lyx-1.3.6cvs configure.in patch
the recent mkdir test didn't make it into configure.in. This patch is sufficient to allow compilation to complete here with configure-2.13. BTW you can add my name to the blanket_permissions.txt file as well. Allan. (ARRae)Index: config/ChangeLog === RCS file: /usr/local/lyx/cvsroot/lyx-devel/config/ChangeLog,v retrieving revision 1.86.2.40 diff -u -p -r1.86.2.40 ChangeLog --- config/ChangeLog18 Feb 2005 17:25:09 - 1.86.2.40 +++ config/ChangeLog22 Feb 2005 05:47:52 - @@ -1,3 +1,7 @@ +2005-02-22 Allan Rae [EMAIL PROTECTED] + + * configure.in: Add missing MKDIR call + 2005-02-17 Michael Schmitt [EMAIL PROTECTED] * qt.m4: detect Qt Win/Free library Index: config/configure.in === RCS file: /usr/local/lyx/cvsroot/lyx-devel/config/Attic/configure.in,v retrieving revision 1.17.2.28 diff -u -p -r1.17.2.28 configure.in --- config/configure.in 2 Feb 2005 13:19:39 - 1.17.2.28 +++ config/configure.in 22 Feb 2005 05:47:52 - @@ -264,6 +264,9 @@ LYX_CHECK_DECL(snprintf, stdio.h) LYX_CHECK_DECL(vsnprintf, stdio.h) LYX_CHECK_DECL(istreambuf_iterator, iterator) +# Check the form of mkdir() +AC_FUNC_MKDIR + dnl This is a slight hack: the tests generated by autoconf 2.52 do not dnl work correctly because of some conflict with stdlib.h with g++ 2.96 dnl We aim to remove this eventually, since we should test as much as
lyx-1.3.6cvs configure.in patch
the recent mkdir test didn't make it into configure.in. This patch is sufficient to allow compilation to complete here with configure-2.13. BTW you can add my name to the blanket_permissions.txt file as well. Allan. (ARRae)Index: config/ChangeLog === RCS file: /usr/local/lyx/cvsroot/lyx-devel/config/ChangeLog,v retrieving revision 1.86.2.40 diff -u -p -r1.86.2.40 ChangeLog --- config/ChangeLog18 Feb 2005 17:25:09 - 1.86.2.40 +++ config/ChangeLog22 Feb 2005 05:47:52 - @@ -1,3 +1,7 @@ +2005-02-22 Allan Rae <[EMAIL PROTECTED]> + + * configure.in: Add missing MKDIR call + 2005-02-17 Michael Schmitt <[EMAIL PROTECTED]> * qt.m4: detect Qt Win/Free library Index: config/configure.in === RCS file: /usr/local/lyx/cvsroot/lyx-devel/config/Attic/configure.in,v retrieving revision 1.17.2.28 diff -u -p -r1.17.2.28 configure.in --- config/configure.in 2 Feb 2005 13:19:39 - 1.17.2.28 +++ config/configure.in 22 Feb 2005 05:47:52 - @@ -264,6 +264,9 @@ LYX_CHECK_DECL(snprintf, stdio.h) LYX_CHECK_DECL(vsnprintf, stdio.h) LYX_CHECK_DECL(istreambuf_iterator, iterator) +# Check the form of mkdir() +AC_FUNC_MKDIR + dnl This is a slight hack: the tests generated by autoconf 2.52 do not dnl work correctly because of some conflict with stdlib.h with g++ 2.96 dnl We aim to remove this eventually, since we should test as much as
Re: lyx-1.3.x/src/lyxlength.C
On Thu, 4 Nov 2004, Jean-Marc Lasgouttes wrote: Jean-Marc == Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Jean-Marc Thanks for the tip. I'd rather use #include Jean-Marc support/snprintf.h since this uses our own implementation Jean-Marc if needed. Can you confirm that it works? Actually it does not work here. Since snprintf is not used anywhere else in 1.3.x, I'd rather rewrite the function using normal ostringstring stuff. When I looked at the backported patch I scratched my head in wonderment as to why it ever used snprintf. As for your original suggestion of support/snprintf.h I didn't even notice it existed and it didn't occur to me to check... I guess I've been away too long. That said I have a deadline of the end of February to submit my thesis so I will continue being absent a little while longer. Actually if someone wanted to be extra helpful they could fix the external inset problem mentioned in ANNOUNCE: - Partly fix the handling of included files. External insets are still not working correctly [bug #605]. I've found a few threads in the archive about this but it seems the more you fix the more you need to fix (at least in 1.4.0cvs). How is this work progressing? Does somebody have a complaint about the following patch? It applies to 1.3.x, easy to transpose to 1.4.0cvs. Works well here. Allan. (ARRae) P.S. Remember I'm not subscribed -- best to CC me.
Re: lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
Angus wrote: This is a gnu make-ism, isn't it? At the moment the XForms frontend builds with non-gnu makes. I don't think that we should change that. Automake generates plenty of entries with the %.a: %.b format. Take a look at your Makefiles. They appear to be libtool related mostly but they're there. What I do know is that my gnu make doesn't find the rule for creating the .C files from the .fd files unless I change it as per the patch. I tried removing the Makefile.am and getting a fresh replacement in case it was somehow corrupted. Didn't make any difference. I tried moving the rule earlier in the generated Makefile -- still doesn't work. Here's what `info make` has to say on the subject of rules: Old-Fashioned Suffix Rules == Suffix rules are the old-fashioned way of defining implicit rules for `make'. Suffix rules are obsolete because pattern rules are more general and clearer. They are supported in GNU `make' for compatibility with old makefiles. ... Here is the old-fashioned way to define the rule for compiling a C source file: .c.o: $(CC) -c $(CFLAGS) $(CPPFLAGS) -o $@ $ ... Introduction to Pattern Rules - A pattern rule contains the character `%' (exactly one of them) in the target; otherwise, it looks exactly like an ordinary rule. The target is a pattern for matching file names; the `%' matches any nonempty substring, while other characters match only themselves. ... Features of GNU `make' ** ... The following features were inspired by various other versions of `make'. In some cases it is unclear exactly which versions inspired which others. * Pattern rules using `%'. This has been implemented in several versions of `make'. We're not sure who invented it first, but it's been spread around a bit. *Note Defining and Redefining Pattern Rules: Pattern Rules. I'm rather suspicious of the bug report as nobody has complained before and, let's face it, most of our users use gnu make. Maybe nobody else has as old a toolchain as I do? Maybe they all use Kayvan's rpms? Maybe they download the distribution which has these files generated? Maybe I shouldn't have `make maintainer-clean`ed my cvs tree? Got any better ideas? My toolchain : gcc-2.95.3 make-3.79.1 autoconf-2.13 automake-1.4 libtool-1.3.5 LyX 1.3.6cvs of Wed, Oct 6, 2004 Built on Nov 4 2004, 14:34:39 Configuration Host type: i686-pc-linux-gnu Special build flags:warnings assertions xforms-image-loader C Compiler: gcc C Compiler flags: -g -O2 C++ Compiler: g++ (2.95.3) C++ Compiler flags: -g -O -Wno-non-template-friend -ftemplate-depth-30 -W -Wall Linker flags: Frontend: xforms libXpm version: 4.11 libforms version: 1.0.0 LyX binary dir: /home/rae/bin LyX files dir: /home/rae/share/lyx Do you still have no-smiley fridays? Allan. (ARRae)
Re: lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
Try this patch instead if you really want to stick to obsoleted rule types. How nobody else has had a problem is a mystery. Allan. (ARRae) Index: Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/frontends/xforms/forms/Makefile.am,v retrieving revision 1.14.2.2 diff -u -p -r1.14.2.2 Makefile.am --- Makefile.am 5 Dec 2003 11:08:11 - 1.14.2.2 +++ Makefile.am 5 Nov 2004 06:26:04 - @@ -52,7 +52,7 @@ libfdesign_la_SOURCES = $(SRCS:.fd=.C) $ # These are used to generate the .[Ch] files from the .fd ones. SHELL = /bin/sh SCRIPT = $(srcdir)/fdfix.sh -SUFFIX = .fd +SUFFIXES = .fd # Make the .Ch files dependent on the fdfix files. $(libfdesign_la_SOURCES): $(srcdir)/fdfix.sh $(srcdir)/fdfix[ch].sed
Re: lyx-1.3.x/src/lyxlength.C
On Thu, 4 Nov 2004, Jean-Marc Lasgouttes wrote: > > "Jean-Marc" == Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes: > > Jean-Marc> Thanks for the tip. I'd rather use #include > Jean-Marc> "support/snprintf.h" since this uses our own implementation > Jean-Marc> if needed. Can you confirm that it works? > > Actually it does not work here. Since snprintf is not used anywhere > else in 1.3.x, I'd rather rewrite the function using normal > ostringstring stuff. When I looked at the backported patch I scratched my head in wonderment as to why it ever used snprintf. As for your original suggestion of support/snprintf.h I didn't even notice it existed and it didn't occur to me to check... I guess I've been away too long. That said I have a deadline of the end of February to submit my thesis so I will continue being absent a little while longer. Actually if someone wanted to be extra helpful they could fix the external inset problem mentioned in ANNOUNCE: - Partly fix the handling of included files. External insets are still not working correctly [bug #605]. I've found a few threads in the archive about this but it seems the more you fix the more you need to fix (at least in 1.4.0cvs). How is this work progressing? > Does somebody have a complaint about the following patch? It applies to > 1.3.x, easy to transpose to 1.4.0cvs. Works well here. Allan. (ARRae) P.S. Remember I'm not subscribed -- best to CC me.
