Re: Call for testing before 0.95.4
On Tue, 13 Nov 2012, Rodolfo García Peñas (kix) wrote: there are two problems: 1. xfig don't have the WM_ICON_NAME set, therefore the function wIconPaint() don't draw the title. You can try to set the value with xprop, reload windowmaker and then the title will appear in xfig's icon. The update pixmap function don't paint the background square because the variable titled is created using the WM_ICON_NAME value too. This may be unrelated and was already discussed but as a summary here's what I think about this: 1. xfig does set WM_ICON_NAME (at least on my system) but it sets it to the name of the edited file which may be empty if "No file" is open. (Note that it is slightly different that the property is not set and it is set but has an empty string as a value.) 2. A possible fallback for missing icon name may be the WM_CLASS which is always set and usually shorter than say WM_NAME (which might also be missing and need a fallback itself) but what you do is up to you. I don't mind having icons without a title and so far you seem to be the only one annoyed by this, at least noone else said anything to this topic. Probably you could make a patch which allows using a fallback but is disabled by default and has to be enabled if one prefers. 2. As I said in my final mail about icons (that was not my last mail about icons :-) ), the icon image and the window image are different. icon image is window->icon and window image is window->net_icon_image. If the user sets the icon, then the selected icon is used in the window (alt+tab) and in the icon, but if is not set, then the icons could be different. I couldn't really follow this and didn't check the code either. Just trying to test it from the users' point of view so I can't comment on this. I sent a patch yesterday. The problem is that the wIconValidateIconSize didn't check the width and the height size for the icon, only checked one of them. The new patch check both values and holds the icon aspect ratio. I've tested your patch but it didn't change anything. gtk-demo still has an icon which is too large for the tile and appears clipped. I did some more testing and found that the Apply button does not seem to do anything (after setting an icon I have to restart the app to have the icon changed) and that the Ignore client supplied icon option seems to work for gtk-demo but not for xfig. Don't know the reason for any of these just letting you know what I saw. Regards, BALATON Zoltan
Re: Call for testing before 0.95.4
On 2012-11-12 23:00, BALATON Zoltan wrote: On Mon, 12 Nov 2012, Carlos R. Mafra wrote: Please do test the #next branch and report any problems (if any). There has been a lot of code churn related to icons. Even though they look relative safe, regressions might be hiding somewhere. I'm not sure if the "Ignore client supplied icon" option in the Icon and Initial Workspace inspector is still working. At least Xfig now has the icon I set in the app icon but displays its own icon in the miniwindow (which is obscuring the tile completely). Hi again, there are two problems: 1. xfig don't have the WM_ICON_NAME set, therefore the function wIconPaint() don't draw the title. You can try to set the value with xprop, reload windowmaker and then the title will appear in xfig's icon. The update pixmap function don't paint the background square because the variable titled is created using the WM_ICON_NAME value too. 2. As I said in my final mail about icons (that was not my last mail about icons :-) ), the icon image and the window image are different. icon image is window->icon and window image is window->net_icon_image. If the user sets the icon, then the selected icon is used in the window (alt+tab) and in the icon, but if is not set, then the icons could be different. Comments about these problems are very welcome. Again, if WM_ICON_NAME is not set, probably we could use something here. Related to this, if the client supplied icon is too big (like for gtk-demo for example) it is simply clipped and not resized. An improvement though that it now seems to be centered at least. Could the icon be resized (maybe before being cached) to improve this. I sent a patch yesterday. The problem is that the wIconValidateIconSize didn't check the width and the height size for the icon, only checked one of them. The new patch check both values and holds the icon aspect ratio. Best regards, kix (snip) Regards, BALATON Zoltan -- ||// //\\// Rodolfo "kix" Garcia ||\\// //\\ http://www.kix.es/ -- To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.org.
Re: Call for testing before 0.95.4
On Mon, 12 Nov 2012, BALATON Zoltan escribió:
> On Mon, 12 Nov 2012, Carlos R. Mafra wrote:
> >Please do test the #next branch and report any problems (if any).
> >There has been a lot of code churn related to icons. Even though they
> >look relative safe, regressions might be hiding somewhere.
