[Bkchem-user] Re: bkchem 0.9.0-pre1 some bugs

[Beda Kosata]

I am reposting this letter to the mailing list because I thing that 
maybe the answers will be interesting to other people and because it 
rebounces off Piotrs mailserver with :

address@hidden on Wed, 24 Nov 2004 08:22:13 +0100
    The e-mail account does not exist at the organization this message
was sent to.  Check the e-mail address, or contact the recipient
directly to find out the correct address.
    < ns.vscht.cz #5.1.1 SMTP; 550 Brak rekordu SPF/MX nadawcy lub brak
poprawnej autoryzacji SMTP!>



Hi Piotr,
thanks for your letter. The first bug is simple to fix, I just forget to
take care of this case.

Piotr Wawrzyniak wrote:
> Hi Beda.
> I've just downloaded your bkchem, and tested it a little. I have found some 
> bugs, that I thing could be corrected in the next pre release. Here what I've 
> found:
> 1) Templates
> When you don't have any templates yet, next change to templates mode (last 
> icon on right), and then press mouse button on the desk, below error appears:
>  
> #v+
> Error: 4
> IndexError Exception in Tk callback
>   Function: <bound method chem_paper._release1 of <paper.chem_paper object at 
> 0xb6f2f3ec>> (type: <type 'instancemethod'>)
>   Args: (<Tkinter.Event instance at 0xb6d9a50c>,)
>   Event type: ButtonRelease (type num: 5)
> Traceback (innermost last):
>   File "/usr/share/python2.3/site-packages/Pmw/Pmw_1_2/lib/PmwBase.py", line 
> 1747, in __call__
>     return apply(self.func, args)
>   File "/usr/share/python2.3/site-packages/bkchem/paper.py", line 225, in 
> _release1
>     None
>   File "/usr/share/python2.3/site-packages/bkchem/modes.py", line 372, in 
> mouse_up
>     None
>   File "/usr/share/python2.3/site-packages/bkchem/modes.py", line 972, in 
> mouse_click
>     None
>   File "/usr/share/python2.3/site-packages/bkchem/modes.py", line 1043, in 
> _get_transformed_template
>     None
>   File "/usr/share/python2.3/site-packages/bkchem/temp_manager.py", line 81, 
> in get_transformed_template
>     None
> IndexError: list index out of range
>
> ================================================
>   Event contents:
>     char: ??
>     delta: 0
>     height: ??
>     keycode: ??
>     keysym: ??
>     keysym_num: ??
>     num: 1
>     send_event: False
>     serial: 18702
>     state: 256
>     time: 624449
>     type: 5
>     widget: .-1224406964.-1224363924.-1225591572.-1225591828
>     width: ??
>     x: 324.0
>     x_root: 475
>     y: 147.0
>     y_root: 348
> #v-
>
> 2) New pdf, ps export:
> When you export using these two new plugins, and when anything is selected, 
> this in pdf/ps file one see also selections, it means these selection points. 
> Maybe it is necessary to do self.paper.unselect_all() at first, or something 
> other, that new library neads.

I know about it and you are right about the solution :)  However I
thought I will maybe leave it there to make it possible to do something
like "PDF screenshot". However it is probably not the way people want to
have it :) so I will fix it and probably add some switch for those who
would like to have the "screenshot mode".

> 3) The templates
> The templates works quite fine. But I have some suggestions. The first maybe 
> you should change a shortcut for selecting template atom. At present both 
> hands are necessary to do it. I know, I overreact a little bit, but this 
> shortcut is imho uncomfortable. Maybe two letters close one another i.e. 
> ctrl+t+r. It could be also good to have this option after pressing the right 
> mouse button.
> And another thing about templates. At present one have to select a template 
> atom. It works fine but, imply many templates for the same molecule (if it 
> has many bonding places). A solution could be to make something like drawing 
> mode. At this mode, if two molecules will be close enough, it should be 
> possible to connect them or to some atom, or bond, for example one could 
> paint a naphthalene using two separate benzens ring, just by moving one to 
> ring to the another. I release that sometimes such a feature could be bad 
> idea, and thats why I suggest to make a special switch i menu for that.  I 
> hope I explained what I am thinking about quite clearly. I haven't look into 
> source code, so I don't know how difficult will it be to implement.
Well, you are right that I could replace the shortcut, I thing something
like C-t would be enough :)  It was there mainly for historical reasons
(this function was in bkchem from version 0.1.0, however not usefull
because no user template mode was supported). On the other hand, you
don't need to push all the keys at one go you can do "C-t" then release
everything and then do "C-1" and it will work, its two independent
keystrokes.
One thing about your template suggestion - it works :)  If you put one 
benzene ring over a bond in another (the tolerance is a few pixels) 
bkchem will automatically concatenate it. In fact the template code and 
the code for concatenation share most of lines.

> 4. It could be useful to make one more option during instalation. Something 
> similar  to DESTDIR, for Makefiles. At present one can use option --root, but 
> it also changes paths in executables (site_config.py, bkchem.py). Such thing 
> could simplify making packages for distribution (rpm, deb, etc). At present I 
> must 'sed' files (I know there are only two) and replaces paths.

I will try to do something about it :)

> Wow I have made quite long letter, I hope not so boring. You do really good 
> job, at present I am writing publication using your software, so keep doing 
> your job. I will help you as much as I can, to make bkchem the better 
> software. 
>
> Best regards
> piotr.
Thanks again for the bugreports and suggestions (and also the icon - its
nice, clean and certainly better than anything I would come with as I
don't have time for it :)

                Cheers
                        BEDA