Re: lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
Angus wrote: > This is a gnu make-ism, isn't it? At the moment the XForms frontend > builds with non-gnu makes. I don't think that we should change that. Automake generates plenty of entries with the "%.a: %.b" format. Take a look at your Makefiles. They appear to be libtool related mostly but they're there. What I do know is that my gnu make doesn't find the rule for creating the .C files from the .fd files unless I change it as per the patch. I tried removing the Makefile.am and getting a fresh replacement in case it was somehow corrupted. Didn't make any difference. I tried moving the rule earlier in the generated Makefile -- still doesn't work. Here's what `info make` has to say on the subject of rules: Old-Fashioned Suffix Rules == "Suffix rules" are the old-fashioned way of defining implicit rules for `make'. Suffix rules are obsolete because pattern rules are more general and clearer. They are supported in GNU `make' for compatibility with old makefiles. ... Here is the old-fashioned way to define the rule for compiling a C source file: .c.o: $(CC) -c $(CFLAGS) $(CPPFLAGS) -o $@ $< ... Introduction to Pattern Rules - A pattern rule contains the character `%' (exactly one of them) in the target; otherwise, it looks exactly like an ordinary rule. The target is a pattern for matching file names; the `%' matches any nonempty substring, while other characters match only themselves. ... Features of GNU `make' ** ... The following features were inspired by various other versions of `make'. In some cases it is unclear exactly which versions inspired which others. * Pattern rules using `%'. This has been implemented in several versions of `make'. We're not sure who invented it first, but it's been spread around a bit. *Note Defining and Redefining Pattern Rules: Pattern Rules. > I'm rather suspicious of the bug report as nobody has complained > before and, let's face it, most of our users use gnu make. Maybe nobody else has as old a toolchain as I do? Maybe they all use Kayvan's rpms? Maybe they download the distribution which has these files generated? Maybe I shouldn't have `make maintainer-clean`ed my cvs tree? Got any better ideas? My toolchain : gcc-2.95.3 make-3.79.1 autoconf-2.13 automake-1.4 libtool-1.3.5 LyX 1.3.6cvs of Wed, Oct 6, 2004 Built on Nov 4 2004, 14:34:39 Configuration Host type: i686-pc-linux-gnu Special build flags:warnings assertions xforms-image-loader C Compiler: gcc C Compiler flags: -g -O2 C++ Compiler: g++ (2.95.3) C++ Compiler flags: -g -O -Wno-non-template-friend -ftemplate-depth-30 -W -Wall Linker flags: Frontend: xforms libXpm version: 4.11 libforms version: 1.0.0 LyX binary dir: /home/rae/bin LyX files dir: /home/rae/share/lyx Do you still have no-smiley fridays? Allan. (ARRae)
Re: lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
Try this patch instead if you really want to stick to obsoleted rule types. How nobody else has had a problem is a mystery. Allan. (ARRae) Index: Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/frontends/xforms/forms/Makefile.am,v retrieving revision 1.14.2.2 diff -u -p -r1.14.2.2 Makefile.am --- Makefile.am 5 Dec 2003 11:08:11 - 1.14.2.2 +++ Makefile.am 5 Nov 2004 06:26:04 - @@ -52,7 +52,7 @@ libfdesign_la_SOURCES = $(SRCS:.fd=.C) $ # These are used to generate the .[Ch] files from the .fd ones. SHELL = /bin/sh SCRIPT = $(srcdir)/fdfix.sh -SUFFIX = .fd +SUFFIXES = .fd # Make the .Ch files dependent on the fdfix files. $(libfdesign_la_SOURCES): $(srcdir)/fdfix.sh $(srcdir)/fdfix[ch].sed
lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
I'm trying to compile lyx-1.3.x cvs and found I need the following change in order to avoid errors from make-3.79.1 which say it doesn't know how to make .C files. BTW, I'm not subscribed so please CC me as I don't get much time to check the archive. Oh and I've looked at 1.4.0cvs and am very impressed although I haven't got time at present to fix bugs there. Still got a thesis to finish writing up. Just a few months to go now. Cheers, Allan. (ARRae) Index: Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/frontends/xforms/forms/Makefile.am,v retrieving revision 1.14.2.2 diff -u -p -r1.14.2.2 Makefile.am --- Makefile.am 5 Dec 2003 11:08:11 - 1.14.2.2 +++ Makefile.am 4 Nov 2004 04:02:06 - @@ -59,5 +59,5 @@ $(libfdesign_la_SOURCES): $(srcdir)/fdfi libfdesign.la: $(libfdesign_la_OBJECTS) $(libfdesign_la_DEPENDENCIES) -.fd.C: +%.C: %.fd $(SHELL) $(SCRIPT) $
lyx-1.3.x/src/lyxlength.C
The recent backported bug fix for bug 1523 breaks compilation because snprintf() is unknown. Note this is lyx-1.3.x cvs I'm talking about. I haven't checked 1.4.x because it won't compile on this machine at all. Index: lyxlength.C === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/lyxlength.C,v retrieving revision 1.26.2.1 diff -u -p -u -p -r1.26.2.1 lyxlength.C --- lyxlength.C 22 Oct 2004 09:08:55 - 1.26.2.1 +++ lyxlength.C 4 Nov 2004 05:00:19 - @@ -23,6 +23,7 @@ #include Lsstream.h #include cstdlib +#include cstdio using std::abs;
lyx-1.3.x/src/frontends/xforms/forms/Makefile.am
I'm trying to compile lyx-1.3.x cvs and found I need the following change in order to avoid errors from make-3.79.1 which say it doesn't know how to make .C files. BTW, I'm not subscribed so please CC me as I don't get much time to check the archive. Oh and I've looked at 1.4.0cvs and am very impressed although I haven't got time at present to fix bugs there. Still got a thesis to finish writing up. Just a few months to go now. Cheers, Allan. (ARRae) Index: Makefile.am === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/frontends/xforms/forms/Makefile.am,v retrieving revision 1.14.2.2 diff -u -p -r1.14.2.2 Makefile.am --- Makefile.am 5 Dec 2003 11:08:11 - 1.14.2.2 +++ Makefile.am 4 Nov 2004 04:02:06 - @@ -59,5 +59,5 @@ $(libfdesign_la_SOURCES): $(srcdir)/fdfi libfdesign.la: $(libfdesign_la_OBJECTS) $(libfdesign_la_DEPENDENCIES) -.fd.C: +%.C: %.fd $(SHELL) $(SCRIPT) $<
lyx-1.3.x/src/lyxlength.C
The recent backported bug fix for bug 1523 breaks compilation because snprintf() is unknown. Note this is lyx-1.3.x cvs I'm talking about. I haven't checked 1.4.x because it won't compile on this machine at all. Index: lyxlength.C === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/lyxlength.C,v retrieving revision 1.26.2.1 diff -u -p -u -p -r1.26.2.1 lyxlength.C --- lyxlength.C 22 Oct 2004 09:08:55 - 1.26.2.1 +++ lyxlength.C 4 Nov 2004 05:00:19 - @@ -23,6 +23,7 @@ #include "Lsstream.h" #include +#include using std::abs;
Re: Aussie status
I took a look at the lovely new web design at: http://devedge.netscape.com/ It all works well so long as you have javascript enabled -- turn off javascript and you have the same problems you're reporting about my mods. That said, the javascript they are using is a modified form of one available from a page I posted earlier in this thread. So it seems if they have to resort to it then we'll have to also. The CSS menus are cool but I won't worry about that just yet. I've got three very busy months ahead of me to get my thesis finished and submitted -- at least as busy as the last month in which I've barely kept up with this thread let alone done anything about the concerns raised in it. So I might look at committing some of simple changes --- bug fixes (formatting mostly) and spelling fixes etc.-- that I've collected for the current web site and that will at least reduce the number of differences between my mods and the main site to just those specific to layout changes. I will try to make time to add a full list of link rel=stylesheet ... / link rel=alternate stylesheet ... / to my site and see what effect if any that has. It's at leaast a step in the needed direction anyway. Allan. (ARRae) seriously contemplating unsubscribing...