>
> I'm not sure if the "Ignore client supplied icon" option in the Icon
> and Initial Workspace inspector is still working. At least Xfig now
> has the icon I set in the app icon but displays its own icon in the
> miniwindow (which is obscuring the tile completely).
Can you try this:
kix@osaka:~/src/wmaker/wmaker-crm/src$ git diff
diff --git a/src/icon.c b/src/icon.c
index ba059ad..39e4b48 100644
--- a/src/icon.c
+++ b/src/icon.c
@@ -359,10 +359,15 @@ RImage *wIconValidateIconSize(RImage *icon, int max_size)
return NULL;
/* We should hold "ICON_BORDER" (~2) pixels to include the icon border
*/
- if ((icon->width - max_size) > -ICON_BORDER ||
- (icon->height - max_size) > -ICON_BORDER) {
- nimage = RScaleImage(icon, max_size - ICON_BORDER,
-(icon->height * (max_size - ICON_BORDER) /
icon->width));
+ if (((max_size - ICON_BORDER) < icon->width) ||
+ ((max_size - ICON_BORDER) < icon->height)) {
+ if (icon->width > icon->height)
+ nimage = RScaleImage(icon, max_size - ICON_BORDER,
+(icon->height * (max_size -
ICON_BORDER) / icon->width));
+ else
+ nimage = RScaleImage(icon, icon->width * (max_size -
ICON_BORDER) / icon->height,
+max_size - ICON_BORDER);
+
RReleaseImage(icon);
icon = nimage;
}
Thanks,
kix
> Related to this, if the client supplied icon is too big (like for
> gtk-demo for example) it is simply clipped and not resized. An
> improvement though that it now seems to be centered at least. Could
> the icon be resized (maybe before being cached) to improve this.
>
> Some other things in Wprefs.app which have been there before but I found
> them now:
>
> In Menu Preferences the "Note: this is annoying" text appears clipped
> which is well... annoying. Could this line be removed and just let the
> users decide what they prefer? (I could make a patch if you want, it's
> localised in one file and some translations.)
>
> Also in Expert User Preferences there are two options:
> "Bounce Appicons when the application wants attention."
> and then a few lines further
> "Do not make Appicons bounce."
> This is confusing and I'm not even sure what these mean and why we have
> two seemingly overlapping options without looking up the code.
>
> Other than these minor things it seems to work well so far and has some
> nice improvements like the fix for the few pixel offset of maximised
> windows due to borders. Thanks for the great work on Window Maker.
>
> Regards,
> BALATON Zoltan
>
>
> --
> To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.info.
--
||// //\\// Rodolfo "kix" Garcia
||\\// //\\ http://www.kix.es/
--
To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.org.
Re: Call for testing before 0.95.4
On Tue, 13 Nov 2012, BALATON Zoltan escribió: > On Tue, 13 Nov 2012, Rodolfo García Peñas (kix) wrote: > >there are two problems: > > > >1. xfig don't have the WM_ICON_NAME set, therefore the function > >wIconPaint() don't draw the title. You can try to set the value > >with xprop, reload windowmaker and then the title will appear in > >xfig's icon. The update pixmap function don't paint the background > >square because the variable titled is created using the > >WM_ICON_NAME value too. > > This may be unrelated and was already discussed but as a summary > here's what I think about this: > > 1. xfig does set WM_ICON_NAME (at least on my system) but it sets it > to the name of the edited file which may be empty if "No file" is > open. (Note that it is slightly different that the property is not > set and it is set but has an empty string as a value.) > > 2. A possible fallback for missing icon name may be the WM_CLASS > which is always set and usually shorter than say WM_NAME (which > might also be missing and need a fallback itself) but what you do is > up to you. I don't mind having icons without a title and so far you > seem to be the only one annoyed by this, at least noone else said > anything to this topic. Probably you could make a patch which allows > using a fallback but is disabled by default and has to be enabled if > one prefers. > > >2. As I said in my final mail about icons (that was not my last > >mail about icons :-) ), the icon image and the window image are > >different. icon image is window->icon and window image is > >window->net_icon_image. If the user sets the icon, then the > >selected icon is used in the window (alt+tab) and in the icon, but > >if is not set, then the icons could be different. > > I couldn't really follow this and didn't check the code either. Just > trying to test it from the users' point of view so I can't comment > on this. > > >I sent a patch yesterday. The problem is that the > >wIconValidateIconSize didn't check the width and the height size > >for the icon, only checked one of them. The new patch check both > >values and holds the icon aspect ratio. > > I've tested your patch but it didn't change anything. gtk-demo still > has an icon which is too large for the tile and appears clipped. I > did some more testing and found that the Apply button does not seem > to