Ridiculous...
Hullo all,
I regard the following scenario which has occurred to me as ridiculous.
Any help would be appreciated.
I had an old CVS version of lyx installed, so I nuked the directory,
binary, and user .lyx directories before upgrading - just to be on the
safe side. (No that the following worked before)
Then I got the latest 1.2.0 cvs snapshot, compiled, did a make, make
install-strip. I then downloaded db2lyx 0.1.4 (As 0.1.5 completely died
- although I did try it first just to make sure. It had issues finding
the DocBook XSL directories, even though the classpaths, other paths etc
were EXACTLY THE SAME as I was using a script, and renaming the
directories - e.g. mv db2lyx-015 db2lyx). I also downloaded xt.jar,
xp.jar, the catalog jars, and mml and put them in:-
/usr/share/mml
/usr/share/catalog
/usr/share/docbook-xml
/usr/share/docbook-xsl/1.49
I setup a catalog file to reflect these (attached). I made a script to
alleviate the problem of classpaths (I don't want to change the system
classpath just for one program) which calls config_xml. This file is
called doit.sh (attached) and resides in /usr/local/share/lyx. I then
edited the patched configure script (attached) to call doit.sh instead
of config_xml.
I now run doit.sh and it works fine (well, doesn't error at least).
On starting lyx it also runs doit.sh (obviously), the output of which is
attached also. Surprise surprise it doesn't work :/
I clicked reconfigure in lyx, then restarted and low and behold the
old freeze-on-splash-screen-when-the-user-has-a-.lyx-directory bug
strikes again! This SURELY can't just be happening to me!?!
The files are attached anyway. If someone could figure out what the hell
is going on it would be much appreciated. I can give more info if you
need it.
Cheers,
Adam.
--
| Adam Fowler |
| Third year undergraduate | TomcatBook Project |
| Computer Science | Author and Administrator |
| University of Wales, Aberystwyth | http://tomcatbook.sourceforge.net|
| Web: http://users.aber.ac.uk/aff9 | [EMAIL PROTECTED] |
PUBLIC -//OASIS//DTD DocBook XML V4.1.2//EN
file:///usr/share/docbook-xml/docbookx.dtd
PUBLIC -//DB2LyX//Norman Walsh's XSL Stylesheet Directory//EN
file:///usr/share/docbook-xsl/1.49
PUBLIC -//DB2LyX//MathML ISO Entities Directory//EN
file:///home/share/mml
#! /bin/sh
# This script is a hand-made configure script. It contains a lot of
# code stolen from GNU autoconf. I removed all the code that was not
# useful for configuring a LyX installation.
some configuration variables
outfile=lyxrc.defaults
rc_entries=
lyx_check_config=yes
lyx_keep_temps=no
srcdir=
lyx_suffix=
Parse the command line
for ac_option do
case $ac_option in
-help | --help | -h)
cat EOF
Usage: configure [options]
Options:
--help show this help lines
--keep-temps keep temporary files (for debug. purposes)
--without-latex-config do not run LaTeX to determine configuration
--with-lyx-suffix=suffix suffix of binary installed files
EOF
exit 0;;
--without-latex-config)
lyx_check_config=no ;;
--keep-temps)
lyx_keep_temps=yes ;;
--with-lyx-suffix*)
lyx_suffix=`echo $ac_option | sed 's,--with-lyx-suffix=,,;s,^,-,'`
esac
done
Checking for some echo oddities
if (echo testing\c; echo 1,2,3) | grep c /dev/null; then
# Stardent Vistra SVR4 grep lacks -e, says [EMAIL PROTECTED]
if (echo -n testing; echo 1,2,3) | sed s/-n/xn/ | grep xn /dev/null; then
ac_n= ac_c='
' ac_t=' '
else
ac_n=-n ac_c= ac_t=
fi
else
ac_n= ac_c='\c' ac_t=
fi
I do not really know why this is useful, but we might as well keep it.
# NLS nuisances.
# Only set these to C if already set. These must not be set unconditionally
# because not all systems understand e.g. LANG=C (notably SCO).
# Fixing LC_MESSAGES prevents Solaris sh from translating var values in `set'!
# Non-C LC_CTYPE values break the ctype check.
if test ${LANG+set} = set; then LANG=C; export LANG; fi
if test ${LC_ALL+set} = set; then LC_ALL=C; export LC_ALL; fi
if test ${LC_MESSAGES+set} = set; then LC_MESSAGES=C; export LC_MESSAGES; fi
if test ${LC_CTYPE+set}= set; then LC_CTYPE=C;export LC_CTYPE;fi
Guess the directory in which configure is located.
ac_prog=$0
srcdir=`echo $ac_prog|sed 's%/[^/][^/]*$%%'`
srcdir=`echo ${srcdir} | sed 's%\([^/]\)/*$%\1%'`
test x$srcdir = x$ac_prog srcdir=.
if test ! -r ${srcdir}/chkconfig.ltx ; then
echo configure: error: cannot find chkconfig.ltx script
exit 1
fi
Adjust PATH for Win32 (Cygwin)
if test x$OSTYPE = xcygwin; then
echo configure: cygwin detected; path correction
srcdir=`cygpath -w
Re: proposed cursor patch
On Tuesday 12 March 2002 8:51 pm, John Levon wrote: On Tue, Mar 12, 2002 at 08:33:07PM +0100, Lars Gullik Bjønnes wrote: | Any objections if I apply it? (and strip the associated X11 stuff from | BufferView::pimpl). I object. This is guii work and we shoulc not do this now. Besides, I am some of the way through a major upheaval, so any unnecessary patches in this area make my life harder Fine. I'll just put in the stuff in the xforms frontend which fixes (hides) the Bad Window message and which sets the color of all four cursors that we use, not just two of them. That _is_ a bug fix, albeit a minor one. I'll also enquire of the xforms list why FL_ObjWin(o) can return 0. Angus
Re: getFontSettings optimization
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars I think this is a better algorithm for font lookup. (at least Lars gprof tells me so) This is probably true, since there are very few font changes in a typical paragraph. What are the getFontSettings numbers down to? I've been wanting to work on getFontSetting after 1.2.0, so I would like to see whether this very simple approach is enough. JMarc
Re: math-macro-arg does not work
Andre == Andre Poenitz [EMAIL PROTECTED] writes:
Andre On Tue, Mar 12, 2002 at 06:15:31PM +0100, Jean-Marc Lasgouttes
Andre wrote: I guess this has not found much testing lately since the
Andre same effect is obtained by 'self-insert \#1' or directly typing
Andre '#1' in the formula.
I tried that, and I get in my file:
\begin_inset FormulaMacro \newcommand{\ind}{\mathbb{1}_{\{#1\}}}
Andre Hm... of course the macro has to be created with the right
Andre number of args from the beginning (M-x math-macro ind 1) ...
Andre but LyX should not crash. This is a bug.
Ahh, you did not tell me that I could give the number of args second
argument as second argument... I see now. LyX should even not let me
enter #1 if the macro has no arguments.
JMarc
Re: guii patch
On Wednesday 13 March 2002 12:38 am, John Levon wrote: Here : http://www.movement.uklinux.net/q1.diff.gz File not found A
Re: math-macro-arg does not work
On Wed, Mar 13, 2002 at 10:19:57AM +0100, Jean-Marc Lasgouttes wrote: Andre Hm... of course the macro has to be created with the right Andre number of args from the beginning (M-x math-macro ind 1) ... Andre but LyX should not crash. This is a bug. Ahh, you did not tell me that I could give the number of args second argument as second argument... I see now. LyX should even not let me enter #1 if the macro has no arguments. I am currently trying to fix that... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: lyxtextclass optimization
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars With the patch below: Lars Flat profile: Excellent. JMarc
Re: cut paste from e-mail
Vitaly == Vitaly Lipatov [EMAIL PROTECTED] writes: Vitaly I do a similar converter for my own needs. I thought I can Vitaly found such converter anywhere, but I can't. Did you hear about Vitaly similar projects? I think I did, but I can't remember now. JMarc
Re: makefile patch
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars This is a think I have been planning to do with the makefiles Lars for a long time. I am sure there are details that can be Lars imporved, but this should make sure that the Makefile.am's are a Lars lot more uniform. Lars Please comment. I'm not sure I like the contents of your include file. +AUTOMAKE_OPTIONS = foreign 1.5 + So you want to force automake 1.5? Can't this wait for 1.3.0cvs? +DISTCLEANFILES= -r *.orig *.rej *~ *.bak lyx.1 core libintl.h config.h config.status +config.cache config.log \ + $(srcdir)/sourcedoc/*~ \ + $(srcdir)/sourcedoc/html $(srcdir)/sourcedoc/latex \ + $(srcdir)/sourcedoc/man $(srcdir)/sourcedoc/Doxyfile Most of the files there can only exist in _some_ directory (toplevel or src). It does not make sense to mention them in a file included everywhere. I don't like that. +MAINTAINERCLEANFILES = $(srcdir)/Makefile.in $(srcdir)/config.h.in \ + $(srcdir)/aclocal.m4 \ + $(srcdir)/configure $(srcdir)/development/lyx.spec \ + $(srcdir)/acinclude.m4 $(srcdir)/sourcedoc/Doxyfile Here also, only $(srcdir)/Makefile.in is generic and should be here. JMarc
Re: getFontSettings optimization
On Wed, Mar 13, 2002 at 11:00:30AM +0100, Jean-Marc Lasgouttes wrote: Andre Concerning the excessive memory consumption: I believe that one Andre could reduce it significantly by converting parts 'far from the Andre cursor' (i.e not the same word or the same line or paragraph) Andre to some more compact form where chars of the same font are Andre collected in some kind of 'string inset'... Are you speaking about the normal paragraph structure? I do not see what could be more compact than the FontTable approach. 'Excessive memory consumption when using mathed's approach without any change'. The current FontTable approach is indeed pretty compact, but I have the suspicion that it is more code involved than necessary for the task as such... Andre' -- André Pönitz .. [EMAIL PROTECTED]
read access to Painter::owner
Can I have a read accessor to painter::owner (or something else that allows me to retrieve the size of the 'interesting' area on screen)? Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: lfuncore.diff
John == John Levon [EMAIL PROTECTED] writes: John can someone apply lfun_core.diff please ? Done in my local tree. JMarc
Re: getFontSettings optimization
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre On Wed, Mar 13, 2002 at 11:00:30AM +0100, Jean-Marc Lasgouttes Andre wrote: Concerning the excessive memory consumption: I believe Andre that one could reduce it significantly by converting parts 'far Andre from the cursor' (i.e not the same word or the same line or Andre paragraph) to some more compact form where chars of the same Andre font are collected in some kind of 'string inset'... Are you speaking about the normal paragraph structure? I do not see what could be more compact than the FontTable approach. Andre 'Excessive memory consumption when using mathed's approach Andre without any change'. The current FontTable approach is indeed Andre pretty compact, but I have the suspicion that it is more code Andre involved than necessary for the task as such... In fact, a cursor (or rather an iterator over a paragraph) could maintain an index in the fonttable for the current cur position. JMarc
RE: read access to Painter::owner
On 13-Mar-2002 Andre Poenitz wrote: Can I have a read accessor to painter::owner (or something else that allows me to retrieve the size of the 'interesting' area on screen)? Well the painters owner should be the BufferView and you have access to it. In which function do you need this? Jug -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Documentation: Instructions translated from Swedish by Japanese for English speaking persons.
Re: read access to Painter::owner
On Wed, Mar 13, 2002 at 11:26:49AM +0100, Juergen Vigna wrote: Well the painters owner should be the BufferView and you have access to it. In which function do you need this? MathXArray::draw(). I just want to skip any work for cell outside the visible workarea. Andre' -- André Pönitz .. [EMAIL PROTECTED]
figinset cruft
This is cruft from the figinset. It should go, go , go!
Angus
Index: src/WorkArea.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/WorkArea.C,v
retrieving revision 1.62
diff -u -p -r1.62 WorkArea.C
--- src/WorkArea.C 28 Feb 2002 14:06:17 - 1.62
+++ src/WorkArea.C 13 Mar 2002 10:42:47 -
-39,8 +39,6 using std::endl;
using std::abs;
using std::hex;
-FL_OBJECT * figinset_canvas;
-
namespace {
inline
-84,8 +82,6 WorkArea::WorkArea(int xpos, int ypos, i
{
fl_freeze_all_forms();
- figinset_canvas = 0;
-
if (lyxerr.debugging(Debug::WORKAREA))
lyxerr Creating work area: +
xpos '+' ypos ' '
-93,14 +89,6 WorkArea::WorkArea(int xpos, int ypos, i
//
FL_OBJECT * obj;
int const bw = int(abs(fl_get_border_width()));
-
- // We really want to get rid of figinset_canvas.