Re: Aussie status
I took a look at the lovely new web design at: http://devedge.netscape.com/ It all works well so long as you have javascript enabled -- turn off javascript and you have the same problems you're reporting about my mods. That said, the javascript they are using is a modified form of one available from a page I posted earlier in this thread. So it seems if they have to resort to it then we'll have to also. The CSS menus are cool but I won't worry about that just yet. I've got three very busy months ahead of me to get my thesis finished and submitted -- at least as busy as the last month in which I've barely kept up with this thread let alone done anything about the concerns raised in it. So I might look at committing some of simple changes --- bug fixes (formatting mostly) and spelling fixes etc.-- that I've collected for the current web site and that will at least reduce the number of differences between my mods and the main site to just those specific to layout changes. I will try to make time to add a full list of to my site and see what effect if any that has. It's at leaast a step in the needed direction anyway. Allan. (ARRae) seriously contemplating unsubscribing...
Re: Aussie status
On Tue, 11 Mar 2003, John Levon wrote: On Tue, Mar 11, 2003 at 03:17:27PM +1000, Allan Rae wrote: How many other pages do you visit that let you change the website design? Several. Admittedly they usually require login :) Are they actually using a cookie then for anything other than a session variable? They are using a session (at least that is what php calls it) and that's server side. But we still need to have a browser revalidate. A new theory at the bottom... Is there some reason we're not presenting several stylesheets and letting the user choose via their browser ? So far the only browser that lets you choose stylesheets is Mozilla. I have however been meaning to add the alternative stylesheets but I'm still trying to find out how we get on with two sheets needed (one for colour and one for layout). Maybe I'll just set them and see what happens. I've tested here with Mozilla-1.1 and your cache preference setting of check when out of date and mozilla does not appear to revalidate a page when it is out of date (for example if max-age=0 is set for the page). If this isn't going to work for one of the most popular browsers used by our audience then I guess we shouldn't enable it. Well, after a bit of a look at the output of phpinfo() I now suspect that the server is effectively overriding my max-age setting. It seems the local server is forcing a max-age of 1 day on all pages including mine so I'll copy my stuff to devel.lyx.org and see what happens there. Allan. (ARRae)
Re: Aussie status
On Tue, 11 Mar 2003, John Levon wrote: > On Tue, Mar 11, 2003 at 03:17:27PM +1000, Allan Rae wrote: > > > How many other pages do you visit that let you change the website > > design? > > Several. Admittedly they usually require login :) Are they actually using a cookie then for anything other than a session variable? They are using a session (at least that is what php calls it) and that's server side. But we still need to have a browser revalidate. A new theory at the bottom... > Is there some reason we're not presenting several stylesheets and > letting the user choose via their browser ? So far the only browser that lets you choose stylesheets is Mozilla. I have however been meaning to add the alternative stylesheets but I'm still trying to find out how we get on with two sheets needed (one for colour and one for layout). Maybe I'll just set them and see what happens. > > I've tested here with Mozilla-1.1 and your cache preference setting of > > "check when out of date" and mozilla does not appear to revalidate a > > page when it is out of date (for example if "max-age=0" is set for the > > page). > > If this isn't going to work for one of the most popular browsers used by > our audience then I guess we shouldn't enable it. Well, after a bit of a look at the output of phpinfo() I now suspect that the server is effectively overriding my max-age setting. It seems the local server is forcing a max-age of 1 day on all pages including mine so I'll copy my stuff to devel.lyx.org and see what happens there. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: Can you send me a cutdown screenshot please. http://movementarian.org/pinmenu.png This should be fixed now. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: I don't have another page I've seen that has this problem How many other pages do you visit that let you change the website design? Do they have cache-control directives in the header? Like: meta http-equiv=Cache-Control content=no-cache (which doesn't mean you can't cache the page BTW) I've tested here with Mozilla-1.1 and your cache preference setting of check when out of date and mozilla does not appear to revalidate a page when it is out of date (for example if max-age=0 is set for the page). The cookies are set correctly but the validation request is never made so the cookies are never read and the page is never regenerated. I have added the above meta line to my site and that appears to be the only way to get mozilla working with this cache-preference. Where working refers to generating a revalidate request to compare what's local to what the server is providing. I'm considering just adding a note to site-preferences page that people should just reload if they don't see all pages how they want them or set their browser to recheck. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: On Sat, Mar 08, 2003 at 05:38:32PM +1000, Allan Rae wrote: What is your setting for: Edit-Preferences-advanced-cache-Compare the page in the cache to the page on the network When the page is out of date. See the other email but I can't even get Mozilla to reload the page when I set a max-age=0 for a page -- which to me means it isn't revalidating when it should be. How many other sites simply change the name of the file in the link to the style sheet/s when you change their config? I have no idea. Why can't you use what lyx.org uses now, which is a lot better ? Because it isn't better. The appended ?long=1 etc. are mishandled by some browsers who cache that URL so that when you set two customise options in a row you don't get what you'd expect -- ie. both settings in operation. And it gets worse when you then click on the first option again. The separate site_preferences.php3 is an attempt to solve that problem and give a clearer idea of what the customise options do. Again I'll ask you if you see a change in the source for the page after a change of site preference. You only need to look at the link lines in the header to see if the names of the css style sheets have changed. If they changed then the cookies worked and the browser has simply failed to redraw. They didn't work, I selected blue, clicked customise, and got a style link to simple-black. This happens reasonably intermittently. Hit reload. Although at present the changes I've made to my site should now work okay once you've refreshed what's in your cache -- you will almost certainly have to do this by hand (M-r or is C-r these days?) unless you set Mozilla to check each time you view a page -- in which case you wouldn't be seeing these problems. Click on a link, then I'm back to the style I chose. It's possible, I suppose, that it's a Mozilla bug, but I strongly doubt it. That's because Mozilla didn't revalidate. Check the cookies and they will be correct. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: > > Can you send me a cutdown screenshot please. > > http://movementarian.org/pinmenu.png This should be fixed now. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: > I don't have another page I've seen that has this problem How many other pages do you visit that let you change the website design? Do they have cache-control directives in the header? Like: (which doesn't mean you can't cache the page BTW) I've tested here with Mozilla-1.1 and your cache preference setting of "check when out of date" and mozilla does not appear to revalidate a page when it is out of date (for example if "max-age=0" is set for the page). The cookies are set correctly but the validation request is never made so the cookies are never read and the page is never regenerated. I have added the above line to my site and that appears to be the only way to get mozilla "working" with this cache-preference. Where "working" refers to generating a revalidate request to compare what's local to what the server is providing. I'm considering just adding a note to site-preferences page that people should just reload if they don't see all pages how they want them or set their browser to recheck. