Resize works fine now, Bill, with TABULA under jgtk on my mac. Thanks. BTW I'm getting 12 pairs of messages as TABULA is loaded like this:
(<unknown>:3206): Gtk-CRITICAL **: gtk_widget_get_parent: assertion `GTK_IS_WIDGET (widget)' failed (<unknown>:3206): Gtk-CRITICAL **: gtk_widget_show: assertion `GTK_IS_WIDGET (widget)' failed I can't recall seeing them before. They appear in jconsole, not in Term. They don't seem to be associated with any error the end-user can see. Resizing does not cause any more to appear. Likewise anything else I've tried with the TABULA form. If you don't recognise where they're coming from, shall I probe deeper? Ian On Sun, Mar 18, 2012 at 5:11 AM, bill lam <bbill....@gmail.com> wrote: > An update to gtkwd for form resize, this version works quite well > for linux, but my tiling WM does not support resie using mouse drag > so that I cannot test it. I hope it will show some improvements in > your Mac. > > Сбт, 17 Мар 2012, Ian Clark писал(а): >> > Unfortunately, not much can be done for form resize. >> >> Doesn't matter too much for TABULA. It's probably only me it affects. >> Haven't tried it under Windows. >> >> AFAICT pmovex applied to the whole form is fast and reliable, and it >> seems to recover a misaligned form. Could a roll-your-own "resize" be >> mashed-up by calling pmovex repeatedly? Or we could go back to the old >> 1980s draggable marquee effect? >> >> >> On Sat, Mar 17, 2012 at 10:17 AM, bill lam <bbill....@gmail.com> wrote: >> > Unfortunately, not much can be done for form resize. >> > >> > modifiers bug caused by a regression in gui/gtk, now fixed. >> > >> > Сбт, 17 Мар 2012, Ian Clark писал(а): >> >> > ** Help should now work >> >> >> >> It does. Thanks. >> >> >> >> >> >> > ** resize is nasty and I don't know its implementation in J602. >> >> > However resize in gtkwd works reasonably well from my experience. >> >> > Can you give some screen shots to show the differencesr? >> >> >> >> See: http://www.jsoftware.com/jwiki/JgtkwdResizeBug >> >> (you can delete this page when you don't need it any more.) >> >> >> >> >> >> > ** Toolbar icon color >> >> >> >> That looks better now, thanks. >> >> >> >> >> >> > ** button: Append in tab: Consts applies itself twice. >> >> >> >> This bug has fixed itself. Thanks. >> >> >> >> >> >> > ** fcopynews >> >> > There is fcopynew in both j6 and j7, can you use it instead? >> >> >> >> Yes, I can code around this. (if "fcopynews" has been discontinued in j7) >> >> >> >> >> >> Just one TABULA bug remains, that I can see... >> >> >> >> shift_tab_ is broken (therefore holding shiftkey down does not alter >> >> the toolbar icons' function) >> >> I'd say sysmodifiers is no longer being updated. >> >> (I think this was working before. But I'm not sure.) >> >> >> >> On Tue, Mar 6, 2012 at 6:22 AM, bill lam <bbill....@gmail.com> wrote: >> >> > Thank you for the report. Please get the update versions. >> >> > >> >> > ** Help should now work >> >> > previously wdselect did not work for tab, now fixed. >> >> > setshow will be no-op for tab group in gtkwd by design. But you probably >> >> > do not neet it because show/hide tabs is now automatically done by >> >> > selecting the tab. >> >> > >> >> > ** resize is nasty and I don't know its implementation in J602. >> >> > However resize in gtkwd works reasonably well from my experience. >> >> > Can you give some screen shots to show the differencesr? >> >> > >> >> > ** Toolbar icon color >> >> > Now it changed to use argb (same as that in j602), the color of + >> >> > and x in toolbar is blue if I am not mistaken. >> >> > >> >> > furthermore, if you change the value of RGBSEQ_j_ defined inside >> >> > gui/gtk to 0 for abgr format, it should still work. >> >> > >> >> > ** button: Append in tab: Consts applies itself twice. >> >> > I cannot reproduce this. Is this mac specific? >> >> > >> >> > ** wd 'setselect panel 