- ::figinset_canvas = figinset_canvas = obj =
- fl_add_canvas(FL_NORMAL_CANVAS,
- xpos + 1, ypos + 1, 1, 1, );
- fl_set_object_boxtype(obj, FL_NO_BOX);
- fl_set_object_resize(obj, FL_RESIZE_ALL);
- fl_set_object_gravity(obj, NorthWestGravity, NorthWestGravity);
// a box
if (lyxerr.debugging(Debug::WORKAREA))
Index: src/WorkArea.h
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/WorkArea.h,v
retrieving revision 1.29
diff -u -p -r1.29 WorkArea.h
--- src/WorkArea.h 7 Feb 2002 16:43:53 - 1.29
+++ src/WorkArea.h 13 Mar 2002 10:42:47 -
-151,8 +151,6 private:
Pixmap workareapixmap;
///
Painter painter_;
- ///
- FL_OBJECT * figinset_canvas;
/// if we call redraw with true needed for locking-insets
bool screen_cleared;
};
Re: read access to Painter::owner
On 13-Mar-2002 Andre Poenitz wrote: MathXArray::draw(). I just want to skip any work for cell outside the visible workarea. Have a look at InsetTabular::draw I'll do that there too. Jug -- -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ Dr. Jürgen VignaE-Mail: [EMAIL PROTECTED] Italienallee 13/N Tel/Fax: +39-0471-450260 / +39-0471-450253 I-39100 Bozen Web: http://www.sad.it/~jug -._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._-._ The Hollywood tradition I like best is called sucking up to the stars. -- Johnny Carson
Re: getFontSettings optimization
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars Each sample counts as 0.01 seconds. [...] getFontSettings is down to .23ms/call from 0.89. This is good. Lars | I've been wanting to work on getFontSetting after 1.2.0, so I Lars would | like to see whether this very simple approach is enough. Lars I have done the same change to other functions that used matchFT Lars and lowerbound. Looks like you create common code that should be factored in some helper function. JMarc
Re: read access to Painter::owner
On Wed, Mar 13, 2002 at 11:52:55AM +0100, Juergen Vigna wrote: MathXArray::draw(). I just want to skip any work for cell outside the visible workarea. Have a look at InsetTabular::draw I'll do that there too. Ah... PainterBase::paperHeight() and ...Width() seem to do the trick. Andre' -- André Pönitz .. [EMAIL PROTECTED]
small stuff
* whitespace changes
* member var with trailing underscore
* one-line comments changed to ///
Ok to apply (together with ChangeLog entry)?
Andre'
--
André Pönitz .. [EMAIL PROTECTED]
Index: Painter.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/Painter.C,v
retrieving revision 1.32
diff -u -p -r1.32 Painter.C
--- Painter.C 27 Feb 2002 09:59:48 - 1.32
+++ Painter.C 13 Mar 2002 11:27:41 -
@@ -48,7 +48,7 @@ Painter::Painter(WorkArea wa)
PainterBase Painter::point(int x, int y, LColor::color c)
{
- XDrawPoint(display, owner.getPixmap(),
+ XDrawPoint(display, owner_.getPixmap(),
lyxColorHandler-getGCForeground(c), x, y);
return *this;
}
@@ -59,7 +59,7 @@ PainterBase Painter::line(int x1, int
enum line_style ls,
enum line_width lw)
{
- XDrawLine(display, owner.getPixmap(),
+ XDrawLine(display, owner_.getPixmap(),
lyxColorHandler-getGCLinepars(ls, lw, col),
x1, y1, x2, y2);
return *this;
@@ -78,7 +78,7 @@ PainterBase Painter::lines(int const *
points[i].y = yp[i];
}
-XDrawLines(display, owner.getPixmap(),
+XDrawLines(display, owner_.getPixmap(),
lyxColorHandler-getGCLinepars(ls, lw, col),
points.get(), np, CoordModeOrigin);
@@ -91,7 +91,7 @@ PainterBase Painter::rectangle(int x,
enum line_style ls,
enum line_width lw)
{
- XDrawRectangle(display, owner.getPixmap(),
+ XDrawRectangle(display, owner_.getPixmap(),
lyxColorHandler-getGCLinepars(ls, lw, col),
x, y, w, h);
return *this;
@@ -101,7 +101,7 @@ PainterBase Painter::rectangle(int x,
PainterBase Painter::fillRectangle(int x, int y, int w, int h,
LColor::color col)
{
- XFillRectangle(display, owner.getPixmap(),
+ XFillRectangle(display, owner_.getPixmap(),
lyxColorHandler-getGCForeground(col), x, y, w, h);
return *this;
}
@@ -117,7 +117,7 @@ PainterBase Painter::fillPolygon(int c
points[i].y = yp[i];
}
- XFillPolygon(display, owner.getPixmap(),
+ XFillPolygon(display, owner_.getPixmap(),
lyxColorHandler-getGCForeground(col), points.get(), np,
Nonconvex, CoordModeOrigin);
@@ -129,7 +129,7 @@ PainterBase Painter::arc(int x, int y,
unsigned int w, unsigned int h,
int a1, int a2, LColor::color col)
{
-XDrawArc(display, owner.getPixmap(),
+XDrawArc(display, owner_.getPixmap(),
lyxColorHandler-getGCForeground(col),
x, y, w, h, a1, a2);
return *this;
@@ -150,7 +150,7 @@ PainterBase Painter::segments(int cons
s[i].x2 = x2[i];
s[i].y2 = y2[i];
}
- XDrawSegments(display, owner.getPixmap(),
+ XDrawSegments(display, owner_.getPixmap(),
lyxColorHandler-getGCLinepars(ls, lw, col),
s.get(), ns);
@@ -162,9 +162,9 @@ PainterBase Painter::pixmap(int x, int
{
XGCValues val;
val.function = GXcopy;
- GC gc = XCreateGC(display, owner.getPixmap(),
+ GC gc = XCreateGC(display, owner_.getPixmap(),
GCFunction, val);
- XCopyArea(display, bitmap, owner.getPixmap(), gc,
+ XCopyArea(display, bitmap, owner_.getPixmap(), gc,
0, 0, w, h, x, y);
XFreeGC(display, gc);
return *this;
@@ -217,7 +217,7 @@ PainterBase Painter::text(int x, int y
GC gc = lyxColorHandler-getGCForeground(f.realColor());
if (f.realShape() != LyXFont::SMALLCAPS_SHAPE) {
lyxfont::XSetFont(display, gc, f);
- XDrawString(display, owner.getPixmap(), gc, x, y, s, ls);
+ XDrawString(display, owner_.getPixmap(), gc, x, y, s, ls);
} else {
LyXFont smallfont(f);
smallfont.decSize().decSize().setShape(LyXFont::UP_SHAPE);
@@ -226,12 +226,12 @@ PainterBase Painter::text(int x, int y
char const c = uppercase(s[i]);
if (c != s[i]) {
lyxfont::XSetFont(display, gc, smallfont);
- XDrawString(display, owner.getPixmap(), gc,
+ XDrawString(display, owner_.getPixmap(), gc,
tmpx, y, c, 1);
tmpx += lyxfont::XTextWidth(smallfont, c, 1);
} else {
lyxfont::XSetFont(display, gc, f);
-
proposed cursor patch try 2
Gives us a well behaved and configurable cursor in the main window as well as
in the dialogs (fixes bug #73).
Angus
Index: src/BufferView_pimpl.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/BufferView_pimpl.C,v
retrieving revision 1.212
diff -u -p -r1.212 BufferView_pimpl.C
--- src/BufferView_pimpl.C 12 Mar 2002 13:27:49 - 1.212
+++ src/BufferView_pimpl.C 13 Mar 2002 11:27:11 -
-107,20 +107,6 void waitForX()
XSync(fl_get_display(), 0);
}
-
-void SetXtermCursor(Window win)
-{
- static Cursor cursor;
- static bool cursor_undefined = true;
- if (cursor_undefined) {
- cursor = XCreateFontCursor(fl_get_display(), XC_xterm);
- XFlush(fl_get_display());
- cursor_undefined = false;
- }
- XDefineCursor(fl_get_display(), win, cursor);
- XFlush(fl_get_display());
-}
-
} // anon namespace
-773,7 +759,7 void BufferView::Pimpl::selectionLost()
void BufferView::Pimpl::enterView()
{
if (active() available()) {
- SetXtermCursor(workarea_.getWin());
+ fl_set_cursor(workarea_.getWin(), XC_xterm);
using_xterm_cursor = true;
}
}
-782,7 +768,9 void BufferView::Pimpl::enterView()
void BufferView::Pimpl::leaveView()
{
if (using_xterm_cursor) {
- XUndefineCursor(fl_get_display(), workarea_.getWin());
+ Cursor const cursor = Dialogs::tooltipsEnabled() ?