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: > On Sat, Mar 08, 2003 at 05:38:32PM +1000, Allan Rae wrote: > > > What is your setting for: > > > > Edit->Preferences->advanced->cache->Compare the page in the > > cache to the page on the network > > When the page is out of date. See the other email but I can't even get Mozilla to reload the page when I set a max-age=0 for a page -- which to me means it isn't revalidating when it should be. > > How many other sites simply change the name of the file in the link to > > the style sheet/s when you change their config? > > I have no idea. Why can't you use what lyx.org uses now, which is a lot > better ? Because it isn't better. The appended "?long=1" etc. are mishandled by some browsers who cache that URL so that when you set two customise options in a row you don't get what you'd expect -- ie. both settings in operation. And it gets worse when you then click on the first option again. The separate site_preferences.php3 is an attempt to solve that problem and give a clearer idea of what the customise options do. > > Again I'll ask you if you see a change in the source for the page > > after a change of site preference. You only need to look at the link > > lines in the header to see if the names of the css style sheets have > > changed. If they changed then the cookies worked and the browser has > > simply failed to redraw. > > They didn't work, I selected blue, clicked customise, and got a style > link to simple-black. This happens reasonably intermittently. Hit reload. Although at present the changes I've made to my site should now work okay once you've refreshed what's in your cache -- you will almost certainly have to do this by hand (M-r or is C-r these days?) unless you set Mozilla to check each time you view a page -- in which case you wouldn't be seeing these problems. > Click on a link, then I'm back to the style I chose. It's possible, I > suppose, that it's a Mozilla bug, but I strongly doubt it. That's because Mozilla didn't revalidate. Check the cookies and they will be correct. Allan. (ARRae)
Re: Aussie status
On Fri, 7 Mar 2003, Alfredo Braunstein wrote: Jean-Marc Lasgouttes wrote: Well I tried to use 'customize' to get away from your `fancy' black theme, but it seems that there is no escape. How is it supposed to work? JMarc The only 'themes' that seem to work are the 'simple blue' and the black ones. As I said at the time: only the simple blue and two black ones exist. The others are planned but I haven't had time to look at the stuff for the last month. Allan. (ARRae)
Re: Aussie status
On Fri, 7 Mar 2003, Garst R. Reese wrote: For me the Navigate menu is behind the sponsor thank you, and does not work. Which browser and version thereof? If you are using netscape 4 there doesn't seem to be much I do about this other than limit to one column. Allan. (ARRae)
Re: toolbars
On Fri, 7 Mar 2003, John Levon wrote: On Fri, Mar 07, 2003 at 08:23:01PM +0100, Alfredo Braunstein wrote: It could also become cyclic. It used to be (which explains why there was only one icon there in the first place) but that's really not good UI (see the list archives) UI smUI. redraw the arrow as a cycle? /--\ | | \--/ (hopefully a little prettier than this nightmare) We could try to setup some kind of contest. yep ... or just see if we can borrow some people from the kde or gnome art teams Better yet, just steal the icons they've already drawn -- with an appropriate reference in CREDITS and maybe README or something -- assuming the icons are also distributed as GPL/LGPL with their corresponding sources. People will recognise the icons and say oh look, a file-opening icon etc. and by association will know what most of the toolbar icons do. And we get sexy high colour icons for free! Well most of them anyway. We just need a few that aren't already covered by existing apps icon sets. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, Garst R. Reese wrote: You guessed it. Mozilla is too bloated for my box = slow. netscape 6,7+ too ugly for words. You try Opera6 and I'll try to make time to get my website update netscape4 friendly. Allan. (ARRae)
Re: Aussie status
I've made a few small changes to the CSS on my site so that the italic headings are gone and run all combos through the CSS validator -- I get colour warnings (color and no background/background-color set and such) that I'll tweak away sometime. I have only validated the XHTML of a few pages (mainly concentrated on my additions) but it looks good -- validation-wise anyway ;-) I also changed the default to simple-blue and removed the non-existant options from site_preferences.php3. Which reminds me that we really should change file extensions to .php from .php3 for our stuff as it works with php3 or php4. I'm pretty sure I haven't used anything specifically php4 but the name change would be good. I also have to either replace the feedback form or tweak it with some CSS. On 7 Mar 2003, Jean-Marc Lasgouttes wrote: Allanhttp://www.itee.uq.edu.au/~rae/www-user/ Well I tried to use 'customize' to get away from your `fancy' black theme, but it seems that there is no escape. How is it supposed to work? Try again. Have you got cookies enabled? Does your browser check for changes everytime you visit a page? What browser and version? Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: On Sat, Mar 08, 2003 at 03:40:47PM +1000, Allan Rae wrote: redraw the arrow as a cycle? /--\ | | \--/ (hopefully a little prettier than this nightmare) Step back a minute. Do you seriously, honestly, think anybody is going to know what on earth that means ? Most Westernised folks at least will be familiar with the recycle triangle. It's a small jump to make a similar association here. If you prefer something more like: \ | -/ could be used which implies that pushing it will move something to the right as far at it will go and then it will come back again. Here's an idea. What about a springy arrow? Most people would know a spring will stretch and return. Perhaps this would be a better analogy? Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: On Sat, Mar 08, 2003 at 04:36:33PM +1000, Allan Rae wrote: I also changed the default to simple-blue and removed the non-existant options from site_preferences.php3. Yay. The pinmenu extends out of the left of the background on my machine. Laying out CSS with a position: fixed div is hell. out of the left? The navbar? how many columns? (two or three?) What does it overlap onto? This really doesn't make sense as the background-color of the body is the bg of the navbar. Hmmm... I've just added background: transparent; to div.navbar and div.sponsor for simple-blue.css. This was one of the colour warnings I mentioned. Which reminds me that we really should change file extensions to .php from .php3 for our stuff as it works with php3 or php4. I'm pretty sure I haven't used anything specifically php4 but the name change would be good. What we really really really should do is make all the files be called index.whatever and use directories for the URLs. But I guess that's too much work, and getting PHP to behave with this and not give 404s is not fun at all. See oprofile.sf.net for an example of what I mean. Well I had other ideas based on making php work for us. It would work much like the search facility on php.net which allows as you suggest and almost any other combination anyone wants. The website is then mostly statically generated (cached actually) and a 404 handler returns the appropriate page based on what a user has requested. It's actually pretty easy to do. I also have to either replace the feedback form or tweak it with some CSS. Kill it ! It's so 1995. Yes I know nobody agreed last time I said this :) We have lists after all. Maybe a few mailto:// in appropriate places -- even in the navbar or sponsor area. The cookie-lossage happens for me too, Mozilla CVS from a couple of weeks ago, cookies fully enabled. I can't see it being cookie lossage. Take a look at the page source code. I have found a number of browsers either require a forced refresh or to have their preferences adjusted so that pages are checked for content change each time. And then there are others that load the changed pages but ignore the fact that they now say a different css style sheet should be loaded and reuse an old cached rendering instead. Mozilla-1.1 doesn't seem to suffer from this. Although it could be a preference setting I've tweaked. Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: On Sat, Mar 08, 2003 at 04:41:48PM +1000, Allan Rae wrote: Most Westernised folks at least will be familiar with the recycle triangle. It's a small jump to make a similar association here. But we cannot have just a recycle triangle, users associate it with re-something. We are not recycling, we're cycling. Okay, I'm being gentle because you have either just woken up or are still partying. Did the suggested arrow look like a triangle? If you thought so then you need to go to bed now. Do not attempt to drive. :) The association I spoke of was that people will realise that something similar is happening. A cycle is formed. A squarish cycle perhaps not a triangular green recycle. If you prefer something more like: \ | -/ And how are you going to fit that, plus something that clearly indicates a llist item, onto even a KDE-sized icon ? You just won't be able to do it clearly. The arrow can overlap the paragraph that is in motion. Here's an idea. What about a springy arrow? Most people would know a spring will stretch and return. Perhaps this would be a better analogy? I have other objections to the cycling behaviour as was, and I'm not super-excited about going therough it again - the archives at marc.theaimsgroup.com are pretty navigable. UI smUI. :) I'm going home now to work on my thesis again. Which is what I was supposed to be doing instead of fiddling with web sites. Allan. (ARRae) Have fun.
