Re: Math fonts & qt5
On Mon, Feb 05, 2018 at 02:25:14PM +, Pavel Sanda wrote: > Pavel Sanda wrote: > > > when I switch from qt4->qt5 some math symbols are no more visible in GUI. > > > E.g. \oiint. Can anyone reproduce? > > > > It actually seems to be related only to \oiint. Other syms from esint work. > > If I copy the gliph to new position from 13 to e.g. 39 and move the pointer > > in symbols to that position, the glyph works again. > > 13 is normally carriage return so perhaps new qt overrides the default value > > of it. > > Little bit of searching suggests that it's another incarnation of bug #8493. > Quick check via looking at stmary10.lyx shows that these codepoints > are now (qt5) broken on top of 9: 10 12 13 88 89 . > > I am not sure whether previous solution via editing ttf files to copy > the codepoint elsewhere is viable strategy when qt devs decide to > ban arbitrary points whenever they are pleased to. Can we at least have a hardcoded list of symbols that are not shown correctly? e.g. can we show "\oiint" as LaTeX in LyX instead of using the broken glyph? I don't know how to do that. In lib/symbols is it possible to still load the esint package but not use the glyph? Scott signature.asc Description: PGP signature
Re: Blurry inserted figures for retina display in Lyx 2.2 Mac
On Mon, Feb 05, 2018 at 08:40:09AM +, Jean-Marc Lasgouttes wrote: > Le 05/02/2018 à 09:22, Stephan Witt a écrit : > > Am 05.02.2018 um 03:07 schrieb Scott Kostyshak: > > > > > > On Sat, Feb 03, 2018 at 10:22:42PM +, Stephan Witt wrote: > > > > > > > I cannot reproduce it either. Let’s see how it goes with the pixmap > > > > cache disabled. > > > > > > Joel and Stephan, can you reproduce with pixmap enabled? > > > > Yes, I can reproduce with pixmap enabled. > > > > Before starting a bisect I’d like to know if the pixmap cache is needed at > > all. > > This is the main question IMO. These days, we cache QTextLayout objects, > which already contain the typeset glyphs. I know nothing about this. If we decide to not use the pixmap cache, would we do a prefs to prefs change at this point in the release process? Or would we just ignore the pref in our internal code? Scott signature.asc Description: PGP signature
Re: Math fonts & qt5
On Mon, Feb 05, 2018 at 12:14:15PM +, Pavel Sanda wrote: > Hi, > > when I switch from qt4->qt5 some math symbols are no more visible in GUI. > E.g. \oiint. Can anyone reproduce? I can also reproduce with Qt 5. Scott signature.asc Description: PGP signature
Re: Math fonts & qt5
Pavel Sanda wrote: > > when I switch from qt4->qt5 some math symbols are no more visible in GUI. > > E.g. \oiint. Can anyone reproduce? > > It actually seems to be related only to \oiint. Other syms from esint work. > If I copy the gliph to new position from 13 to e.g. 39 and move the pointer > in symbols to that position, the glyph works again. > 13 is normally carriage return so perhaps new qt overrides the default value > of it. Little bit of searching suggests that it's another incarnation of bug #8493. Quick check via looking at stmary10.lyx shows that these codepoints are now (qt5) broken on top of 9: 10 12 13 88 89 . I am not sure whether previous solution via editing ttf files to copy the codepoint elsewhere is viable strategy when qt devs decide to ban arbitrary points whenever they are pleased to. Pavel
Re: Math fonts & qt5
Le 05/02/2018 à 13:52, Pavel Sanda a écrit : Jean-Pierre Chrétien wrote: Le 05/02/2018 ? 13:14, Pavel Sanda a écrit : Hi, when I switch from qt4->qt5 some math symbols are no more visible in GUI. E.g. \oiint. Can anyone reproduce? Yes, I have to toggle Instant preview to see it, and when I click on the drawing, it disappears again. That was with 2.3.0rc2, the pdf output is all right. So when you compile with qt4 it work again, right? The only version I have here with qt4 is 2.2.4dev, and yes the symbol is rendered ok in the gui with instant preview deactivated. -- Jean-Pierre
Re: Math fonts & qt5
Pavel Sanda wrote: > Hi, > > when I switch from qt4->qt5 some math symbols are no more visible in GUI. > E.g. \oiint. Can anyone reproduce? It actually seems to be related only to \oiint. Other syms from esint work. If I copy the gliph to new position from 13 to e.g. 39 and move the pointer in symbols to that position, the glyph works again. 13 is normally carriage return so perhaps new qt overrides the default value of it. Pavel
Re: Math fonts & qt5
Jean-Pierre Chrétien wrote: > Le 05/02/2018 ? 13:14, Pavel Sanda a écrit : >> Hi, >> when I switch from qt4->qt5 some math symbols are no more visible in GUI. >> E.g. \oiint. Can anyone reproduce? > > Yes, I have to toggle Instant preview to see it, and when I click on the > drawing, it disappears again. So when you compile with qt4 it work again, right? Pavel
Re: Math fonts & qt5
Le 05/02/2018 à 13:14, Pavel Sanda a écrit : Hi, when I switch from qt4->qt5 some math symbols are no more visible in GUI. E.g. \oiint. Can anyone reproduce? Yes, I have to toggle Instant preview to see it, and when I click on the drawing, it disappears again. Debian stretch, lyx-2.3.0rc2, Qt 5.7.1 -- Jean-Pierre
Math fonts & qt5
Hi, when I switch from qt4->qt5 some math symbols are no more visible in GUI. E.g. \oiint. Can anyone reproduce? Debian here. Pavel
Re: Blurry inserted figures for retina display in Lyx 2.2 Mac
Le 05/02/2018 à 09:22, Stephan Witt a écrit : Am 05.02.2018 um 03:07 schrieb Scott Kostyshak: On Sat, Feb 03, 2018 at 10:22:42PM +, Stephan Witt wrote: I cannot reproduce it either. Let’s see how it goes with the pixmap cache disabled. Joel and Stephan, can you reproduce with pixmap enabled? Yes, I can reproduce with pixmap enabled. Before starting a bisect I’d like to know if the pixmap cache is needed at all. This is the main question IMO. These days, we cache QTextLayout objects, which already contain the typeset glyphs. JMarc
Re: Blurry inserted figures for retina display in Lyx 2.2 Mac
Am 05.02.2018 um 03:07 schrieb Scott Kostyshak: > > On Sat, Feb 03, 2018 at 10:22:42PM +, Stephan Witt wrote: > >> I cannot reproduce it either. Let’s see how it goes with the pixmap cache >> disabled. > > Joel and Stephan, can you reproduce with pixmap enabled? Yes, I can reproduce with pixmap enabled. Before starting a bisect I’d like to know if the pixmap cache is needed at all. Stephan