+XC_question_arrow : FL_DEFAULT_CURSOR;
+fl_set_cursor(workarea_.getWin(), cursor);
using_xterm_cursor = false;
}
}
Re: small stuff
On Wednesday 13 March 2002 11:35 am, Lars Gullik Bjønnes wrote: Andre Poenitz [EMAIL PROTECTED] writes: | * whitespace changes | * member var with trailing underscore I try to stay with the rule that only private class variables get the trailing underscore. How hard is it to make owner_ private? If you're modifying Painter, then perhaps you could merge image() and pixmap() as we no longer have a figinset. Angus
Re: small stuff
On Wed, Mar 13, 2002 at 12:35:06PM +0100, Lars Gullik Bjønnes wrote: How hard is it to make owner_ private? It is only use a couple of times to call a single function: owner_.getPixmap() So we could have a private 'owner_' and a protected 'getPixmap' or a protected 'owner() const' Your call. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: small stuff
On Wed, Mar 13, 2002 at 11:41:59AM +, Angus Leeming wrote: How hard is it to make owner_ private? If you're modifying Painter, then perhaps you could merge image() and pixmap() as we no longer have a figinset. Uh.. it was just trivial stuff that caught my eye when hunting for 'paperHeight()'. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: small stuff
On Wednesday 13 March 2002 11:45 am, Andre Poenitz wrote: On Wed, Mar 13, 2002 at 11:41:59AM +, Angus Leeming wrote: How hard is it to make owner_ private? If you're modifying Painter, then perhaps you could merge image() and pixmap() as we no longer have a figinset. Uh.. it was just trivial stuff that caught my eye when hunting for 'paperHeight()'. Well this is just trivial cruft too... A
Re: small stuff
On Wednesday 13 March 2002 11:43 am, Andre Poenitz wrote: On Wed, Mar 13, 2002 at 12:35:06PM +0100, Lars Gullik Bjønnes wrote: How hard is it to make owner_ private? It is only use a couple of times to call a single function: owner_.getPixmap() So we could have a private 'owner_' and a protected 'getPixmap' Removes Painter.C's dependency on WorkArea... or a protected 'owner() const' Your call. Andre'
Re: small stuff
On Wednesday 13 March 2002 11:50 am, Lars Gullik Bjønnes wrote: Andre Poenitz [EMAIL PROTECTED] writes: | On Wed, Mar 13, 2002 at 12:35:06PM +0100, Lars Gullik Bjønnes wrote: How hard is it to make owner_ private? | It is only use a couple of times to call a single function: |owner_.getPixmap() | So we could have |a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. Incidentally, these are historical cruft too by the look of things... aleem@pneumon:src- grep -n dummy Painter*.[Ch] PainterBase.C:21:int PainterBase::dummy1 = 0; PainterBase.C:22:int PainterBase::dummy2 = 0; PainterBase.C:23:int PainterBase::dummy3 = 0; PainterBase.h:41: static int dummy1; PainterBase.h:43: static int dummy2; PainterBase.h:45: static int dummy3; Angus
Re: small stuff
On Wed, Mar 13, 2002 at 12:50:25PM +0100, Lars Gullik Bjønnes wrote: | So we could have |a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. Uh... I just see we need Pixmap defined... We do not to #include X11/X.h for that, do we? | or |a protected 'owner() const' The better way than? Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: small stuff
On Wed, Mar 13, 2002 at 11:59:32AM +, Angus Leeming wrote: Incidentally, these are historical cruft too by the look of things... aleem@pneumon:src- grep -n dummy Painter*.[Ch] PainterBase.C:21:int PainterBase::dummy1 = 0; PainterBase.C:22:int PainterBase::dummy2 = 0; PainterBase.C:23:int PainterBase::dummy3 = 0; PainterBase.h:41: static int dummy1; PainterBase.h:43: static int dummy2; PainterBase.h:45: static int dummy3; Ok. I think just removing these and doing the whitespace changes without the variable renaming is uncontroversial and can be done right now? Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: small stuff
On Wednesday 13 March 2002 11:58 am, Andre Poenitz wrote: On Wed, Mar 13, 2002 at 12:50:25PM +0100, Lars Gullik Bjønnes wrote: | So we could have |a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. Uh... I just see we need Pixmap defined... We do not to #include X11/X.h for that, do we? it's just a long int isn't it? Here at least. Angus
Re: small stuff
Angus == Angus Leeming [EMAIL PROTECTED] writes: Angus On Wednesday 13 March 2002 11:50 am, Lars Gullik Bjønnes wrote: Andre Poenitz [EMAIL PROTECTED] writes: | On Wed, Mar 13, 2002 at 12:35:06PM +0100, Lars Gullik Bjønnes wrote: How hard is it to make owner_ private? | It is only use a couple of times to call a single function: | owner_.getPixmap() | So we could have | a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. Angus Incidentally, these are historical cruft too by the look of Angus things... Indeed, it looks like it is. JMArc
Re: figinset cruft
Angus == Angus Leeming [EMAIL PROTECTED] writes: Angus This is cruft from the figinset. It should go, go , go! Angus Go ahead. JMarc
Re: makefile patch
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars I just saw that we used foreign 1.4 in a couple of files, I'll Lars be happy with just foreign OK. Lars It makes things easier, and although we get some clutter and Lars unwanted stuff in include.am, we get a much nicer Makefile.am. I think it is better to add the clutter in the Makefiles.am that need it. Isn't it possible to do VAR += foo bar in Makefiles? JMarc
Re: small stuff
On Wed, Mar 13, 2002 at 01:05:13PM +0100, Lars Gullik Bjønnes wrote: I am stil for the getPixmap solution, but leave the includes as is. Then it should work right? No, if we have Pixmap getPixmap() we need a full definition for Pixmap (which is a typedef to some XID, so it can't even be sensible passed as reference). The definition sits in X11/X.h withi has 700 lines. Apart from that PainterBase should be completely ignorant of X... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: small stuff
On Wed, Mar 13, 2002 at 12:07:46PM +, Angus Leeming wrote: Uh... I just see we need Pixmap defined... We do not to #include X11/X.h for that, do we? it's just a long int isn't it? Here at least. Depends on architecture after reading X11/X.h. I would not simply use a 'long'... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: configure question
Angus == Angus Leeming [EMAIL PROTECTED] writes: Angus Having just updated my tree, JMarc's change to configure.in Angus triggered an automatic rerun of configure. In the past this has Angus failed to run correctly, because configure has insisted on Angus using g++ rather than deccxx. I have, therefore, had to run the Angus configure script as CXX=deccxx configure etc If you upgrade to autoconf 2.5x, you will be able to write configure CXX=deccxx and configure will remember this value when reconfiguring. Angus I have, however, one query. Here is a section of the configure Angus messages: Angus checking for a good enough C++ compiler... g++ Angus checking for c++... (cached) deccxx Angus checking whether the C++ compiler (deccxx ) works... yes Angus why does it check for g++ at all in this case? Could you try the following patch? JMarc pa-conf Description: Binary data
Re: Painter diff
On Wednesday 13 March 2002 12:25 pm, Lars Gullik Bjønnes wrote:
Index: Painter.h
===
private:
+ /// For the figinset
+ PainterBase pixmap(int x, int y, int w, int h, Pixmap bitmap);
+
Errr, why? Helper functions are fine if they're used by more than one
function but this is just daft. You've already proved to yourself that pixmap
is used only with Painter.C, so now's the time to dump it.
Angus
PainterBase Painter::pixmap(int x, int y, int w, int h, Pixmap bitmap)
{
XGCValues val;
val.function = GXcopy;
GC gc = XCreateGC(display, owner.getPixmap(),
GCFunction, val);
XCopyArea(display, bitmap, owner.getPixmap(), gc,
0, 0, w, h, x, y);
XFreeGC(display, gc);
return *this;
}
PainterBase Painter::image(int x, int y, int w, int h,
grfx::GImage const image)
{
return pixmap(x, y, w, h, image.getPixmap());
}
Re: math-macro-arg does not work
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre On Tue, Mar 12, 2002 at 02:39:11PM +0100, Jean-Marc Lasgouttes Andre wrote: I get a message `a truly unknown func' with math-macro-arg, whereas math-macro works well. Andre I guess this has not found much testing lately since the same Andre effect is obtained by 'self-insert \#1' or directly typing '#1' Andre in the formula. Andre Should we just remove the LFUN? OK, I think you should just remove the lfun. BTW, there are some metrics problems with macros. Example attached. JMarc newfile1.lyx Description: Binary data
Re: configure question
On Wednesday 13 March 2002 12:28 pm, Jean-Marc Lasgouttes wrote:
Angus Having just updated my tree, JMarc's change to configure.in
Angus triggered an automatic rerun of configure. In the past this has
Angus failed to run correctly, because configure has insisted on
Angus using g++ rather than deccxx. I have, therefore, had to run the
Angus configure script as CXX=deccxx configure etc
If you upgrade to autoconf 2.5x, you will be able to write
configure CXX=deccxx
and configure will remember this value when reconfiguring.
Angus I have, however, one query. Here is a section of the configure
Angus messages:
Angus checking for a good enough C++ compiler... g++
Angus checking for c++... (cached) deccxx
Angus checking whether the C++ compiler (deccxx ) works... yes
Angus why does it check for g++ at all in this case?
Could you try the following patch?
JMarc
Thanks for the info, JMarc. I will try it. Incidentally, I found that, in
order to get the configure test to run properly:
LYX_USE_XFORMS_IMAGE_LOADER
I had to define
CPPFLAGS=-I/usr/local/include
rather than
configure --extra-includes=...
My configure script is currently:
#!/bin/sh
HOMEDIR=/usr/users/aleem/OTHERS_CODE/lyx/devel
export HOMEDIR
echo HOMEDIR is ${HOMEDIR}
CPP=cpp
export CPP
echo CPP is ${CPP}
CPPFLAGS=-I/usr/local/include
export CPPFLAGS
echo CPPFLAGS is ${CPPFLAGS}
CC=deccc
export CC
echo CC is ${CC}
CFLAGS=-O2
export CFLAGS
echo CFLAGS is ${CFLAGS}
CXX=deccxx
export CXX
echo CXX is ${CXX}
CXXFLAGS=-msg_display_number -msg_disable 11,193,236,261,401,611 -w0 -ptr
${HOMEDIR}/lyx_cxx_repository -O2
export CXXFLAGS
echo CXXFLAGS is ${CXXFLAGS}
CONFIGURE=./configure --prefix=/usr/local --with-version-suffix
--with-included-gettext -with-frontend=xforms
${CONFIGURE}
Re: guii patch
On Wed, Mar 13, 2002 at 12:38:17AM +, John Levon wrote: http://www.movement.uklinux.net/q1.diff.gz q1.diff.bz2 john
Re: Painter diff
On Wednesday 13 March 2002 12:55 pm, Lars Gullik Bjønnes wrote: [EMAIL PROTECTED] (Lars Gullik Bjønnes) writes: | Angus Leeming [EMAIL PROTECTED] writes: | | On Wednesday 13 March 2002 12:25 pm, Lars Gullik Bjønnes wrote: Index: Painter.h === private: + /// For the figinset + PainterBase pixmap(int x, int y, int w, int h, Pixmap bitmap); + | | Errr, why? Helper functions are fine if they're used by more than one | | function. | In general I strongly disagree with that statement. Well of course I do too. In general. I chose my words poorly. | | but this is just daft. You've already proved to yourself that pixmap | | is used only with Painter.C, so now's the time to dump it. so you are more happy with this patch then: Yes. In this case I think that the helper function isn't a help. Angus
Re: [Devel] Bug list update again
On Wednesday 13 March 2002 12:40 pm, Jean-Marc Lasgouttes wrote: Herbert == Herbert Voss [EMAIL PROTECTED] writes: - Citation labels should reflect their content, i.e. if only the year is chosen in the citation dialog, then the label should show the year afterwards. (#114) Herbert attached a patch which fixes a bug. Herbert - now the choice_style also changes when changing the bibkey And now multiple citations are not taken in account anymore, if I read the patch correctly. Oops... I hadn't noticed this. Sorry, Herbert! I'm afraid I agree with Jean-Marc. + // we save the combination after|label string const after = fl_get_input(dialog_-input_after); - controller().params().setOptions(after); + controller().params().setOptions(after+ + '|'+strip(frontStrip( + fl_get_choice_text(dialog_-choice_style; This is a fudge. I had something like this very temporarily and then ditched it. The proper fix is to have insetcitation support multiple options. Anyway, this shouldn't go in FormCitation but in ControlCitation. The insetcitation stuff is also nasty. I think that the real thing to do is to create a BibTeX cache, analogous to the graphics cache and to do such BibTeX stuff there. I plan to use BibTeX itself to parse the .bib files in 1.3.0. We get such stuff for free then. Try the attached demo code (which I'm sure I've sent to you in the past!) Angus trial2.C.bz2 Description: BZip2 compressed data
q1.diff.bz2
Hi, I read the patch quickly, and it looks fine to me. Of course, a ChangeLog would be nice, because it is practically impossible to see what changes you have done to files that were moved. Greets, Asger
Re: [Devel] Bug list update again
On Wed, 13 Mar 2002, Angus Leeming wrote: On Wednesday 13 March 2002 12:40 pm, Jean-Marc Lasgouttes wrote: Herbert == Herbert Voss [EMAIL PROTECTED] writes: - Citation labels should reflect their content, i.e. if only the year is chosen in the citation dialog, then the label should show the year afterwards. (#114) Herbert attached a patch which fixes a bug. Herbert - now the choice_style also changes when changing the bibkey And now multiple citations are not taken in account anymore, if I read the patch correctly. Oops... I hadn't noticed this. Sorry, Herbert! I'm afraid I agree with Jean-Marc. I can not follow. Why are multiple citations no more possible?? Herbert This is a fudge. I had something like this very temporarily and then ditched it. The proper fix is to have insetcitation support multiple options. Anyway, this shouldn't go in FormCitation but in ControlCitation. this is another question ... Try the attached demo code (which I'm sure I've sent to you in the past!) again: the patch makes some people happy who like more wysiwyg. And the argument of JMarc that a cryptic key is more useful than the real output, is not an argument! that is all! When you rewrite the citation stuff, you can delete what you want, what has this to do with this patch? Herbert
Re: guii patch
On Wednesday 13 March 2002 12:55 pm, John Levon wrote: On Wed, Mar 13, 2002 at 12:38:17AM +, John Levon wrote: http://www.movement.uklinux.net/q1.diff.gz q1.diff.bz2 Thank you. One tiny thing on the #include front: you don't need WorkArea.h in BufferView_pimpl.h anymore. As Asger says, it's hard to see what is changed and what is just moved into frontends. Am I right in saying that you've changed little here other than move files into the frontends? The odd method rather than member variable. An #included the correct Painter.h. One final thing: why not create a frontends/X11 dir for Painter etc that don't depend explicitly on xforms and have a USE_RAW_X11 variable in the frontends Makefile.am. 1.3 is going to be fun! Angus