11' >> >> > wdselect now fixed >> >> > >> >> > ** fcopynews >> >> > There is fcopynew in both j6 and j7, can you use it instead? >> >> > >> >> > I think you test it on mac, just want to confirm you are using gtk2. >> >> > size and position of gtkwd under gtk3 are rather ugly. TODO. >> >> > >> >> > Пнд, 05 Мар 2012, Ian Clark писал(а): >> >> >> Congratulation to Bill Lam for gui/gtkwd, which has given a flying >> >> >> start to porting math/tabula to j7, not to mention many treasured j6 >> >> >> demos. >> >> >> >> >> >> A few j6/j7 wd incompatibilities revealed by math/tabula running under >> >> >> jgtk... >> >> >> >> >> >> To see these: >> >> >> load 'math/tabula' NB. same in both j6 and jgtk >> >> >> >> >> >> ** Help -won't go to Info tab (in fact all verbs which are meant to >> >> >> change tabs don't do so). >> >> >> Menu: Help > Help for 1-char commands -executes: hlpc_tab_'' >> >> >> setshow_tab_'' has become inoperative (setshow 4 should open >> >> >> tab 4 ("inf")) >> >> >> because wd'setshow inf 4' and wd'setselect tabs 4' appear to be >> >> >> no-ops. >> >> >> >> >> >> ** resize goes wrong, especially if bottom-right corner dragged too >> >> >> fast. >> >> >> (affects controls with: rightmove and bottommove) >> >> >> >> >> >> ** Toolbar icons' colors are corrupted because rgb codes get >> >> >> interpreted as bgr >> >> >> (this affects: glpixels, but I don't know what else) >> >> >> >> >> >> ** button: Append in tab: Consts applies itself twice. >> >> >> But, changing... >> >> >> consts_cappend_button=: newc >> >> >> to... >> >> >> consts_cappend_button=: smoutput bind 'hi' >> >> >> -shows the handler is only called once by the mouseclick. >> >> >> But somehow, newc gets called twice. >> >> >> (I suspect it's tied up with the wd-operation taking place in a >> >> >> currently hidden tab.) >> >> >> >> >> >> ** wd 'setselect panel 11' >> >> >> --(if line 11 is absent) causes the message to appear in the >> >> >> associated jconsole: >> >> >> (<unknown>:9592): Gtk-CRITICAL **: gtk_list_store_get_path: >> >> >> assertion >> >> >> `iter->stamp == GTK_LIST_STORE (tree_model)->stamp' failed >> >> >> whereas j6 forgives this "error". >> >> >> >> >> >> I'm hoping the location of each bug will be clear from the above. For >> >> >> those that aren't, I'll happily isolate the bug into a demo script. >> >> >> >> >> >> I note too that verb: fcopynews_z_ (used by: cal and formerly loaded >> >> >> by: require 'files') is undefined in jgtk. However fcopynew_z_ is >> >> >> still there, but now defined in stdlib. Is the absence of fcopynews in >> >> >> jgtk deliberate, or an oversight? >> >> >> ---------------------------------------------------------------------- >> >> >> For information about J forums see http://www.jsoftware.com/forums.htm >> >> > >> >> > -- >> >> > regards, >> >> > ==================================================== >> >> > GPG key 1024D/4434BAB3 2008-08-24 >> >> > gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3 >> >> > ---------------------------------------------------------------------- >> >> > For information about J forums see http://www.jsoftware.com/forums.htm >> >> ---------------------------------------------------------------------- >> >> For information about J forums see http://www.jsoftware.com/forums.htm >> > >> > -- >> > regards, >> > ==================================================== >> > GPG key 1024D/4434BAB3 2008-08-24 >> > gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3 >> > ---------------------------------------------------------------------- >> > For information about J forums see http://www.jsoftware.com/forums.htm >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > -- > regards, > ==================================================== > GPG key 1024D/4434BAB3 2008-08-24 > gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3 > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
