[fityk-devel] New version of xylib code has been submitted

4 views
Skip to first unread message

张鹏

unread,
Aug 17, 2007, 10:20:01 PM8/17/07
to Marcin Wojdyr, fityk-dev ML
Hi, Marcin:

The following are the changes from last version:
/////////////////////////////////////////////////////////////////////////////////////////////////////////////
* Changed to use Class Column to represent a 1-D data
* If SPE is linear, set fixed_step = true
* Export_xy_file is modified with necessary changes, and output column
names of x, y, stddev if they exist
* Depressed exceptions in check(): only return a boolean value.
* Changed to Read multiple columns in a range of pdCIF format, the
uncertainty is treated as a separate column
* Moved stream& arg from Ctors to load_data() member functions. Got
rid of the stream &/* member f in DataSet, passed to other members
functions if necessary
* Added the clear() function in DataSet.
* Re-wrote one of the uxd-like formats: Rigaku_dat
* Some other trivial changes
/////////////////////////////////////////////////////////////////////////////////////////////////////////////

Next to do:
I will do in the following order. If you think the 2nd one is more
important, tell me ASAP.

* PANalytical's data format
* find more sample files from other source to test existing code

BTW:
I have only re-written one of the UxdLikeDataSet - Rigaku_dat format.
I'd like to know what you think of such an implementation. If you
think this is fine, I will apply the same strategy to the UXD format.

Thanks

--
Regards,

Peng ZHANG (张鹏)

Marcin Wojdyr

unread,
Aug 18, 2007, 4:28:09 AM8/18/07
to 张鹏, fityk-dev ML
On 8/17/07, 张鹏 <zhangp...@gmail.com> wrote:
> Next to do:
> I will do in the following order. If you think the 2nd one is more
> important, tell me ASAP.
>
> * PANalytical's data format
> * find more sample files from other source to test existing code

There is only a few days before GSOC final evaluation. The library is
not tested well, I'm sure there are bugs there.

>
> BTW:
> I have only re-written one of the UxdLikeDataSet - Rigaku_dat format.
> I'd like to know what you think of such an implementation. If you
> think this is fine, I will apply the same strategy to the UXD format.

it's OK.
The value returned by new should _not_ be checked, see
http://www.cprogramming.com/tips/showTip.php?tip=64&count=30&page=0

Marcin

--
Marcin Wojdyr | http://www.unipress.waw.pl/~wojdyr/

Reply all
Reply to author
Forward
0 new messages