Re: [Devel] Bug list update again
On Wednesday 13 March 2002 1:19 pm, Herbert Voss wrote: again: the patch makes some people happy who like more wysiwyg. And the argument of JMarc that a cryptic key is more useful than the real output, is not an argument! that is all! When you rewrite the citation stuff, you can delete what you want, what has this to do with this patch? It's a nasty fudge. Ask Lars what he thinks of nasty fudges. Angus
Re: makefile patch
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars Ok, does this patch taste better? Definitely. We have to be sure though that it works for `reasonable' non-GNU makes. But we'll find out soon enough... A small nit: I do not like the name include.am. I'd rather have something like Make.rules or Makefile.common, or whatever. Of course, .am is used by automake for makefile snippets, so maybe you are right. JMarc
Re: [Devel] Bug list update again
Herbert == Herbert Voss [EMAIL PROTECTED] writes: Herbert again: the patch makes some people happy who like more Herbert wysiwyg. And the argument of JMarc that a cryptic key is more Herbert useful than the real output, is not an argument! that is all! Hmm, this is the kind of argument that is difficult to discuss :) To express things differently, I am not sure that showing everything wysiwyg is a panacea. For example, I would hate to see equation numbers in mathed showed as just numbers instead of labels. But may is it just because I am used to give meaningful names to my labels. And for citations, seeing [1] in the lyx window instead of [Key] is certainly not a plus, forgetting for now that I do not get anything. JMarc
Re: makefile patch
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars | Lars Ok, does this patch taste better? Lars | Definitely. We have to be sure though that it works for Lars `reasonable' | non-GNU makes. But we'll find out soon enough... Lars | A small nit: I do not like the name include.am. I'd rather Lars have | something like Make.rules or Makefile.common, or Lars whatever. Lars | Of course, .am is used by automake for makefile snippets, so Lars maybe you | are right. Lars would common.am be better? Yes. JMarc
Re: [Devel] Bug list update again
Jean-Marc Lasgouttes wrote: To express things differently, I am not sure that showing everything wysiwyg is a panacea. For example, I would hate to see equation numbers in mathed showed as just numbers instead of labels. But may is it just because I am used to give meaningful names to my labels. we are talking about citations and natbib. And for citations, seeing [1] in the lyx window instead of [Key] is certainly not a plus, forgetting for now that I do not get anything. did I ever said, that you'll see [1] as label?? I suppose that you don't tried the patch but give arguments against ... we are talking about natbib and not the plain version of an bst-file. Herbert -- http://www.lyx.org/help/
Re: guii patch
On Wed, Mar 13, 2002 at 01:34:49PM +, Angus Leeming wrote: One tiny thing on the #include front: you don't need WorkArea.h in BufferView_pimpl.h anymore. indeed ! As Asger says, it's hard to see what is changed and what is just moved into frontends. Am I right in saying that you've changed little here other than move files into the frontends? The odd method rather than member variable. An #included the correct Painter.h. yes everything is pretty much the same except WorkArea, screen got virtualised. And of course PainterBase became Painter :) I can do diffs against individual files if you like... One final thing: why not create a frontends/X11 dir for Painter etc that don't depend explicitly on xforms and have a USE_RAW_X11 variable in the frontends Makefile.am. I'm not sure of the point of this - are you planning on writing a pure Xlib frontend ? 1.3 is going to be fun! I suppose my next step is a proof of concept qt2... john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Painter diff
On Wed, Mar 13, 2002 at 01:25:14PM +0100, Lars Gullik Bjønnes wrote: These are the changes I have in my tree now: Index: PainterBase.C I can't help but think that you committed this thoroughly cosmetic patch just to make my life harder. john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: guii patch
On Wednesday 13 March 2002 2:37 pm, John Levon wrote: I'm not sure of the point of this - are you planning on writing a pure Xlib frontend ? Not me. But that't not the point. xforms in xforms, qt2 in qt2 and X11 in X11. It costs us nothing and leads to improved clarity. 1.3 is going to be fun! I suppose my next step is a proof of concept qt2... That will indeed be fun. Have you made any more progress in understanding KLyX or was it too a hack? Angus
Re: Painter diff
On Wednesday 13 March 2002 2:42 pm, John Levon wrote: On Wed, Mar 13, 2002 at 01:25:14PM +0100, Lars Gullik Bjønnes wrote: These are the changes I have in my tree now: Index: PainterBase.C I can't help but think that you committed this thoroughly cosmetic patch just to make my life harder. john Along the same lines, can you try my BufferView_pimpl.C patch in your tree proposed cursor patch try 2. Presumably those fl_set_cursor calls should become calls to a new WorkArea method? Thoughts? Angus
Compilation error
Once again, my compiler (egcs 1.1.2) refuses to compile lyx: g++ -DHAVE_CONFIG_H -I. -I. -I../../../src -I../../../images -I../../../src/ -I../../../src/frontends/ -I../../../src/frontends/controllers -I../../.. -I../../.. -I../../../boost -isystem /usr/X11R6/include -O -fno-rtti -fno-exceptions -W -Wall -Wp,-MD,.deps/FormPreferences.pp -c FormPreferences.C FormBaseDeprecated.h: In method void FormPreferences::Colors::build()': FormBaseDeprecated.h:113: static void FormBaseDeprecated::setPrehandler(struct FL_OBJECT *)' is protected FormPreferences.C:492: within this context FormBaseDeprecated.h:113: static void FormBaseDeprecated::setPrehandler(struct FL_OBJECT *)' is protected etc.
Re: Painter diff
On Wed, Mar 13, 2002 at 03:47:33PM +0100, Lars Gullik Bjønnes wrote: Yup. OK -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Painter diff
On Wed, Mar 13, 2002 at 02:50:41PM +, Angus Leeming wrote: Along the same lines, can you try my BufferView_pimpl.C patch in your tree proposed cursor patch try 2. Presumably those fl_set_cursor calls should become calls to a new WorkArea method? I suspect so, but I'm not sure what yet. For now that stuff is just #if 0'd out. I don't think it's a great idea to have te cursor stuff in screen either ... But this obviously isn't the time to be able to maintain a separate tree. I don't really mind it's only a couple of hours work wasted :) john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: guii patch
On Wed, Mar 13, 2002 at 02:44:55PM +, Angus Leeming wrote: I'm not sure of the point of this - are you planning on writing a pure Xlib frontend ? Not me. But that't not the point. xforms in xforms, qt2 in qt2 and X11 in X11. It costs us nothing and leads to improved clarity. I guess you are probably right. Doesn't xforms have drawing functions ? Why aren't we using them ? I suppose my next step is a proof of concept qt2... That will indeed be fun. Have you made any more progress in understanding KLyX or was it too a hack? I haven't looked more yet. The current situation is that BV manages the scrollbar. This is OK but prevents the use of e.g. QScrollArea. This is not a big problem as we can easily have a custom widget that includes a QScrollBar, and do the manipulation ourselves. Later on the interface can be improved further so Qt can manage the scrollbar itself... john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Painter diff
On Wednesday 13 March 2002 3:41 pm, John Levon wrote: On Wed, Mar 13, 2002 at 02:50:41PM +, Angus Leeming wrote: Along the same lines, can you try my BufferView_pimpl.C patch in your tree proposed cursor patch try 2. Presumably those fl_set_cursor calls should become calls to a new WorkArea method? I suspect so, but I'm not sure what yet. For now that stuff is just #if 0'd out. I don't think it's a great idea to have te cursor stuff in screen either ... But this obviously isn't the time to be able to maintain a separate tree. I don't really mind it's only a couple of hours work wasted :) john Sure it is. This stuff _should_ go in a branch. Then we can hack away happily. Angus
Re: Painter diff
On Wed, Mar 13, 2002 at 03:59:11PM +, Angus Leeming wrote: Sure it is. This stuff _should_ go in a branch. Then we can hack away happily. Yeah but I don't want to spend ages merging. You had enough trouble with that during MVC branch, from the sounds of it. Of course the branch will have to cope with all the queued 1.3 patches coming as well. Hopefully these are mostly core code and shouldn't hit us too badly. So perhaps we should make a branch now yes. Especially as I appear to have exhausted my supply of 1.2 bugfixes... I expect the branch will last 6 months or so before the interfaces are good enough to hit the main tree but I suppose that depends on whether you hit one of your massive coding sessions in the mean time :) regards john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Compilation error
On Wednesday 13 March 2002 3:36 pm, Dekel Tsur wrote:
Once again, my compiler (egcs 1.1.2) refuses to compile lyx:
g++ -DHAVE_CONFIG_H -I. -I. -I../../../src -I../../../images -I../../../src/
-I../../../src/frontends/ -I../../../src/frontends/controllers -I../../..
-I../../.. -I../../../boost -isystem /usr/X11R6/include -O -fno-rtti
-fno-exceptions -W -Wall -Wp,-MD,.deps/FormPreferences.pp -c
FormPreferences.C
FormBaseDeprecated.h: In method void FormPreferences::Colors::build()':
FormBaseDeprecated.h:113: static void
FormBaseDeprecated::setPrehandler(struct FL_OBJECT *)' is protected
FormPreferences.C:492: within this context
FormBaseDeprecated.h:113: static void
FormBaseDeprecated::setPrehandler(struct FL_OBJECT *)' is protected
... and it's problem is what exactly?
class FormBaseDeprecated {
protected:
static void setPrehandler(FL_OBJECT * ob);
};
class FormBaseBI : public FormBaseDeprecated {
};
class FormPreferences : public FormBaseBI {
};
Sigh... Patches welcome of course.
Angus
Re: [Devel] Bug list update again
On Wednesday 13 March 2002 4:11 pm, Herbert Voss wrote: Jean-Marc Lasgouttes wrote: Herbert == Herbert Voss [EMAIL PROTECTED] writes: Herbert why should they get an empty button?? I thought you tried Herbert it?? I tried it. I sent an example file. I can even add a screeshot if you do not believe me. you forget to add your chosen bibentry from the bibfile into the left browser (in the citation gui) ... you always get an empty label, with and without my patch because it has nothing to do with that! Herbert ;-) However, I don't think the key is empty on screen in such a situation. Doesn't it have brackets []? ... Yup. Angus
Re: Painter diff
On Wed, Mar 13, 2002 at 05:13:56PM +0100, Lars Gullik Bjønnes wrote: Especially not for stuff that is main tasks for the next development round. hmm. This would mean you would have to accept patches that are not perfect. I've tried to look for ways to do the Andre thing of small gradual changes but I can't see an easy way: there will always be some point where a big fat patch has to happen as far as I can see. But then, I'm lame... john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: guii patch
On Wed, Mar 13, 2002 at 05:07:15PM +0100, Lars Gullik Bjønnes wrote: I have said this before, and I am sure you have seen it, but you have probaly not understood the implications: We are using xforms wrong. We are not calling xforms functions as a result some events all the time, this confuses xforms internal completely. Some time ago we did all the drawing with xforms native functions (no direct X11 functions), this resulted in spurious lines, colors etc. all over the lyx window. In fact we are using /X11/ wrong. We should only ever draw in response to an expose event indeed. I suspect this is something that is going to have to get fixed for GUII (good !). So in fact then we can go back to using its drawing routines if it makes sense. thanks john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [patch] (html)URL - inset
Herbert == Herbert Voss [EMAIL PROTECTED] writes: Back to this one... Believe it or not, I would like to resolve it before 1.2.0. Herbert Jean-Marc Lasgouttes wrote: To rephrase myself: could this html check button produce code using hyperref \href instead of html.sty version? Would that work with latex2html? Answering to my own question, it seems that the macros in html.sty are mostly wrappers around hyperref's \href. I'd rather force the use of hyperref (which is somewhat standard) than force people to convert to html using latex2html. Herbert If you want to leran a lot of latex than use package Herbert hyperref. You can be sure, that something doesn't work when Herbert using a lot of other packages ... I know there are lot of problems in interactions between hyperref and other packages. I hoped that the situation was better today than a few years back. A way to avoid these incompatibility would be to use the draft option of hyperref, but latex2html does not understand hyperref anyway :( Herbert the html package supports hyperref, so it's no problem to Herbert start with this html and have a look what users really need. To summarize we do not have many choices: 1/ the patch you propose: use html.sty, which will only work for latex2html. The reason I am reluctant to do so is that I know that people favor other converters wrt latex2html, and we will basically break support for these converters. 2/ using hyperref in some nice way. We will need hyperref support for pdf, anyway. So this may be a longterm solution, but not viable right now since latex2html just ignores it. 3/ tweaking the latex generation code to produce latex suitable for such or such use (latex2html, whatever). This would be the best, but even more longterm. So it seems we are left with your approach :) I have a question on the patch, though: the functions LFUN_URL, LFUN_HTML and LFUN_HTMLURL seems to all do the same thing after your change? Do you mean that we shall remove all of the except one? Does someone know what were the differences between them? Also it would be nice if the patch did not try to remove double LF btween methods and change tab-indentation to 4char indentation :) But this is secondary. I cannot really comment on the interface changes, since this is not what I know best. JMarc
C++ question
If I have a mapkey, value some_map, can I create a vector of the keys also,
vectorkey const ref_vec? I can create a vectorkey const * ptr_vec.