Re: Aussie status
The three-column and the two-column styles use different padding/margin methods to achieve the same thing. Does one work and not the other? In either case they work in Mozilla-1.1 and as far as I can see if Mozilla-HEAD gets this bit wrong it's a new bug in Mozilla. I have noticed that sometimes Mozilla and others have their own ideas of when they can collapse padding etc. But it seems your newest email is saying that the sponsor div's contents are spilling onto the content div. Hmmm That would mean that Mozilla thinks that 160px isn't wide enough to fit the text. Bollocks to that. Can you send me a cutdown screenshot please. And this time I really do have to get going. Won't be in email contact again until Tuesday. Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: But the thing *does* cycle !! And _your_ problem is? (...recorded in the mail archives somewhere) Allan. (ARRae) Cycling environment depth ROCKS!
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: [...] I don't have another page I've seen that has this problem Must be our rewriting of a single line that does it. What is your setting for: Edit-Preferences-advanced-cache-Compare the page in the cache to the page on the network I have every time I view the page and have no problems with Mozilla-1.1 How many other sites simply change the name of the file in the link to the style sheet/s when you change their config? Again I'll ask you if you see a change in the source for the page after a change of site preference. You only need to look at the link lines in the header to see if the names of the css style sheets have changed. If they changed then the cookies worked and the browser has simply failed to redraw. Allan. (ARRae)
Re: Aussie status
On Fri, 7 Mar 2003, Alfredo Braunstein wrote: > Jean-Marc Lasgouttes wrote: > > > Well I tried to use 'customize' to get away from your `fancy' black > > theme, but it seems that there is no escape. How is it supposed to > > work? > > > > JMarc > > The only 'themes' that seem to work are the 'simple blue' and the black > ones. As I said at the time: only the simple blue and two black ones exist. The others are planned but I haven't had time to look at the stuff for the last month. Allan. (ARRae)
Re: Aussie status
On Fri, 7 Mar 2003, Garst R. Reese wrote: > For me the Navigate menu is behind the sponsor thank you, and does not > work. Which browser and version thereof? If you are using netscape 4 there doesn't seem to be much I do about this other than limit to one column. Allan. (ARRae)
Re: toolbars
On Fri, 7 Mar 2003, John Levon wrote: > On Fri, Mar 07, 2003 at 08:23:01PM +0100, Alfredo Braunstein wrote: > > > It could also become cyclic. > > It used to be (which explains why there was only one icon there in the > first place) but that's really not good UI (see the list archives) UI smUI. redraw the arrow as a cycle? /->-\ | | \-<-/ (hopefully a little prettier than this nightmare) > > We could try to setup some kind of contest. > > yep ... or just see if we can borrow some people from the kde or gnome > art teams Better yet, just steal the icons they've already drawn -- with an appropriate reference in CREDITS and maybe README or something -- assuming the icons are also distributed as GPL/LGPL with their corresponding sources. People will recognise the icons and say "oh look, a file-opening icon" etc. and by association will know what most of the toolbar icons do. And we get sexy high colour icons for free! Well most of them anyway. We just need a few that aren't already covered by existing apps icon sets. Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, Garst R. Reese wrote: > You guessed it. Mozilla is too bloated for my box = slow. > netscape 6,7+ too ugly for words. You try Opera6 and I'll try to make time to get my website update netscape4 "friendly." Allan. (ARRae)
Re: Aussie status
I've made a few small changes to the CSS on my site so that the italic headings are gone and run all combos through the CSS validator -- I get colour warnings (color and no background/background-color set and such) that I'll tweak away sometime. I have only validated the XHTML of a few pages (mainly concentrated on my additions) but it looks good -- validation-wise anyway ;-) I also changed the default to simple-blue and removed the non-existant options from site_preferences.php3. Which reminds me that we really should change file extensions to .php from .php3 for our stuff as it works with php3 or php4. I'm pretty sure I haven't used anything specifically php4 but the name change would be good. I also have to either replace the feedback form or tweak it with some CSS. On 7 Mar 2003, Jean-Marc Lasgouttes wrote: > Allan>http://www.itee.uq.edu.au/~rae/www-user/ > > Well I tried to use 'customize' to get away from your `fancy' black > theme, but it seems that there is no escape. How is it supposed to > work? Try again. Have you got cookies enabled? Does your browser check for changes everytime you visit a page? What browser and version? Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: > On Sat, Mar 08, 2003 at 03:40:47PM +1000, Allan Rae wrote: > > > redraw the arrow as a cycle? > > > > /->-\ > > | | > > \-<-/ > > > > (hopefully a little prettier than this nightmare) > > Step back a minute. Do you seriously, honestly, think anybody is going > to know what on earth that means ? Most Westernised folks at least will be familiar with the "recycle" triangle. It's a small jump to make a similar association here. If you prefer something more like: --->-\ | <-/ could be used which implies that pushing it will move something to the right as far at it will go and then it will come back again. Here's an idea. What about a springy arrow? Most people would know a spring will stretch and return. Perhaps this would be a better analogy? Allan. (ARRae)
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: > On Sat, Mar 08, 2003 at 04:36:33PM +1000, Allan Rae wrote: > > I also changed the default to simple-blue and removed the non-existant > > options from site_preferences.php3. > > Yay. The pinmenu extends out of the left of the background on my > machine. Laying out CSS with a position: fixed div is hell. out of the left? The navbar? how many columns? (two or three?) What does it overlap onto? This really doesn't make sense as the background-color of the body is the bg of the navbar. Hmmm... I've just added background: transparent; to div.navbar and div.sponsor for simple-blue.css. This was one of the colour warnings I mentioned. > > Which reminds me that we really should change file extensions to .php > > from .php3 for our stuff as it works with php3 or php4. I'm pretty > > sure I haven't used anything specifically php4 but the name change > > would be good. > > What we really really really should do is make all the files be called > index. and use directories for the URLs. But I guess that's > too much work, and getting PHP to behave with this and not give 404s is > not fun at all. See oprofile.sf.net for an example of what I mean. Well I had other ideas based on making php work for us. It would work much like the search facility on php.net which allows as you suggest and almost any other combination anyone wants. The website is then mostly statically generated (cached actually) and a 404 handler returns the appropriate page based on what a user has requested. It's actually pretty easy to do. > > I also have to either replace the feedback form or tweak it with some > > CSS. > > Kill it ! It's so 1995. Yes I know nobody agreed last time I said this > :) We have lists after all. Maybe a few mailto:// in appropriate places -- even in the navbar or sponsor area. > The cookie-lossage happens for me too, Mozilla CVS from a couple of > weeks ago, cookies fully enabled. I can't see it being cookie lossage. Take a look at the page source code. I have found a number of browsers either require a forced refresh or to have their preferences adjusted so that pages are checked for content change each time. And then there are others that load the changed pages but ignore the fact that they now say a different css style sheet should be loaded and reuse an old cached rendering instead. Mozilla-1.1 doesn't seem to suffer from this. Although it could be a preference setting I've tweaked. Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: > On Sat, Mar 08, 2003 at 04:41:48PM +1000, Allan Rae wrote: > > > Most Westernised folks at least will be familiar with the "recycle" > > triangle. It's a small jump to make a similar association here. > > But we cannot have just a recycle triangle, users associate it with > "re-something". We are not "recycling", we're "cycling". Okay, I'm being gentle because you have either just woken up or are still partying. Did the suggested arrow look like a triangle? If you thought so then you need to go to bed now. Do not attempt to drive. :) The "association" I spoke of was that people will realise that something similar is happening. A cycle is formed. A squarish cycle perhaps not a triangular green recycle. > > If you prefer something more like: > > > > --->-\ > > | > ><-/ > > And how are you going to fit that, plus something that clearly indicates > a llist item, onto even a KDE-sized icon ? You just won't be able to do > it clearly. The arrow can overlap the paragraph that is in motion. > > > Here's an idea. What about a springy arrow? Most people would know a > > spring will stretch and return. Perhaps this would be a better > > analogy? > > I have other objections to the cycling behaviour as was, and I'm not > super-excited about going therough it again - the archives at > marc.theaimsgroup.com are pretty navigable. UI smUI. :) I'm going home now to work on my thesis again. Which is what I was supposed to be doing instead of fiddling with web sites. Allan. (ARRae) Have fun.
