Re: help request on proposed VCL code patch
On 12/14/2015 06:46 AM, Lionel Elie Mamane wrote: > Was Caolan's commit, you said you would test "not before this weekend" > (a month ago), and to please poke you again "by Sunday" but nobody > poked you. (The poke fell through the list of things *I* have to do.) > > http://lists.freedesktop.org/archives/libreoffice/2015-November/071038.html Ah yes. My mistake (thanks for the reference.hundred+ emails a day makes things get buried). Is a test still really needed or is the regression thereand therefore we're past the stage of testing? Best, Joel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Mon, Dec 14, 2015 at 07:13:43AM -0800, Joel Madero wrote: > On 12/14/2015 06:46 AM, Lionel Elie Mamane wrote: >> Was Caolan's commit, you said you would test "not before this weekend" >> (a month ago), and to please poke you again "by Sunday" but nobody >> poked you. (The poke fell through the list of things *I* have to do.) >> http://lists.freedesktop.org/archives/libreoffice/2015-November/071038.html > Ah yes. My mistake (thanks for the reference.hundred+ emails a > day makes things get buried). Is a test still really needed or is > the regression thereand therefore we're past the stage of > testing? Regression is there. We are about to release 5.0.4.2 as final with that regression. -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Mon, Dec 14, 2015 at 08:37:55PM +, Caolán McNamara wrote: > On Mon, 2015-12-14 at 15:13 +0100, Lionel Elie Mamane wrote: >> Well, you did, but apparently we didn't catch up on it and now we are >> about to ship 5.0.4 with a significant regression compared to 5.0.3 >> :-( >> See https://bugs.documentfoundation.org/96482 >> Do you think you can take a look at it "fast" or should we revert? > No way I can look at it fast now. Though I suggest checking in master > if you revert my fix does it then work as well as it historically did ? > i.e. did the various idle+event twiddling work restore things to the > old flickering but successful move behavior ? Joel's testing says that yes. I assume there is something to backport to make the same happen on libreoffice-5-0? -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Wed, Nov 11, 2015 at 10:59:46AM +, Caolán McNamara wrote: > On Fri, 2015-10-30 at 11:31 +0100, Lionel Elie Mamane wrote: >> There is patch attached to tdf#94069, nominally a rather annoying >> Base UI bug (crash!), but which touches the VCL event loop (not >> specific to Base) in a way that I don't understand the consequences >> of > If I look at the behaviour in an older version of LibreOffice pre idle > tweaking then on moving that widget it "jumps" around a lot and is > shown rendering in the original location and then the new one and so > on. So it didn't really work right in the first place as far as I can > see. > Looking at the UnoControl PushButton etc which don't suffer from this > problem I see that when setDesignMode is called for everything else > they set the "real" widget to hidden (it gets drawn via a different > mechanism than the real Paint in this case) so when its moved around it > doesn't generate Invalidate events. > So, I think the right solution may be to make the GridControl behave > like the other UnoControl widgets in that design mode turns off > visibility. So I've done that as > a7816853bad55ada597092c16ba9a0a761e067d0 and it seems to work well. > I don't really "use" these things day to day, so testing appreciated > for this and follow up 44daaebf835bb60fb7e442e928cd30191f15af52 to > tweak how the grid control is laid out, to ensure I haven't busted > anything horribly. Well, you did, but apparently we didn't catch up on it and now we are about to ship 5.0.4 with a significant regression compared to 5.0.3 :-( See https://bugs.documentfoundation.org/96482 Do you think you can take a look at it "fast" or should we revert? -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On 12/14/2015 06:13 AM, Lionel Elie Mamane wrote: > Do you think you can take a look at it "fast" or should we revert? I don't remember why/how I was involved with this? Did I say I would test it and somehow it fell through the list of things that I have to do? What should I "look at"? Was it my commit? Is there something I can do to fix it? I'm happy to test in a couple hours but need a bit more guidance. Best, Joel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Mon, Dec 14, 2015 at 06:37:36AM -0800, Joel Madero wrote: > On 12/14/2015 06:13 AM, Lionel Elie Mamane wrote: >> Do you think you can take a look at it "fast" or should we revert? > I don't remember why/how I was involved with this? Did I say I would > test it and somehow it fell through the list of things that I have to > do? What should I "look at"? Was Caolan's commit, you said you would test "not before this weekend" (a month ago), and to please poke you again "by Sunday" but nobody poked you. (The poke fell through the list of things *I* have to do.) http://lists.freedesktop.org/archives/libreoffice/2015-November/071038.html To be fair, several comments on the bug "tested, works now" did not mention this regression... -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Mon, Dec 14, 2015 at 06:59:27PM +0100, Alexander Thurgood wrote: > Le 14/12/2015 16:56, Lionel Elie Mamane a écrit : >> Regression is there. We are about to release 5.0.4.2 as final with >> that regression. > Looked at tdf#94069 and tested on master 5.2 alpha - I get no crash when > moving the grid control around in a form in form design mode. > What is/was the regression ? The regression is tdf#96482 -- Lionel ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
Le 14/12/2015 16:56, Lionel Elie Mamane a écrit : Hi all, > Regression is there. We are about to release 5.0.4.2 as final with > that regression. > Looked at tdf#94069 and tested on master 5.2 alpha - I get no crash when moving the grid control around in a form in form design mode. What is/was the regression ? Alex ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Mon, 2015-12-14 at 15:13 +0100, Lionel Elie Mamane wrote: > Well, you did, but apparently we didn't catch up on it and now we are > about to ship 5.0.4 with a significant regression compared to 5.0.3 > :-( > > See https://bugs.documentfoundation.org/96482 > > Do you think you can take a look at it "fast" or should we revert? No way I can look at it fast now. Though I suggest checking in master if you revert my fix does it then work as well as it historically did ? i.e. did the various idle+event twiddling work restore things to the old flickering but successful move behavior ? C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: help request on proposed VCL code patch
On Fri, 2015-10-30 at 11:31 +0100, Lionel Elie Mamane wrote: > Hi, > > There is patch attached to tdf#94069, nominally a rather annoying > Base UI bug (crash!), but which touches the VCL event loop (not > specific to Base) in a way that I don't understand the consequences > of If I look at the behaviour in an older version of LibreOffice pre idle tweaking then on moving that widget it "jumps" around a lot and is shown rendering in the original location and then the new one and so on. So it didn't really work right in the first place as far as I can see. Looking at the UnoControl PushButton etc which don't suffer from this problem I see that when setDesignMode is called for everything else they set the "real" widget to hidden (it gets drawn via a different mechanism than the real Paint in this case) so when its moved around it doesn't generate Invalidate events. So, I think the right solution may be to make the GridControl behave like the other UnoControl widgets in that design mode turns off visibility. So I've done that as a7816853bad55ada597092c16ba9a0a761e067d0 and it seems to work well. I don't really "use" these things day to day, so testing appreciated for this and follow up 44daaebf835bb60fb7e442e928cd30191f15af52 to tweak how the grid control is laid out, to ensure I haven't busted anything horribly. C. ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