I'm thinking of my Bibtex data base of course.
/** All Bibliography keys derive from this base class, and need only define
* code() to be instantiated.
*/
class BibKey {
public:
enum Code {
BIBLIO_CODE,
BIBTEX_CODE
};
virtual ~BibKey() {}
virtual BibKey::Code code() const = 0;
std::string const get() const { return key_; }
private:
std::string key_;
};
class BibtexKey : public BibKey {
public:
BibKey::Code code() const { return BibKey::BIBTEX_CODE; }
};
/** The BibTeX database stores a map of the data so,
* where Bibitem is the info
*/
typedef std::mapBibtexKey, Bibitem BMap;
///
BMap data_;
/** But it also stores a vector of the keys to pass back to the
* citation dialog. Note that the citation dialog gets a vector of Bibkeys,
* not BibtexKeys or BiblioKeys.
*/
std::vectorBibKey const * keys_;
Once I have the data, I can create the keys_, so:
keys_.resize(data_.size());
std::vectorBibKey const *::iterator key_it = keys_.begin();
BMap::const_iterator map_it = data_.begin();
BMap::const_iterator map_end = data_.end();
for (; map_it != map_end; ++map_it, ++key_it) {
*key_it = map_it-first;
}
But I can't do it if the keys_ is a vector of references, not pointers. Is
this impossible?
Angus
Re: C++ question
On Wednesday 13 March 2002 6:08 pm, Lars Gullik Bjønnes wrote: Angus Leeming [EMAIL PROTECTED] writes: | If I have a mapkey, value some_map, can I create a vector of the keys also, | vectorkey const ref_vec? I can create a vectorkey const * ptr_vec. | I'm thinking of my Bibtex data base of course. You cannot reset an reference; string a(Hello There!); string b = a; string c(How are you?); b = c; // error Then ponder about what the problem with vectorType const is. Gotcha! I create a vector of references and then I try to set their value. Thank you. A
Keyboard handling
ouch ! Klyx punts the issue by not using Qt's key event structure. I think we need a buffer view slot : void, string const text, unsigned int count, LKeySym where LKeySym is our own enum/class for all the special keys like F1, Up, PageDown etc. The reason I don't want to use X key events in the Qt frontend is mainly so we can use setKeyCompression(true). It can be a significant benefit to be able to squish 2 or 3 characters together as one event. I've got a little Qt prog for receiving key events, and it looks like compression will work even in the presence of composing LyX funcs. Of course for real compose, Qt should be doing this for us already. Of course there are the things like getiso() and stuff which really need to be fixed. Can we do anything usefel with the unicode inside the QString from the Qt event ? thanks john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Keyboard handling
On Wed, Mar 13, 2002 at 09:44:55PM +, John Levon wrote: stupid question, but how do normal keypresses like 'a' end up so that a lookup() in the kbmap returns LFUN_INSERT ??? thanks john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
[PATCH] Lar's optimisation font patch broken
Load Intro.lyx and look at the You tell LyX what you're doing.
A fencepost error, but they accumulate in the saved version.
The fix is attached. I don't know/care if it's right.
thanks
john
--
I am a complete moron for forgetting about endianness. May I be
forever marked as such.
Index: src/ChangeLog
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/ChangeLog,v
retrieving revision 1.622
diff -u -r1.622 ChangeLog
--- src/ChangeLog 13 Mar 2002 18:23:38 - 1.622
+++ src/ChangeLog 13 Mar 2002 22:26:47 -
@@ -1,3 +1,8 @@
+2002-03-13 John Levon [EMAIL PROTECTED]
+
+ * paragraph.C: fix fencepost error in getFontSettings() from
+ last patch
+
2002-03-13 Lars Gullik Bjønnes [EMAIL PROTECTED]
* paragraph.C: change several algorithm to be more appripriate for
Index: src/paragraph.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/paragraph.C,v
retrieving revision 1.197
diff -u -r1.197 paragraph.C
--- src/paragraph.C 13 Mar 2002 18:23:37 - 1.197
+++ src/paragraph.C 13 Mar 2002 22:26:51 -
@@ -595,7 +595,7 @@
Pimpl::FontList::const_iterator cit = pimpl_-fontlist.begin();
Pimpl::FontList::const_iterator end = pimpl_-fontlist.end();
for (; cit != end; ++cit) {
- if (cit-pos() pos)
+ if (cit-pos() = pos)
break;
}
@@ -1690,6 +1690,7 @@
}
if (moving_arg)
os \\protect ;
+
os \n;
}
texrow.newline();
Re: [PATCH] Lar's optimisation font patch broken
On Wed, Mar 13, 2002 at 11:54:53PM +0100, Lars Gullik Bjønnes wrote: This fix should be applied to all places where I did this change then. OK, I'll let you do that because I don't know how :) Currently you the first characeter of a word is not getting set to a font change either. Oh and you can't change the dummy position's font, OR apply a font change from the dialog to a selection... If you're quick, this might only get reported 4 or 5 times :) regards john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 12:08:35AM +0100, Lars Gullik Bjønnes wrote: btw. My name is not Lar... Ha ! And my name isn't Levon. So I suppose we are equal (on a side note: Is that a name anywhere?) My English skills are inversely proportional to the speed of my internet connection ... For some reason I choose to write all my email via a ssh connection. building with your patch, john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 12:00:12AM +0100, Lars Gullik Bjønnes wrote: This should fix this and the rest of them, please test. It fixes the problems Ive mentioned, but there is another: the '!' icon is broken. Apply Italic font change via the dialog to the current cursor position, and it changes, but the icon doesn't. Perhaps this is one of my changes ? regards john
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 12:21:09AM +0100, Lars Gullik Bjønnes wrote: If I just move the cursor over some '!' text, the toolbar button gets pressed. Yes this works OK for me, but just typing after changes doesn't i.e. the icons don't represent what the next char will look like john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 12:35:58AM +0100, Lars Gullik Bjønnes wrote: Either I do not understand what you mean, or I don't have the problem. Actually this might be on older bug, it happens on an older 1.2 build ... start new doc. type some text. change character dialog to italic and apply (no change in button) type some text (italic) BUT the emph button is not on. Now press emph button (to select it) and type some text. Note something odd ? The button is the exact opposite in operation to what you would expect ! regards john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 12:50:29AM +0100, Lars Gullik Bjønnes wrote: cursor in front of 1: icon on hmm are you sure that the icon becomes selected when you click apply from the dialog ? If you're moving the cursor with arrow keys then the problem won't be seen. Another problem (UserGuide as usual bring out all the nasty bugs) : #6 0x80c4d38 in Paragraph::size (this=0x0) at paragraph.C:2119 #7 0x80e9cb1 in LyXText::selectionAsString (this=0x8518870, buffer=0x83891c0, label=false) at text2.C:1012 #8 0x8058d6f in BufferView::Pimpl::selectionRequested (this=0x83c0330) at BufferView_pimpl.C:752 so tmpcur.par() is 0. Ideas ?? john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Thu, Mar 14, 2002 at 01:12:53AM +0100, Lars Gullik Bjønnes wrote: | hmm are you sure that the icon becomes selected when you click apply | from the dialog ? Dialog? You said nothing about using a dialog... Well apart from the two or three times I mentioned using it, including the step by step :P How should I move it then? mouse? don't move it at all ! new doc type a apply dialog italic type b -- icon doesn't change press emph icon (so it becomes on) type c -- c is in normal text but icon says otherwise no, not now must sleep brain say god night now God Night ! john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Wed, Mar 13, 2002 at 08:36:20PM -0400, Garst R. Reese wrote: You mean from Layout-Character I presume. When I do that the dialog remains up and I have to close it. You have to close it ?? Why ? But yes, I see the problem also. The OK thanks toggle on the icon gets out of step with reality. Let's hope Lars gets a good night's sleep. Or a god night's sleep, even better john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Wed, Mar 13, 2002 at 09:24:25PM -0400, Garst R. Reese wrote: Well, apply leaves it up and in the way and it's easier to close than switch views. Maybe this is not a problem with more pixels. Or you could just press alt-tab or whatever your WM does to switch window stacking order... regards john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: [PATCH] Lar's optimisation font patch broken
On Wed, Mar 13, 2002 at 09:37:33PM -0400, Garst R. Reese wrote: I am a complete moron. hey we can form a club ! You have to say so in your signature too though, or you can't join :) john -- I am a complete moron for forgetting about endianness. May I be forever marked as such.
Re: Lyx
On Thu, 14 Mar 2002, Marianne Doyle wrote: Dear Dr Rae, I'm still only a mister :( but I'm working on it! Ian Mortimer from the physics dept said you may be able to help me with a lyx problem. Is there any way to overcome the problem of not being able to use the number key pad inside tables. At the moment if you use the number key pad the number appear outside the table. There is a fix for this problem in the sources for the up-coming 1.2.0 release. That release is still a few weeks (a month, maybe two?) away though and it is probably a bit too volatile for you to consider using the bleeding edge for your work (I am, but I also help write it). Jürgen, is this an easy fix you could backport to 1.1.6? Allan. (ARRae)
[] lyx-devel ?
Title: emailsale [lyx-devel´ÔÀÇ ¸ÞÀÏÁÖ¼Ò´Â ÀÌ À¥ÆäÀÌÁö¿¡¼ º¸°í ¹ß¼ÛÇß½À´Ï´Ù. Ç÷¯½º ³ëºí·¹½º(www.Plusnoblesse.com) ¿¡¼´Â ¸íÇ° °¡°ÝÀÎÇÏ µî À¯¿ëÇÑ Á¤º¸¸¦ Á¦°øÇÒ ¿¹Á¤À̳ª ¼ö½ÅÀ» ¿øÄ¡ ¾ÊÀ¸½Ã¸é ´Ù½Ã ¿î¿µÀÚÀÇ ¸ÞÀÏÀ» ¹Þ°Ô µÇ´Â ÀÏÀÌ ¾øµµ·Ï ÇÏ°Ú½À´Ï´Ù.]
Re: Tooltips revamp
On 11 Mar 2002, Jean-Marc Lasgouttes wrote: Allan == Allan Rae [EMAIL PROTECTED] writes: Oops not so easy. I'll need another static signal in Dialogs.h static bool SigC::Signal0bool Dialogs::tooltipsEnabled; Allan Why? The tooltips should have a lyxrc entry and you can toggle Allan based on that value instead. Otherwise someone who wants/hates Allan tooltips has to turn them on/off in every session! No, because the tooltips are designed to be lengthy, helpful and annoying when you do not need them. So it is reasonable to turn them off by default. I don't see a problem. Sure, some of the tooltips are long but you have to have your mouse over a button for them to appear. I have no problem with this. The thing I do have a problem with is that it is not really a What's this? mechanism. The new cursor shape implies that I should click on something that I'm asking about -- as learned by using KDE just once. AFAIAC we should just allow enable/disable of the tooltips and leave the cursor alone. The code that is in CVS works fine otherwise (ie. get rid of the ? cursor and all is excellent). BTW, we will need an icon for this toggle, since operating tooltips is a mouse-based thing. Seems to be getting more and more complicated. BTW, I'm inclined to think that if we must have a menu entry for this then is should be Edit-Tooltips [] (located near the preferences stuff perhaps). As it is now it looks like it is a part of the documentation and worse yet the proximity and similarity to the Documents menu had me wondering if it wasn't an indicator of a file I had opened. Can we set the marker for the current buffer in the documents menu to a different colour or shape so it doesn't look like a toggle? Allan. (ARRae)
lyx-1.1.6fix4 bug: can't use cite.sty
I am trying to use cite.sty in lyx. Unfortunately thje packages lsiterd in the latex preamble of lyx are inserted after everything else. If cite shows up after babel is is ignored! Any ideas how to fix this? Ulrich
Re: [Devel] Bug list update again
On Wed, 13 Mar 2002, Herbert Voss wrote: Jean-Marc Lasgouttes wrote: To express things differently, I am not sure that showing everything wysiwyg is a panacea. For example, I would hate to see equation numbers in mathed showed as just numbers instead of labels. But may is it just because I am used to give meaningful names to my labels. we are talking about citations and natbib. The analogy is still valid though. And for citations, seeing [1] in the lyx window instead of [Key] is certainly not a plus, forgetting for now that I do not get anything. did I ever said, that you'll see [1] as label?? I suppose that you don't tried the patch but give arguments against ... He sent a screen shot didn't he? we are talking about natbib and not the plain version of an bst-file. What about non-natbib, regular old numbered citations like JMarc and I use? I haven't tested your patch but if JMarc is able to provide screenshots which don't show anything useful for a citation then I'm also going to be annoyed. I certainly don't want a [1]. I'd much rather have a [Key]. Will you answer this challenge then if I ask it instead of JMarc? What do you get if you turn off natbib? What do the citations appear as? I've just spent half the day catching up on 500 emails from the last two days and gotten to this conversation last so please excuse my lack of time to test your patch today -- I've only just finished absorbing most of the major changes in CVS. Allan. (ARRae)
[PATCH] Fix GNOME compilation Part 1
Hey guys, I've just spent some time getting the gtkmm frontend to compile again. I don't quite know how to work the build system so I've split the task into two. The attached patch fixes the compilation problems, please review and apply. However, linking still fails, John Levon tells me I need something like src/frontends/qt2/xforms but I'm having trouble getting that to work. If I haven't got it going by this time tommorow I'll post my work to date and ask for help. Cheers for now -- | Michael Koziarski |Conventional wisdom is often | | Data Engineer, Linux user | long on convention and short | | Objectivist. | on wisdom -- | | http://www.koziarski.com| Warren E. Buffett, BRK.A |
hmpf
I think I do this more than anyone else
--
| Michael Koziarski |Conventional wisdom is often |
| Data Engineer, Linux user | long on convention and short |
| Objectivist. | on wisdom -- |
| http://www.koziarski.com| Warren E. Buffett, BRK.A |
Index: ChangeLog
===
RCS file: /cvs/lyx/lyx-devel/ChangeLog,v
retrieving revision 1.823
diff -u -r1.823 ChangeLog
--- ChangeLog 2002/03/13 13:24:48 1.823
+++ ChangeLog 2002/03/14 07:35:02
@@ -1,3 +1,8 @@
+2002-03-14 Michael A. Koziarski [EMAIL PROTECTED]
+
+ * configure.in: First stage of fixing the gnome compile problems.