Re: Aussie status
The three-column and the two-column styles use different padding/margin methods to achieve the same thing. Does one work and not the other? In either case they work in Mozilla-1.1 and as far as I can see if Mozilla-HEAD gets this bit wrong it's a new bug in Mozilla. I have noticed that sometimes Mozilla and others have their own ideas of when they can collapse padding etc. But it seems your newest email is saying that the sponsor div's contents are spilling onto the content div. Hmmm That would mean that Mozilla thinks that 160px isn't wide enough to fit the text. Bollocks to that. Can you send me a cutdown screenshot please. And this time I really do have to get going. Won't be in email contact again until Tuesday. Allan. (ARRae)
Re: toolbars
On Sat, 8 Mar 2003, John Levon wrote: > But the thing *does* cycle !! And _your_ problem is? (...recorded in the mail archives somewhere) Allan. (ARRae) Cycling environment depth ROCKS!
Re: Aussie status
On Sat, 8 Mar 2003, John Levon wrote: [...] > I don't have another page I've seen that has this problem Must be our rewriting of a single line that does it. What is your setting for: Edit->Preferences->advanced->cache->Compare the page in the cache to the page on the network I have "every time I view the page" and have no problems with Mozilla-1.1 How many other sites simply change the name of the file in the link to the style sheet/s when you change their config? Again I'll ask you if you see a change in the source for the page after a change of site preference. You only need to look at the link lines in the header to see if the names of the css style sheets have changed. If they changed then the cookies worked and the browser has simply failed to redraw. Allan. (ARRae)
Re: Aussie status
On 6 Mar 2003, Lars Gullik Bjønnes wrote: I have just moved the web pages over to aussie. [...] Please report all and any problems to me. The Customise cookies don't seem to work anymore. Perhaps you have: auto_globals = false (which I admit is a good thing) but our scripts currently require to work properly. My modded website is meant for use with auto_globals = false but it needs some more updating and the CSS to be updated to reflect previous comments and then needs validating. http://www.itee.uq.edu.au/~rae/www-user/ Allan. (ARRae)
Re: Aussie status
On 6 Mar 2003, Lars Gullik Bjønnes wrote: > I have just moved the web pages over to aussie. [...] > Please report all and any problems to me. The Customise cookies don't seem to work anymore. Perhaps you have: auto_globals = false (which I admit is a good thing) but our scripts currently require to work properly. My modded website is meant for use with auto_globals = false but it needs some more updating and the CSS to be updated to reflect previous comments and then needs validating. http://www.itee.uq.edu.au/~rae/www-user/ Allan. (ARRae)
Re: collapsing LFUN_XYZ_APPLY together
On Wed, 26 Feb 2003, Andre Poenitz wrote: On Wed, Feb 26, 2003 at 04:09:37PM +, Angus Leeming wrote: As John has noted, these differnet LFUNs are sufficiently similar to make the thought of collapsing them together attractive. My question: should I use these switches or should I use if-blocks? case LFUN_INDEX_APPLY: case LFUN_REF_APPLY: { dispatch_result result = UNDISPATCHED; string name; switch (ev.action) { case LFUN_INDEX_APPLY: name = index; break; case LFUN_REF_APPLY: name = ref; break; default: break; } What about a function string getNameOfTheBeast(whatever) { if (whatever == LFUN_INDEX_APPLY) return index; if (whatever == LFUN_REF_APPLY) return ref; } ? How about a map or array lookup instead? switch (ev.action) { case LFUN_INDEX_APPLY: inset = new InsetIndex(params); break; case LFUN_REF_APPLY: inset = new InsetRef(params, *buffer_); break; default: break; } This can't go to the factory? This is it's purpose after all... Or how about auto-registration of insets with the factory -- probably a bit outlandish -- but that might allow a scripting language to define insets also. Allan. (ARRae)
Re: keybindings
On Wed, 26 Feb 2003, Angus Leeming wrote: I am clearly a brain dead zombie. Surely, you meant muttonhead? Allan. (ARRae)
Re: collapsing LFUN_XYZ_APPLY together
On Wed, 26 Feb 2003, Andre Poenitz wrote: > On Wed, Feb 26, 2003 at 04:09:37PM +, Angus Leeming wrote: > > As John has noted, these differnet LFUNs are sufficiently similar to make > > the thought of collapsing them together attractive. > > > > My question: should I use these switches or should I use if-blocks? > > > case LFUN_INDEX_APPLY: > > case LFUN_REF_APPLY: { > > dispatch_result result = UNDISPATCHED; > > > > string name; > > switch (ev.action) { > > case LFUN_INDEX_APPLY: name = "index"; break; > > case LFUN_REF_APPLY: name = "ref"; break; > > default: break; > > } > > What about a function > > string getNameOfTheBeast(whatever) > { > if (whatever == LFUN_INDEX_APPLY) > return "index"; > if (whatever == LFUN_REF_APPLY) > return "ref"; > } > > ? How about a map or array lookup instead? > > switch (ev.action) { > > case LFUN_INDEX_APPLY: > > inset = new InsetIndex(params); > > break; > > case LFUN_REF_APPLY: > > inset = new InsetRef(params, *buffer_); > > break; > > default: > > break; > > } > > This can't go to the factory? This is it's purpose after all... Or how about auto-registration of insets with the factory -- probably a bit outlandish -- but that might allow a scripting language to define insets also. Allan. (ARRae)
Re: keybindings
On Wed, 26 Feb 2003, Angus Leeming wrote: > I am clearly a brain dead zombie. Surely, you meant muttonhead? Allan. (ARRae)
Re: PATCH: inverse search.
On 25 Feb 2003, Jean-Marc Lasgouttes wrote: Joao Is there someone with a recent tetex that can confirm if the Joao option --src-specials to latex is enabled? Note that since LyX generates the LaTeX code by itself, it would not be too difficult to add src-special by ourselves, for example at each new paragraph or new line. There is a small problem here that I noticed when playing with this stuff about 2 weeks ago. If we generate the src-specials ourselves we have to provide a filename and a paragraph number. Sounds simple enough _but_ if we use LyX's paragraph id as the paragraph number we can have non-sequential numbering in the output file (since if we insert a paragraph at the start of the document it won't be numbered 0 or 1 it will n+1 where n is the number of paragraphs that have been inserted into the buffer starting at load time). This situation is okay so long as we do the reverse search in the same buffer as generated the dvi. If we keep a separate tally of the number of paragraphs written and then use the lyxserver to do the reverse lookup we must count paragraphs from the beginning of the buffer and not lookup the paragraph id. This might seem minor but it does complicate things. That said the source specials do offer a character offset entry which would allow us to setup reverse links from any inset also. Allan. (ARRae)
Re: Timeout.[Ch]
On 26 Feb 2003, Lars Gullik Bjønnes wrote: What has Matthias to do with anything? (He is not my contact at Trolltech.) he's the only name we know associated with Trolltech. Allan. (ARRae)
Re: PATCH: inverse search.
