows\SysWOW64 folder ==> C:\Windows\SysWOW64 \libeay32.dll and
ssleay32.dll are loaded, that causes a crash.
I checked that :
- In GdalConfiguration.cs, GdalConfiguration(), my PATH env var is
correctly overriden
var gdalPath = Path.Combine(executingDirectory, "gdal-2.0.1");
var nativePath = Path.Combine(gdalPath, GetPlatform());
// Prepend native path to environment path, to ensure the
// right libs are being used.
var path = Environment.GetEnvironmentVariable("PATH");
path = nativePath + ";" + Path.Combine(nativePath, "plugins") + ";" + path;
Environment.SetEnvironmentVariable("PATH", path);
- Using a file access monitoring tool (ProcessMonitor, from
Microsoft), I observe that dll search order is :
o myApp.exe directory
o C:\windows\ SysWOW64
o C:\windows\System32
o C:\...\MyApp\gdal-2.0.1
ð This is not compliant with PATH overriding
After browsing internet (including the mailing-list :)), I don't want to :
- Compile gdal statically,
- Copy gdal dependencies in myApp.exe root directory,
- Comment some lines of my makefile in order to avoid dependencies
such as libeay32.dll or ssleay32.dll.
Have you got ideas about this problem ?
--
Thank you, Jérome Siot
(Excuse me for my awful English, I'm French !)
--
___
gdal-dev mailing list
gdal-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/gdal-dev
Re: [gdal-dev] GSOC 2016
Hi Sarthak,
The first version is not working (do you test it?):
https://github.com/sarthak-0415/gdal/commit/36344cc26f23202cb289390322c1d295697136bd#diff-31df0e62d00ca09f9f11ad2f29e94b54R2541
Here you try to get array value with index -1. You need to set
ppszDbname = NULL no DB name present in input parameters.
The second variant is not working too:
>>> ds = gdal.Open('PG:')
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_M_construct null not valid
In both cases there is a problem here:
https://github.com/sarthak-0415/gdal/blob/6264d3fc52242fdce858547cc3a0312b04fd638b/gdal/frmts/postgisraster/postgisrasterdataset.cpp#L2743
Also look there ppszDbname is using, as before modifications the code
expect that ppszDbname cannot be NULL.
Best regards,
Dmitry
15.03.2016 13:13, sarthak agarwal пишет:
Hey Dmitry ,
As discussed on the IRC yesterday,
I made the changes in the code.
I build two versions of the code
1.
The changes suggested by you (to use the old trunk code and remove
the additional checks) link
travis .
a. in this version the Dbname is left empty if not provided by the
user.
2.
The version in which we
a. if the Dbname is provided by the user then ppzDbname=Dbname.
b. else use the psql env var PGDATABASE
c. else use the Username as the database name.
d. if nothing is available then pass empty string.
e link
travis
I think both version should work
Regards,
Sarthak
___
gdal-dev mailing list
gdal-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/gdal-dev