+ Linking problems to be fixed next.
+
2002-03-12 Jean-Marc Lasgouttes [EMAIL PROTECTED]
* README: add bulgarian
Index: configure.in
===
RCS file: /cvs/lyx/lyx-devel/configure.in,v
retrieving revision 1.97
diff -u -r1.97 configure.in
--- configure.in2002/03/12 11:32:24 1.97
+++ configure.in2002/03/14 07:35:11
@@ -198,6 +198,8 @@
`
;;
gnome)
+LYX_PATH_XPM
+LYX_PATH_XFORMS
AM_PATH_GTKMM(1.2.1,,
AC_MSG_ERROR(Cannot find GTK--: Please install Version 1.2.1+))
AM_PATH_GNOMEMM
@@ -217,8 +219,8 @@
FRONTEND=xforms gnome
FRONTEND_GUILIB=gnome/*.lo
FRONTEND_LDFLAGS=\$(GNOMEMM_LIBDIR)
-FRONTEND_INCLUDES=-I\${srcdir}/gnome \$(GNOMEMM_INCLUDEDIR) \$(GTKMM_CFLAGS)
`gnome-config --cflags libglade gnomeui`
-FRONTEND_LIBS=\$(GNOMEMM_LIBS) \$(GTKMM_LIBS) `gnome-config --libs libglade
gnomeui`;;
+FRONTEND_INCLUDES=-I\${srcdir}/gnome -I\${srcdir}/xforms \$(GNOMEMM_INCLUDEDIR)
+\$(GTKMM_CFLAGS) `gnome-config --cflags libglade gnomeui`
+FRONTEND_LIBS=@XPM_LIB@ @XFORMS_LIB@ \$(GNOMEMM_LIBS) \$(GTKMM_LIBS)
+`gnome-config --libs libglade gnomeui`;;
qt2)
LYX_PATH_XPM
LYX_PATH_XFORMS
Index: po/POTFILES.in
===
RCS file: /cvs/lyx/lyx-devel/po/POTFILES.in,v
retrieving revision 1.268
diff -u -r1.268 POTFILES.in
--- po/POTFILES.in 2002/03/13 18:24:42 1.268
+++ po/POTFILES.in 2002/03/14 07:35:24
@@ -9,7 +9,6 @@
src/CutAndPaste.C
src/debug.C
src/exporter.C
-src/ext_l10n.h
src/FontLoader.C
src/frontends/controllers/biblio.C
src/frontends/controllers/ButtonController.h
Index: src/frontends/gnome/ChangeLog
===
RCS file: /cvs/lyx/lyx-devel/src/frontends/gnome/ChangeLog,v
retrieving revision 1.38
diff -u -r1.38 ChangeLog
--- src/frontends/gnome/ChangeLog 2002/03/11 22:47:40 1.38
+++ src/frontends/gnome/ChangeLog 2002/03/14 07:35:32
@@ -1,3 +1,10 @@
+2002-03-14 Michael A. Koziarski [EMAIL PROTECTED]
+
+ * Timeout_pimpl.C
+ * Timeout_pimpl.h: implemented running to fix compilation, linking
+ fix still to come
+
+
2002-03-11 Lars Gullik Bjønnes [EMAIL PROTECTED]
* docygen fixes.
Index: src/frontends/gnome/Timeout_pimpl.C
===
RCS file: /cvs/lyx/lyx-devel/src/frontends/gnome/Timeout_pimpl.C,v
retrieving revision 1.2
diff -u -r1.2 Timeout_pimpl.C
--- src/frontends/gnome/Timeout_pimpl.C 2002/03/11 22:47:40 1.2
+++ src/frontends/gnome/Timeout_pimpl.C 2002/03/14 07:35:33
@@ -4,6 +4,7 @@
* Read COPYING
*
* \author Baruch Even
+ * \author Michael Koziarski
*/
#include config.h
@@ -28,7 +29,11 @@
stop();
}
-
+bool Timeout::Pimpl::running() const
+{
+return running_;
+}
+
void Timeout::Pimpl::start()
{
if (conn_.connected()) {
@@ -40,12 +45,14 @@
SigC::slot(this, Timeout::Pimpl::timeoutEvent),
owner_-timeout_ms
);
+ running_ = true;
}
void Timeout::Pimpl::stop()
{
conn_.disconnect();
+ running_ = false;
}
Index: src/frontends/gnome/Timeout_pimpl.h
===
RCS file: /cvs/lyx/lyx-devel/src/frontends/gnome/Timeout_pimpl.h,v
retrieving revision 1.4
diff -u -r1.4 Timeout_pimpl.h
--- src/frontends/gnome/Timeout_pimpl.h 2002/03/11 22:47:40 1.4
+++ src/frontends/gnome/Timeout_pimpl.h 2002/03/14 07:35:33
@@ -31,7 +31,10 @@
/// stop the timer
void stop();
/// reset
- void reset();
+void reset();
+/// Is the timer running?
+bool running() const;
+
public:
/// The timeout signal, this gets called when the timeout passed.
@@ -41,6 +44,8 @@
Timeout * owner_;
/// Timer connection
SigC::Connection conn_;
+/// Used for running as
Re: makefile patch
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: | Lars This is a think I have been planning to do with the makefiles | Lars for a long time. I am sure there are details that can be | Lars imporved, but this should make sure that the Makefile.am's are a | Lars lot more uniform. | Lars Please comment. | I'm not sure I like the contents of your include file. | +AUTOMAKE_OPTIONS = foreign 1.5 | + | So you want to force automake 1.5? Can't this wait for 1.3.0cvs? I just saw that we used foreign 1.4 in a couple of files, I'll be happy with just foreign | +DISTCLEANFILES= -r *.orig *.rej *~ *.bak lyx.1 core libintl.h config.h |config.status config.cache config.log \ | + $(srcdir)/sourcedoc/*~ \ | + $(srcdir)/sourcedoc/html $(srcdir)/sourcedoc/latex \ | + $(srcdir)/sourcedoc/man $(srcdir)/sourcedoc/Doxyfile | Most of the files there can only exist in _some_ directory (toplevel | or src). It does not make sense to mention them in a file included | everywhere. I don't like that. That is true, but it just makes things easier. | +MAINTAINERCLEANFILES = $(srcdir)/Makefile.in $(srcdir)/config.h.in \ | + $(srcdir)/aclocal.m4 \ | + $(srcdir)/configure $(srcdir)/development/lyx.spec \ | + $(srcdir)/acinclude.m4 $(srcdir)/sourcedoc/Doxyfile | Here also, only $(srcdir)/Makefile.in is generic and should be here. It makes things easier, and although we get some clutter and unwanted stuff in include.am, we get a much nicer Makefile.am. -- Lgb
Re: getFontSettings optimization
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes:
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes:
| Lars I think this is a better algorithm for font lookup. (at least
| Lars gprof tells me so)
| This is probably true, since there are very few font changes in a
| typical paragraph. What are the getFontSettings numbers down to?
Each sample counts as 0.01 seconds.
% cumulative self self total
time seconds secondscalls us/call us/call name
7.52 0.31 0.31 2043074 0.15 0.29
LyXTextClass::operator[](std::string const) const
6.80 0.59 0.28 1397195 0.20 1.31 LyXText::getFont(Buffer const*,
Paragraph*, int) const
6.31 0.85 0.26 1401073 0.19 0.23
Paragraph::getFontSettings(BufferParams const, int) const
| I've been wanting to work on getFontSetting after 1.2.0, so I would
| like to see whether this very simple approach is enough.
I have done the same change to other functions that used matchFT and
lowerbound.