do anything (after setting an icon I have to restart the app to > have the icon changed) and that the Ignore client supplied icon > option seems to work for gtk-demo but not for xfig. Don't know the > reason for any of these just letting you know what I saw. Yes, I installed gtk-demo and you are right :-/ The problem is the icon is not resized. @@ -476,6 +478,8 @@ static RImage *get_wwindow_image_from_wmhints(WWindow *wwin, WIcon *icon) (hints->flags & IconMaskHint) ? hints->icon_mask : None); + image = wIconValidateIconSize(image, wPreferences.icon_size); + return image; } I will check tomorrow the other methods to create icons, to check if they include the icon size validation. Anyway, I am so happy with the new code, because as you can see, is more easy, but yes, some things could be better. Alt+tab show a correct icon too. Thanks A LOT for your report. kix PS. The patch I sent yesterday is ok too. You can test it with xcalc, the icon without that patch is not correctly painted. PS2. I will reply your other questions tomorrow. > Regards, > BALATON Zoltan -- To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.org.
Re: Call for testing before 0.95.4
On Mon, 12 Nov 2012, BALATON Zoltan wrote: On Mon, 12 Nov 2012, Carlos R. Mafra wrote: I should look at git blame output to be sure, but iirc it is not enough to not tick the "Bounce Appicons when..." option for the appicons not to bounce. Someone sent a patch to explicitly disable bouncing afterwards. I think that the current need for the two options are a sign of some mistake in the code. Maybe or at least the text describing the options should be improved to make it clearer why there are two of them. Trying to improve this I ended up with a patch that reorganises the options and gathers all bounce options at one place I've found on the Misc. Ergonomic Prefs page. The two options previously there were not really related and one of them about focus has a better place on the corresponding page where it is moved now and the other one related to XKB (which did not even show up on my config) was moved to expert instead. I think this makes it a bit better but I still could not find other labels for the options (nor am I sure what they really do after looking at the code). But maybe it's less confusing now and we can live with it. (Ususally people change it once after installing and then forget about it anyway.) Regards, BALATON Zoltan -- To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.org.
Re: Call for testing before 0.95.4
On Wed, 14 Nov 2012, Rodolfo García Peñas wrote: Anyway, I am so happy with the new code, because as you can see, is more easy, but yes, some things could be better. Alt+tab show a correct icon too. Thanks A LOT for your report. I'm happy with the improvements too and the problems I've sent are not that serious indeed but fixing them could make it even better. PS2. I will reply your other questions tomorrow. Don't rush it. I think we don't have a hard deadline and we can make the release a few days or weeks later if that allows us to fix more bugs and make a better release. So take your time to think about it and sleep when you need. :-) Regards, BALATON Zoltan
Re: Call for testing before 0.95.4
../../src/actions.c:38:22: error: colormap.h: No such file or directory ../../src/actions.c: In function 'wSetFocusTo': ../../src/actions.c:159: warning: implicit declaration of function 'wColormapInstallForWindow' make[4]: *** [actions.o] Error 1 make[4]: Leaving directory `/tmp/wmaker-crm/WindowMaker-0.95.3/_build/src' make[3]: *** [all] Error 2 make[3]: Leaving directory `/tmp/wmaker-crm/WindowMaker-0.95.3/_build/src' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/tmp/wmaker-crm/WindowMaker-0.95.3/_build' make[1]: *** [all] Error 2 make[1]: Leaving directory `/tmp/wmaker-crm/WindowMaker-0.95.3/_build' make: *** [distcheck] Error 1 zsh: exit 2 make distcheck make distcheck 36.28s user 7.02s system 105% cpu 41.046 total make distcheck is exactly what I was looking for, thank you! -- John H. Robinson, IV On 2012-11-14, at 18:18, BALATON Zoltan wrote: > On Wed, 14 Nov 2012, John H. Robinson, IV wrote: >> Please find attached a diff to add the missing .h files from src/Makefile.am >> One way to find these: >> $ git checkout next >> $ ./autogen.sh >> $ ./configure >> $ make dist-gzip >> $ tar xzf WindowMaker-0.95.3.tar.gz >> $ cd WindowMaker-0.95.3 >> $ ./configure >> $ make >> There has to be an easier way. > > How about 'make distcheck'? (I haven't tried it but this is the official > autotools way that's supposed to do that.) > > Regards, > BALATON Zoltan
Re: Call for testing before 0.95.4
Quoth Paul Seelig, Yet another rather strange glitch: - open three terminal windows - repeatedly restart wmaker - all windows slowly drift to the left and up by just a few pixels That sounds similar to the bug which was fixed in b657f15344f193d70747689f96cefddc8baa2181. I'll take a look at it. -- To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.org.