On 25 Feb 2003, Jean-Marc Lasgouttes wrote: > Joao> Is there someone with a recent tetex that can confirm if the > Joao> option --src-specials to latex is enabled? > > Note that since LyX generates the LaTeX code by itself, it would not > be too difficult to add src-special by ourselves, for example at each > new paragraph or new line. There is a small problem here that I noticed when playing with this stuff about 2 weeks ago. If we generate the src-specials ourselves we have to provide a filename and a paragraph number. Sounds simple enough _but_ if we use LyX's paragraph id as the paragraph number we can have non-sequential numbering in the output file (since if we insert a paragraph at the start of the document it won't be numbered 0 or 1 it will n+1 where n is the number of paragraphs that have been inserted into the buffer starting at load time). This situation is okay so long as we do the reverse search in the same buffer as generated the dvi. If we keep a separate tally of the number of paragraphs written and then use the lyxserver to do the reverse lookup we must count paragraphs from the beginning of the buffer and not lookup the paragraph id. This might seem minor but it does complicate things. That said the source specials do offer a character offset entry which would allow us to setup reverse links from any inset also. Allan. (ARRae)
Re: Timeout.[Ch]
On 26 Feb 2003, Lars Gullik Bjønnes wrote: > What has Matthias to do with anything? > (He is not my contact at Trolltech.) he's the only name we know associated with Trolltech. Allan. (ARRae)
Re: The image loader queue
On Sat, 22 Feb 2003, Alfredo Braunstein wrote: Allan Rae wrote: A 20 second scan of the code in LoadQueue.C suggests that push_front() and pop_front() is the reason your queue is acting like a stack. Try pushing on one end and popping off the other. Sorry Allan, I've tried to answer you by mail but with no luck. It is supposed to work like this, because we want newly added images to be loaded first (except from when loading a document, they came because of user interaction: scrolling, editing, etc). Ah. There you go then. I should have spent 30 seconds instead ;-) It seems you now have this all ironed out anyway. Allan. (ARRae)
Re: The image loader queue
On Sat, 22 Feb 2003, Alfredo Braunstein wrote: > Allan Rae wrote: > > > A 20 second scan of the code in LoadQueue.C suggests that push_front() > > and pop_front() is the reason your queue is acting like a stack. > > > > Try pushing on one end and popping off the other. > > Sorry Allan, I've tried to answer you by mail but with no luck. It is > supposed to work like this, because we want newly added images to be loaded > first (except from when loading a document, they came because of user > interaction: scrolling, editing, etc). Ah. There you go then. I should have spent 30 seconds instead ;-) It seems you now have this all ironed out anyway. Allan. (ARRae)
Re: The image loader queue
On Fri, 21 Feb 2003, Alfredo Braunstein wrote: 1) I'm still investigating how to solve the startup problem. (images get loaded in reverse order). Clues welcomed. A 20 second scan of the code in LoadQueue.C suggests that push_front() and pop_front() is the reason your queue is acting like a stack. Try pushing on one end and popping off the other. Allan. (ARRae)
Re: The image loader queue
On Fri, 21 Feb 2003, Alfredo Braunstein wrote: > 1) I'm still investigating how to solve the startup problem. (images get > loaded in reverse order). Clues welcomed. A 20 second scan of the code in LoadQueue.C suggests that push_front() and pop_front() is the reason your queue is acting like a stack. Try pushing on one end and popping off the other. Allan. (ARRae)
Re: some C++ queries
On Tue, 18 Feb 2003, Andre Poenitz wrote: On Tue, Feb 18, 2003 at 02:44:11PM +, Angus Leeming wrote: I think I have just about cleaned up the controllers code so that it is now transparent. As a result of this clean-up I have managed to replace all those boost::signals in frontends/Dialogs.h with normal C++ class methods. Good stuff. We could have done without signals from the beginning (in theory at least) but I wanted to be sure we were as flexible and independent as possible in each of the frontends. That said, I agree with André that this news is indeed good stuff. How much difference does it make to compile times? WRT the remainder of the discussion I like where the two of you are heading with your discussion and support your latest suggestions. Allan. (ARRae)
Re: some C++ queries
On Tue, 18 Feb 2003, Andre Poenitz wrote: > On Tue, Feb 18, 2003 at 02:44:11PM +, Angus Leeming wrote: > > I think I have just about cleaned up the controllers code so that it is now > > transparent. As a result of this clean-up I have managed to replace all > > those boost::signals in frontends/Dialogs.h with normal C++ class methods. > > Good stuff. We could have done without signals from the beginning (in theory at least) but I wanted to be sure we were as flexible and independent as possible in each of the frontends. That said, I agree with André that this news is indeed "good stuff". How much difference does it make to compile times? WRT the remainder of the discussion I like where the two of you are heading with your discussion and support your latest suggestions. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Fri, 14 Feb 2003, John Levon wrote: On Thu, Jan 30, 2003 at 02:15:37PM +1000, Allan Rae wrote: +2003-01-27 Allan Rae [EMAIL PROTECTED] + + * insetinclude.C (loadIfNeeded): included files might be under + VCS control so we need loadLyXFile() not readFile() for that. + Are you still planning to apply this at some point ? I guess you tested it Yes, but I haven't gotten around to it yet. Will update again and commit today. Yes, it works for me. Can't think of anything I might do to break it -- loadLyXFile() just adds VCS handling a little bit extra then calls readFile() and the existing tests in loadIfNeeded() cover all possibilities of an alert dialog in both functions. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Fri, 14 Feb 2003, Allan Rae wrote: Yes, but I haven't gotten around to it yet. Will update again and commit today. Haven't had time after all. Feel free to make the one line change yourself. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Fri, 14 Feb 2003, John Levon wrote: > On Thu, Jan 30, 2003 at 02:15:37PM +1000, Allan Rae wrote: > > > +2003-01-27 Allan Rae <[EMAIL PROTECTED]> > > + > > + * insetinclude.C (loadIfNeeded): included files might be under > > + VCS control so we need loadLyXFile() not readFile() for that. > > + > > Are you still planning to apply this at some point ? I guess you tested > it Yes, but I haven't gotten around to it yet. Will update again and commit today. Yes, it works for me. Can't think of anything I might do to break it -- loadLyXFile() just adds VCS handling a little bit extra then calls readFile() and the existing tests in loadIfNeeded() cover all possibilities of an alert dialog in both functions. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Fri, 14 Feb 2003, Allan Rae wrote: > Yes, but I haven't gotten around to it yet. Will update again and > commit today. Haven't had time after all. Feel free to make the one line change yourself. Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, Alfredo Braunstein wrote: John Levon wrote: On Tue, Feb 11, 2003 at 09:08:25AM +0100, Alfredo Braunstein wrote: Making shorcuts consistent would accelerate user interaction. In fact, I find a PITA every cleverness in choosing shorcuts. In the Layout-Document-Layout dialog tab (Qt frontend) we have for instance Options|t without any reason for not using O. Or Page style|s when P would be much more intuitive. And Float placement|p, same thing. This doesn't sound like cleverness to me, but silly choices. That's my point. Non-standard choices are bad, so let's automatize the thing. Or at least make it in a standard way. I'm surprised that noone has mentioned the _existing_ shortcut creation tool which is included in the source tree at: development/tools/scgen.pl Such tools are good for initial creation of a dialog but as others have pointed out they are likely to change allocations once a dialog has been reordered (for tab ordering purposes) or extended. Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, John Levon wrote: On Tue, Feb 11, 2003 at 11:59:28AM +0100, Lars Gullik Bj?nnes wrote: [...] | I have never been able to use that interface, and it buys us precisely | nothing, and costs a lot. Poor John, failed at Emacs but rules with vi. When are you vi guys going to get your acts together and put a vi-style command interface or are you quietly admitting defeat here also. :P Gets rid of the popup/dialog clutter. By replacing it with some clutter on the buffer instead, which : 1) is non-standard 2) cannot use the frontend's architecture 3) requires more code I don't know about you, but I do not fancy coding all the stuff like tab navigation etc. by hand. Can't Qt create a dialog as part of a canvas? That is, can't you just put everything on a scrolling-canvas instead of a separate popup dialog box? (which all the Qt/KDE apps I remember using seem to always want to make modal) Then the only difference is the stuff that is covered up or blocked by silly dialogs -- in fact it seems that since Qt can float toolbars the same could effectively be done with dialogs -- the user can choose to pull them out of the workarea and decorate the rest of their screen with them if they want to (or vice versa). Allan. (ARRae)
Re: HELP! I can't get the *%^% website to accept lgt!