Index: paragraph.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/paragraph.C,v
retrieving revision 1.196
diff -u -p -r1.196 paragraph.C
--- paragraph.C 7 Mar 2002 10:51:45 - 1.196
+++ paragraph.C 13 Mar 2002 09:23:46 -
@@ -581,11 +581,21 @@ LyXFont const Paragraph::getFontSettings
pos_type pos) const
{
lyx::Assert(pos = size());
-
+
+#if 0
Pimpl::FontTable search_font(pos, LyXFont());
Pimpl::FontList::const_iterator cit = lower_bound(pimpl_-fontlist.begin(),
pimpl_-fontlist.end(),
search_font, Pimpl::matchFT());
+#else
+ Pimpl::FontList::const_iterator cit = pimpl_-fontlist.begin();
+ Pimpl::FontList::const_iterator end = pimpl_-fontlist.end();
+ for (; cit != end; ++cit) {
+ if (cit-pos() pos)
+ break;
+ }
+#endif
+
LyXFont retfont;
if (cit != pimpl_-fontlist.end()) {
retfont = cit-font();
@@ -680,20 +690,36 @@ Paragraph::highestFontInRange(pos_type s
return def_size;
LyXFont::FONT_SIZE maxsize = LyXFont::SIZE_TINY;
+#if 0
Pimpl::FontTable end_search(endpos, LyXFont());
Pimpl::FontList::const_iterator end_it =
lower_bound(pimpl_-fontlist.begin(),
pimpl_-fontlist.end(),
end_search, Pimpl::matchFT());
+#else
+ Pimpl::FontList::const_iterator end_it = pimpl_-fontlist.begin();
+ Pimpl::FontList::const_iterator end = pimpl_-fontlist.end();
+ for (; end_it != end; ++end_it) {
+ if (end_it-pos() endpos)
+ break;
+ }
+#endif
if (end_it != pimpl_-fontlist.end())
++end_it;
+#if 0
Pimpl::FontTable start_search(startpos, LyXFont());
Pimpl::FontList::const_iterator cit =
lower_bound(pimpl_-fontlist.begin(),
pimpl_-fontlist.end(),
start_search, Pimpl::matchFT());
-
+#else
+ Pimpl::FontList::const_iterator cit = pimpl_-fontlist.begin();
+ for (; cit != end; ++cit) {
+ if (cit-pos() startpos)
+ break;
+ }
+#endif
for (; cit != end_it; ++cit) {
LyXFont::FONT_SIZE size = cit-font().size();
if (size == LyXFont::INHERIT_SIZE)
@@ -757,13 +783,23 @@ void Paragraph::setFont(pos_type pos, Ly
// in a new kernel. (Asger)
// Next search font table
+#if 0
Pimpl::FontTable search_font(pos, LyXFont());
Pimpl::FontList::iterator it = lower_bound(pimpl_-fontlist.begin(),
pimpl_-fontlist.end(),
search_font, Pimpl::matchFT());
unsigned int i = it - pimpl_-fontlist.begin();
bool notfound = it == pimpl_-fontlist.end();
-
+#else
+ Pimpl::FontList::iterator it = pimpl_-fontlist.begin();
+ Pimpl::FontList::iterator endit = pimpl_-fontlist.end();
+ for (; it != endit; ++it) {
+ if (it-pos() pos)
+ break;
+ }
+ unsigned int i = std::distance(pimpl_-fontlist.begin(), it);
+ bool notfound = (it == endit);
+#endif
if (!notfound pimpl_-fontlist[i].font() == font)
return;
--
Lgb
Re: getFontSettings optimization
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: | Lars Each sample counts as 0.01 seconds. | [...] | getFontSettings is down to .23ms/call from 0.89. This is good. | Lars | I've been wanting to work on getFontSetting after 1.2.0, so I | Lars would | like to see whether this very simple approach is enough. | Lars I have done the same change to other functions that used matchFT | Lars and lowerbound. | Looks like you create common code that should be factored in some | helper function. Perhaps, but I am not going to do that now. -- Lgb
Re: small stuff
Andre Poenitz [EMAIL PROTECTED] writes: | * whitespace changes | * member var with trailing underscore I try to stay with the rule that only private class variables get the trailing underscore. How hard is it to make owner_ private? -- Lgb
Re: small stuff
Andre Poenitz [EMAIL PROTECTED] writes: | On Wed, Mar 13, 2002 at 12:35:06PM +0100, Lars Gullik Bjønnes wrote: How hard is it to make owner_ private? | It is only use a couple of times to call a single function: |owner_.getPixmap() | So we could have |a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. | or |a protected 'owner() const' | Your call. | Andre' | -- | André Pönitz .. [EMAIL PROTECTED] -- Lgb
Re: small stuff
Andre Poenitz [EMAIL PROTECTED] writes: | On Wed, Mar 13, 2002 at 12:50:25PM +0100, Lars Gullik Bjønnes wrote: | So we could have |a private 'owner_' and a protected 'getPixmap' a protected getPixmap is then probably best. | Uh... I just see we need Pixmap defined... | We do not to #include X11/X.h for that, do we? | or |a protected 'owner() const' | The better way than? I am stil for the getPixmap solution, but leave the includes as is. Then it should work right? -- Lgb
Re: makefile patch
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: | Lars I just saw that we used foreign 1.4 in a couple of files, I'll | Lars be happy with just foreign | OK. | Lars It makes things easier, and although we get some clutter and | Lars unwanted stuff in include.am, we get a much nicer Makefile.am. | I think it is better to add the clutter in the Makefiles.am that need | it. Isn't it possible to do | VAR += foo bar | in Makefiles? Remember that Makefile.am is not a makefile... I have no idea if it works or not. -- Lgb
Painter diff
These are the changes I have in my tree now:
Index: PainterBase.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.C,v
retrieving revision 1.10
diff -u -p -r1.10 PainterBase.C
--- PainterBase.C 30 Jul 2001 11:55:47 - 1.10
+++ PainterBase.C 13 Mar 2002 12:23:00 -
-18,9 +18,6
#include WorkArea.h
#include font.h
-int PainterBase::dummy1 = 0;
-int PainterBase::dummy2 = 0;
-int PainterBase::dummy3 = 0;
int PainterBase::paperMargin() const
{
-67,23 +64,23 PainterBase PainterBase::buttonFrame(i
{
// Width of a side of the button
int d = 2;
-
+
fillRectangle(x, y, w, d, LColor::top);
fillRectangle(x, (y+h-d), w, d, LColor::bottom);
-
+
// Now a couple of trapezoids
int x1[4], y1[4];
-
- x1[0] = x+d; y1[0] = y+d;
- x1[1] = x+d; y1[1] = (y+h-d);
- x1[2] = x; y1[2] = y+h;
+
+ x1[0] = x + d; y1[0] = y + d;
+ x1[1] = x + d; y1[1] = (y + h - d);
+ x1[2] = x; y1[2] = y + h;
x1[3] = x; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::left);
- x1[0] = (x+w-d); y1[0] = y+d;
- x1[1] = (x+w-d); y1[1] = (y+h-d);
- x1[2] = x+w; y1[2] = (y+h-d);
- x1[3] = x+w; y1[3] = y;
+ x1[0] = (x + w - d); y1[0] = y + d;
+ x1[1] = (x + w - d); y1[1] = (y + h - d);
+ x1[2] = x + w; y1[2] = (y + h - d);
+ x1[3] = x + w; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::right);
return *this;
Index: PainterBase.h
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.h,v
retrieving revision 1.17
diff -u -p -r1.17 PainterBase.h
--- PainterBase.h 27 Feb 2002 09:59:49 - 1.17
+++ PainterBase.h 13 Mar 2002 12:23:00 -
-36,13 +36,6 namespace grfx {
*/
class PainterBase {
-protected:
-///
- static int dummy1;
- ///
- static int dummy2;
- ///
- static int dummy3;
public:
///
enum line_width {
-51,7 +44,7 public:
///
line_thick
};
-
+
///
enum line_style {
///
-61,13 +54,13 public:
///
line_onoffdash
};
-
+
///
explicit PainterBase(WorkArea wa) : owner(wa) {}
///
virtual ~PainterBase() {}
-
+
/* Screen geometry */
///
int paperMargin() const;
-82,7 +75,7 public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/** Draw the lines between the lines in xp and yp.
xp and yp are arrays of points, and np is the
number of them. */
-91,13 +84,13 public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/// Here xp and yp are arrays of points
virtual PainterBase fillPolygon(
int const * xp, int const * yp,
int np,
LColor::color = LColor::foreground) = 0;
-
+
/// Draw lines from x1,y1 to x2,y2. They are arrays
virtual PainterBase segments(
int const * x1, int const * y1,
-117,7 +110,7 public:
virtual PainterBase circle(
int x, int y, unsigned int d,
LColor::color = LColor::foreground);
-
+
/// Draw an ellipse
virtual PainterBase ellipse(
int x, int y,
-143,7 +136,7 public:
/// A filled rectangle with the shape of a 3D button
virtual PainterBase button(int x, int y, int w, int h);
-
+
///
virtual PainterBase buttonFrame(int x, int y, int w, int h);
-155,22 +148,22 public:
/// Draw a string at position x, y (y is the baseline)
virtual PainterBase text(int x, int y,
string const str, LyXFont const f) = 0;
-
+
/** Draw a string at position x, y (y is the baseline)
This is just for fast drawing */
virtual PainterBase text(int x, int y, char const * str, size_t l,
- LyXFont const f) = 0;
-
+ LyXFont const f) = 0;
+
/// Draw a char at position x, y (y is the baseline)
virtual PainterBase text(int x, int y, char c, LyXFont const f)=0;
-
+
/** Draws a string and encloses it inside a rectangle. */
PainterBase rectText(int x, int baseline,
string const string,
LyXFont const font,
LColor::color back,
Re: Painter diff
Angus Leeming [EMAIL PROTECTED] writes:
| On Wednesday 13 March 2002 12:25 pm, Lars Gullik Bjønnes wrote:
Index: Painter.h
===
private:
+/// For the figinset
+PainterBase pixmap(int x, int y, int w, int h, Pixmap bitmap);
+
| Errr, why? Helper functions are fine if they're used by more than one
| function.
In general I strongly disagree with that statement.
| but this is just daft. You've already proved to yourself that pixmap
| is used only with Painter.C, so now's the time to dump it.
It is used in image... and the minimal change is this one.
| Angus
|
| PainterBase Painter::pixmap(int x, int y, int w, int h, Pixmap bitmap)
| {
| XGCValues val;
| val.function = GXcopy;
| GC gc = XCreateGC(display, owner.getPixmap(),
| GCFunction, val);
| XCopyArea(display, bitmap, owner.getPixmap(), gc,
| 0, 0, w, h, x, y);
| XFreeGC(display, gc);
| return *this;
| }
| PainterBase Painter::image(int x, int y, int w, int h,
| grfx::GImage const image)
| {
| return pixmap(x, y, w, h, image.getPixmap());
| }
--
Lgb
Re: Painter diff
[EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
| Angus Leeming [EMAIL PROTECTED] writes:
| | On Wednesday 13 March 2002 12:25 pm, Lars Gullik Bjønnes wrote:
Index: Painter.h
===
private:
+ /// For the figinset
+ PainterBase pixmap(int x, int y, int w, int h, Pixmap bitmap);
+
| | Errr, why? Helper functions are fine if they're used by more than one
| | function.
| In general I strongly disagree with that statement.
| | but this is just daft. You've already proved to yourself that pixmap
| | is used only with Painter.C, so now's the time to dump it.
so you are more happy with this patch then:
Index: PainterBase.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.C,v
retrieving revision 1.10
diff -u -p -r1.10 PainterBase.C
--- PainterBase.C 30 Jul 2001 11:55:47 - 1.10
+++ PainterBase.C 13 Mar 2002 12:54:44 -
@@ -18,9 +18,6 @@
#include WorkArea.h
#include font.h
-int PainterBase::dummy1 = 0;
-int PainterBase::dummy2 = 0;
-int PainterBase::dummy3 = 0;
int PainterBase::paperMargin() const
{
@@ -67,23 +64,23 @@ PainterBase PainterBase::buttonFrame(i
{
// Width of a side of the button
int d = 2;
-
+
fillRectangle(x, y, w, d, LColor::top);
fillRectangle(x, (y+h-d), w, d, LColor::bottom);
-
+
// Now a couple of trapezoids
int x1[4], y1[4];
-
- x1[0] = x+d; y1[0] = y+d;
- x1[1] = x+d; y1[1] = (y+h-d);
- x1[2] = x; y1[2] = y+h;
+
+ x1[0] = x + d; y1[0] = y + d;
+ x1[1] = x + d; y1[1] = (y + h - d);
+ x1[2] = x; y1[2] = y + h;
x1[3] = x; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::left);
- x1[0] = (x+w-d); y1[0] = y+d;
- x1[1] = (x+w-d); y1[1] = (y+h-d);
- x1[2] = x+w; y1[2] = (y+h-d);
- x1[3] = x+w; y1[3] = y;
+ x1[0] = (x + w - d); y1[0] = y + d;
+ x1[1] = (x + w - d); y1[1] = (y + h - d);
+ x1[2] = x + w; y1[2] = (y + h - d);
+ x1[3] = x + w; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::right);
return *this;
Index: PainterBase.h
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.h,v
retrieving revision 1.17
diff -u -p -r1.17 PainterBase.h
--- PainterBase.h 27 Feb 2002 09:59:49 - 1.17
+++ PainterBase.h 13 Mar 2002 12:54:44 -
@@ -36,13 +36,6 @@ namespace grfx {
*/
class PainterBase {
-protected:
-///
- static int dummy1;
- ///
- static int dummy2;
- ///
- static int dummy3;
public:
///
enum line_width {
@@ -51,7 +44,7 @@ public:
///
line_thick
};
-
+
///
enum line_style {
///
@@ -61,13 +54,13 @@ public:
///
line_onoffdash
};
-
+
///
explicit PainterBase(WorkArea wa) : owner(wa) {}
///
virtual ~PainterBase() {}
-
+
/* Screen geometry */
///
int paperMargin() const;
@@ -82,7 +75,7 @@ public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/** Draw the lines between the lines in xp and yp.
xp and yp are arrays of points, and np is the
number of them. */
@@ -91,13 +84,13 @@ public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/// Here xp and yp are arrays of points
virtual PainterBase fillPolygon(
int const * xp, int const * yp,
int np,
LColor::color = LColor::foreground) = 0;
-
+
/// Draw lines from x1,y1 to x2,y2. They are arrays
virtual PainterBase segments(
int const * x1, int const * y1,
@@ -117,7 +110,7 @@ public:
virtual PainterBase circle(
int x, int y, unsigned int d,
LColor::color = LColor::foreground);
-
+
/// Draw an ellipse
virtual PainterBase ellipse(
int x, int y,
@@ -143,7 +136,7 @@ public:
/// A filled rectangle with the shape of a 3D button
virtual PainterBase button(int x, int y, int w, int h);
-
+
///
virtual PainterBase buttonFrame(int x, int y, int w, int h);
@@ -155,22 +148,22 @@ public:
/// Draw a string at position x, y (y is the baseline)
virtual PainterBase text(int x, int y,
string const str, LyXFont const f) = 0;
-
+
/** Draw a string at position x, y (y is the baseline)
This is just for fast drawing */
virtual PainterBase text(int x, int y, char const * str, size_t l,
- LyXFont const f) = 0;
-
+ LyXFont const f) = 0;
+
/// Draw a char at position x, y (y is the baseline)
virtual PainterBase text(int x, int y, char c, LyXFont const f)=0;
-
+
/** Draws a string and encloses it inside a rectangle. */
PainterBase rectText(int x, int baseline,
string const string,
LyXFont const font,
LColor::color back,
LColor::color frame);
-
+
/** Draw a string and encloses it inside a button frame. */
PainterBase buttonText(int x, int baseline, string const s,
LyXFont const font);
Index: Painter.C
Re: Painter diff
Angus Leeming [EMAIL PROTECTED] writes:
| Yes. In this case I think that the helper function isn't a help.