(fwd) Re: Call for testing before 0.95.4
I didn't see this mail in the mail list archives. Was sent ok?
if is ok for you I will sent the patch. I will work today in the xfig problem.
Cheers,
kix
--
||// //\\// Rodolfo "kix" Garcia
||\\// //\\ http://www.kix.es/
--- Begin Message ---
On Mon, 12 Nov 2012, BALATON Zoltan escribió:
> On Mon, 12 Nov 2012, Carlos R. Mafra wrote:
> >Please do test the #next branch and report any problems (if any).
> >There has been a lot of code churn related to icons. Even though they
> >look relative safe, regressions might be hiding somewhere.
>
> I'm not sure if the "Ignore client supplied icon" option in the Icon
> and Initial Workspace inspector is still working. At least Xfig now
> has the icon I set in the app icon but displays its own icon in the
> miniwindow (which is obscuring the tile completely).
Can you try this:
kix@osaka:~/src/wmaker/wmaker-crm/src$ git diff
diff --git a/src/icon.c b/src/icon.c
index ba059ad..39e4b48 100644
--- a/src/icon.c
+++ b/src/icon.c
@@ -359,10 +359,15 @@ RImage *wIconValidateIconSize(RImage *icon, int max_size)
return NULL;
/* We should hold "ICON_BORDER" (~2) pixels to include the icon border
*/
- if ((icon->width - max_size) > -ICON_BORDER ||
- (icon->height - max_size) > -ICON_BORDER) {
- nimage = RScaleImage(icon, max_size - ICON_BORDER,
-(icon->height * (max_size - ICON_BORDER) /
icon->width));
+ if (((max_size - ICON_BORDER) < icon->width) ||
+ ((max_size - ICON_BORDER) < icon->height)) {
+ if (icon->width > icon->height)
+ nimage = RScaleImage(icon, max_size - ICON_BORDER,
+(icon->height * (max_size -
ICON_BORDER) / icon->width));
+ else
+ nimage = RScaleImage(icon, icon->width * (max_size -
ICON_BORDER) / icon->height,
+max_size - ICON_BORDER);
+
RReleaseImage(icon);
icon = nimage;
}
Thanks,
kix
> Related to this, if the client supplied icon is too big (like for
> gtk-demo for example) it is simply clipped and not resized. An
> improvement though that it now seems to be centered at least. Could
> the icon be resized (maybe before being cached) to improve this.
>
> Some other things in Wprefs.app which have been there before but I found
> them now:
>
> In Menu Preferences the "Note: this is annoying" text appears clipped
> which is well... annoying. Could this line be removed and just let the
> users decide what they prefer? (I could make a patch if you want, it's
> localised in one file and some translations.)
>
> Also in Expert User Preferences there are two options:
> "Bounce Appicons when the application wants attention."
> and then a few lines further
> "Do not make Appicons bounce."
> This is confusing and I'm not even sure what these mean and why we have
> two seemingly overlapping options without looking up the code.
>
> Other than these minor things it seems to work well so far and has some
> nice improvements like the fix for the few pixel offset of maximised
> windows due to borders. Thanks for the great work on Window Maker.
>
> Regards,
> BALATON Zoltan
>
>
> --
> To unsubscribe, send mail to wmaker-dev-unsubscr...@lists.windowmaker.info.
--
||// //\\// Rodolfo "kix" Garcia
||\\// //\\ http://www.kix.es/
--- End Message ---