I have yet another patch for you too look at, the thing missing now is
to get rid of the XChar2b stuff. We want to use wchar_t in interfaces.
(and if wchar_t is not supported, then we should not support XChar2b
stuff either...)
Index: PainterBase.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.C,v
retrieving revision 1.10
diff -u -p -r1.10 PainterBase.C
--- PainterBase.C 30 Jul 2001 11:55:47 - 1.10
+++ PainterBase.C 13 Mar 2002 13:10:48 -
@@ -18,9 +18,6 @@
#include WorkArea.h
#include font.h
-int PainterBase::dummy1 = 0;
-int PainterBase::dummy2 = 0;
-int PainterBase::dummy3 = 0;
int PainterBase::paperMargin() const
{
@@ -67,23 +64,23 @@ PainterBase PainterBase::buttonFrame(i
{
// Width of a side of the button
int d = 2;
-
+
fillRectangle(x, y, w, d, LColor::top);
fillRectangle(x, (y+h-d), w, d, LColor::bottom);
-
+
// Now a couple of trapezoids
int x1[4], y1[4];
-
- x1[0] = x+d; y1[0] = y+d;
- x1[1] = x+d; y1[1] = (y+h-d);
- x1[2] = x; y1[2] = y+h;
+
+ x1[0] = x + d; y1[0] = y + d;
+ x1[1] = x + d; y1[1] = (y + h - d);
+ x1[2] = x; y1[2] = y + h;
x1[3] = x; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::left);
- x1[0] = (x+w-d); y1[0] = y+d;
- x1[1] = (x+w-d); y1[1] = (y+h-d);
- x1[2] = x+w; y1[2] = (y+h-d);
- x1[3] = x+w; y1[3] = y;
+ x1[0] = (x + w - d); y1[0] = y + d;
+ x1[1] = (x + w - d); y1[1] = (y + h - d);
+ x1[2] = x + w; y1[2] = (y + h - d);
+ x1[3] = x + w; y1[3] = y;
fillPolygon(x1, y1, 4, LColor::right);
return *this;
Index: PainterBase.h
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/PainterBase.h,v
retrieving revision 1.17
diff -u -p -r1.17 PainterBase.h
--- PainterBase.h 27 Feb 2002 09:59:49 - 1.17
+++ PainterBase.h 13 Mar 2002 13:10:48 -
@@ -36,13 +36,6 @@ namespace grfx {
*/
class PainterBase {
-protected:
-///
- static int dummy1;
- ///
- static int dummy2;
- ///
- static int dummy3;
public:
///
enum line_width {
@@ -51,7 +44,7 @@ public:
///
line_thick
};
-
+
///
enum line_style {
///
@@ -61,13 +54,13 @@ public:
///
line_onoffdash
};
-
+
///
explicit PainterBase(WorkArea wa) : owner(wa) {}
///
virtual ~PainterBase() {}
-
+
/* Screen geometry */
///
int paperMargin() const;
@@ -82,7 +75,7 @@ public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/** Draw the lines between the lines in xp and yp.
xp and yp are arrays of points, and np is the
number of them. */
@@ -91,13 +84,13 @@ public:
LColor::color = LColor::foreground,
enum line_style = line_solid,
enum line_width = line_thin) = 0;
-
+
/// Here xp and yp are arrays of points
virtual PainterBase fillPolygon(
int const * xp, int const * yp,
int np,
LColor::color = LColor::foreground) = 0;
-
+
/// Draw lines from x1,y1 to x2,y2. They are arrays
virtual PainterBase segments(
int const * x1, int const * y1,
@@ -117,7 +110,7 @@ public:
virtual PainterBase circle(
int x, int y, unsigned int d,
LColor::color = LColor::foreground);
-
+
/// Draw an ellipse
virtual PainterBase ellipse(
int x, int y,
@@ -143,7 +136,7 @@ public:
/// A filled rectangle with the shape of a 3D button
virtual PainterBase button(int x, int y, int w, int h);
-
+
///
virtual PainterBase buttonFrame(int x, int y, int w, int h);
@@ -155,22 +148,22 @@ public:
/// Draw a string at position x, y (y is the baseline)
virtual PainterBase text(int x, int y,
string const str, LyXFont const f) = 0;
-
+
/** Draw a string at position x, y (y is the baseline)
This is just for fast drawing */
virtual PainterBase text(int x, int y, char const * str, size_t l,
- LyXFont const f) = 0;
-
+ LyXFont const f) = 0;
+
/// Draw a char at position x, y (y is the baseline)
virtual PainterBase text(int x, int y, char c, LyXFont const f)=0;
-
+
/** Draws a string and encloses it inside a rectangle. */
PainterBase rectText(int x, int baseline,
string const string,
LyXFont const font,
LColor::color back,
LColor::color frame);
-
+
/** Draw a string and encloses it inside a button frame. */
PainterBase buttonText(int x, int baseline, string const s,
LyXFont const font);
Index: Painter.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/Painter.C,v
retrieving revision 1.32
diff -u -p -r1.32 Painter.C
--- Painter.C 27 Feb 2002 09:59:48 - 1.32
+++ Painter.C 13 Mar 2002 13:10:48 -
@@ -34,13 +34,26 @@
#include cmath
+
using std::endl;
using std::max;
Re: makefile patch
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: | Lars This is a think I have been planning to do with the makefiles | Lars for a long time. I am sure there are details that can be | Lars imporved, but this should make sure that the Makefile.am's are a | Lars lot more uniform. | Lars Please comment. | I'm not sure I like the contents of your include file. Ok, does this patch taste better? Index: Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/Makefile.am,v retrieving revision 1.28 diff -u -p -r1.28 Makefile.am --- Makefile.am 11 Mar 2002 23:02:03 - 1.28 +++ Makefile.am 13 Mar 2002 13:38:48 - @@ -1,19 +1,20 @@ -AUTOMAKE_OPTIONS = foreign -DISTCLEANFILES = -r *.orig *.rej *~ *.bak lyx.1 core \ - config.status config.cache config.log \ +include $(top_srcdir)/config/include.am + +DISTCLEANFILES += lyx.1 config.status config.cache config.log \ $(srcdir)/sourcedoc/*~ \ $(srcdir)/sourcedoc/html $(srcdir)/sourcedoc/latex \ $(srcdir)/sourcedoc/man $(srcdir)/sourcedoc/Doxyfile -MAINTAINERCLEANFILES = $(srcdir)/Makefile.in $(srcdir)/aclocal.m4 \ + +MAINTAINERCLEANFILES += $(srcdir)/aclocal.m4 \ $(srcdir)/configure $(srcdir)/development/lyx.spec \ $(srcdir)/acinclude.m4 $(srcdir)/sourcedoc/Doxyfile + SUBDIRS = intl po sigc++ boost src lib EXTRA_DIST = ANNOUNCE INSTALL.OS2 INSTALL.autoconf README.OS2 \ UPGRADING lyx.man acconfig.h autogen.sh \ config development images sourcedoc -#ETAGS_ARGS = --language-force=c++ man_MANS = lyx.1 ## Needed by bindist @@ -57,10 +58,6 @@ bindist: doxydoc: sourcedoc/Doxyfile cd sourcedoc; \ doxygen ./Doxyfile - -#sourcedoc/Doxyfile: sourcedoc/Doxyfile.in config.status -# cd $(top_builddir) \ -# CONFIG_FILES=$@ CONFIG_HEADERS= $(SHELL) ./config.status lgbtags: etags --totals=yes --recurse=yes -o TAGS $(top_srcdir)/* Index: config/include.am === RCS file: config/include.am diff -N config/include.am --- /dev/null 1 Jan 1970 00:00:00 - +++ config/include.am 13 Mar 2002 13:38:48 - @@ -0,0 +1,10 @@ +AUTOMAKE_OPTIONS = foreign + +DISTCLEANFILES= -r *.orig *.rej *~ *.bak core + +MAINTAINERCLEANFILES = $(srcdir)/Makefile.in + +ETAGS_ARGS = --lang=c++ + +BOOST_INCLUDES = -I$(top_srcdir)/boost +SIGC_INCLUDES = -I$(top_srcdir) -I$(top_builddir) Index: src/Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/Makefile.am,v retrieving revision 1.113 diff -u -p -r1.113 Makefile.am --- src/Makefile.am 12 Mar 2002 17:15:44 - 1.113 +++ src/Makefile.am 13 Mar 2002 13:38:49 - @@ -1,7 +1,11 @@ -AUTOMAKE_OPTIONS = foreign +include $(top_srcdir)/config/include.am + +DISTCLEANFILES += config.h libintl.h + +MAINTAINERCLEANFILES += $(srcdir)/config.h.in + SUBDIRS = mathed insets graphics support frontends -DISTCLEANFILES= *.orig *.rej *~ *.bak core libintl.h config.h -MAINTAINERCLEANFILES = $(srcdir)/Makefile.in $(srcdir)/config.h.in + bin_PROGRAMS = lyx lyx_DEPENDENCIES = mathed/libmathed.o insets/libinsets.o \ @@ -22,9 +26,7 @@ EXTRA_DIST = config.h.in stamp-h.in chea Variables.h \ counters.C \ counters.h - -ETAGS_ARGS = --lang=c++ -BOOST_INCLUDES = -I$(top_srcdir)/boost + INCLUDES = $(SIGC_CFLAGS) $(BOOST_INCLUDES) $(PSPELL_INCLUDES) localedir = $(datadir)/locale BUILT_SOURCES = version.C @@ -66,7 +68,6 @@ lyx_SOURCES = \ LyXAction.h \ LyXView.C \ LyXView.h \ - Makefile.in \ MenuBackend.C \ MenuBackend.h \ Painter.C \ Index: src/frontends/Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/frontends/Makefile.am,v retrieving revision 1.26 diff -u -p -r1.26 Makefile.am --- src/frontends/Makefile.am 1 Mar 2002 15:08:12 - 1.26 +++ src/frontends/Makefile.am 13 Mar 2002 13:38:49 - @@ -1,15 +1,10 @@ -AUTOMAKE_OPTIONS = foreign 1.4 -DISTCLEANFILES= *.orig *.rej *~ *.bak core -MAINTAINERCLEANFILES = $(srcdir)/Makefile.in -# xforms has to be fixed here as we still need object files in this directory -# for the compilation of other frontends, so until that we have to let this -# here. +include $(top_srcdir)/config/include.am + SUBDIRS = controllers @FRONTEND@ -ETAGS_ARGS = --lang=c++ -BOOST_INCLUDES = -I$(top_srcdir)/boost -INCLUDES = $(FRONTEND_INCLUDES) -I$(srcdir)/.. $(SIGC_CFLAGS) $(BOOST_INCLUDES) -LIBS = + noinst_LTLIBRARIES = libfrontends.la + +INCLUDES = $(FRONTEND_INCLUDES) -I$(srcdir)/.. $(SIGC_CFLAGS) $(BOOST_INCLUDES) libfrontends_la_LIBADD= @FRONTEND_GUILIB@ \ controllers/*.lo Index: src/frontends/xforms/Makefile.am === RCS file:
Re: [Devel] Bug list update again
Angus Leeming [EMAIL PROTECTED] writes: | On Wednesday 13 March 2002 1:19 pm, Herbert Voss wrote: again: the patch makes some people happy who like more wysiwyg. And the argument of JMarc that a cryptic key is more useful than the real output, is not an argument! that is all! When you rewrite the citation stuff, you can delete what you want, what has this to do with this patch? | It's a nasty fudge. Ask Lars what he thinks of nasty fudges. fudge? Isn't that some kind of dessert? -- Lgb