On Tue, 11 Feb 2003, Angus Leeming wrote: On Tuesday 11 February 2003 6:41 pm, Jean-Marc Lasgouttes wrote: Angus == Angus Leeming [EMAIL PROTECTED] writes: Angus Why not move lgt into the top level dir? I thought about it, but before moving things around in cvs, how would you organize the menus in the sidebar? As it is now? Yes. The logical structure of the site to the user need bear no relation to its physical layout. That's up to our discretion. You should create a redirection page from the old lgt.html to the new lgt.php3 though, or we'll be guilty of the spurious generation of broken links horror. Sorry but this won't work properly because of the way we create the navbar at present. Better to move /about/lgt/ to /lgt/ I also want to rewrite the subdir handling code but haven't gotten around to it yet. I'll make it a priority next week. Allan. (ARRae)
Re: [PATCH] changes: LColor
On Tue, 11 Feb 2003, José Matos wrote: On Tuesday 11 February 2003 08:54, Angus Leeming wrote: Allan Rae wrote: What I'd like to know is how did Angus know I have purple on my desktop (as icon backgrounds and title bars as it happens) as well as black and gold? Let's just call it an educated guess based on the evidence available. AKA, a shot in the dark ;-) You hited the gold spot on the darkness. Ok, I'm starting to see a pattern here. ;-) Please tell all. I might have got a colour wrong somewhere. Allan. (ARRae)
Re: tex2lyx
On Tue, 11 Feb 2003, Angus Leeming wrote: On Tuesday 11 February 2003 3:54 pm, Andre Poenitz wrote: On Tue, Feb 11, 2003 at 03:52:31PM +, Angus Leeming wrote: André, here's a patch to get it to work on this little test case. Else it get's lost in an infinite loop in handle_opts (There are none to documentclass...). You are big enough to apply this by yourself, aren't you? ;-) Nope! There's a girl I know who thinks I regress with each passing day ;-) Better bring that wedding day forward a few months then. ;-) Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Fri, 7 Feb 2003, Andre Poenitz wrote: Currently it looks like neither Asger, Angus and John could come if we insisted June 20th. Currently I have (5 - ok, 0 - not ok): I think I'm going to be a 0 for all dates this year unless I win the lottery. I am however making good progress on my thesis -- but I have a cunning plan to get me a new distraction this Friday. Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Wed, 12 Feb 2003, Andre Poenitz wrote: On Wed, Feb 12, 2003 at 02:14:32PM +1000, Allan Rae wrote: I think I'm going to be a 0 for all dates this year unless I win the lottery. Is that the price of spending the whole life upside down? No the price for spending my life upside down is that it takes about 7 hours of flying in a 747 to leave the Australian land-mass behind and be flying over water (toward Singapore or Hong Kong) -- and no, the wheels are not in contact with terra firma during those 7 hours. About 7.5hrs to Singapore or 8.5hrs to Hong Kong (I think -- never been that way before). Then 12.5hrs Singapore to Milan or maybe Munich a little longer to Frankfurt. Actually the main reason for the 0 rating is that I've quit my job and am surviving on what little savings I have in an effort to get this thesis finished without too many distractions. I wonder whether RyanAir will go beyond Europe some time. Yesterday I've seen Frankfurt-Glasgow for 7 EUR... How much extra for a seat? And how much on top of that for a seat belt? And how much more on top of that for a pressurised cabin? Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Wed, 12 Feb 2003, Andre Poenitz wrote: My last trip a few weeks ago was Berlin-London for ~50 EUR (one way, including taxes) with a company called buzz and it was ok. No food on board (well, there was, but you would have to pay extra) but I really don't care about that for a two hours journey... That's about what we pay here for a similar length domestic flight (between capital cities -- Brisbane to Melbourne (usually 2hrs in a 767)). That's flying about half the length of the eastern seaboard of Australia -- just so you get some feel for the scale. Flying anywhere other than between capital cities the prices get much more hideous. Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, Alfredo Braunstein wrote: > John Levon wrote: > > > On Tue, Feb 11, 2003 at 09:08:25AM +0100, Alfredo Braunstein wrote: > > > >> Making shorcuts consistent would accelerate user interaction. In fact, I > >> find a PITA every cleverness in choosing shorcuts. In the > >> Layout->Document->Layout dialog tab (Qt frontend) we have for instance > >> "Options|t" without any reason for not using O. Or "Page style|s" when P > >> would be much more intuitive. And "Float placement|p", same thing. > > > > This doesn't sound like cleverness to me, but silly choices. > > That's my point. Non-standard choices are bad, so let's automatize the > thing. Or at least make it in a standard way. I'm surprised that noone has mentioned the _existing_ shortcut creation tool which is included in the source tree at: development/tools/scgen.pl Such tools are good for initial creation of a dialog but as others have pointed out they are likely to change allocations once a dialog has been reordered (for tab ordering purposes) or extended. Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, John Levon wrote: > On Tue, Feb 11, 2003 at 11:59:28AM +0100, Lars Gullik Bj?nnes wrote: [...] > > | I have never been able to use that interface, and it buys us precisely > > | nothing, and costs a lot. Poor John, failed at Emacs but rules with vi. When are you vi guys going to get your acts together and put a vi-style command interface or are you quietly admitting defeat here also. :P > > Gets rid of the popup/dialog clutter. > > By replacing it with some clutter on the buffer instead, which : > > 1) is non-standard > 2) cannot use the frontend's architecture > 3) requires more code > > I don't know about you, but I do not fancy coding all the stuff like tab > navigation etc. by hand. Can't Qt create a "dialog" as part of a canvas? That is, can't you just put everything on a scrolling-canvas instead of a separate popup dialog box? (which all the Qt/KDE apps I remember using seem to always want to make modal) Then the only difference is the stuff that is covered up or blocked by silly dialogs -- in fact it seems that since Qt can float toolbars the same could effectively be done with dialogs -- the user can choose to pull them out of the workarea and decorate the rest of their screen with them if they want to (or vice versa). Allan. (ARRae)
Re: HELP! I can't get the *%&^%& website to accept lgt!
On Tue, 11 Feb 2003, Angus Leeming wrote: > On Tuesday 11 February 2003 6:41 pm, Jean-Marc Lasgouttes wrote: > > > "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes: > > > > Angus> Why not move lgt into the top level dir? > > > > I thought about it, but before moving things around in cvs, how would > > you organize the menus in the sidebar? As it is now? > > Yes. The logical structure of the site to the user need bear no relation to > its physical layout. That's up to our discretion. You should create a > redirection page from the old lgt.html to the new lgt.php3 though, or we'll > be guilty of the spurious generation of broken links horror. Sorry but this won't work properly because of the way we create the navbar at present. Better to move /about/lgt/ to /lgt/ I also want to rewrite the subdir handling code but haven't gotten around to it yet. I'll make it a priority next week. Allan. (ARRae)
Re: [PATCH] changes: LColor
On Tue, 11 Feb 2003, José Matos wrote: > On Tuesday 11 February 2003 08:54, Angus Leeming wrote: > > Allan Rae wrote: > > > What I'd like to know is how did Angus know I have purple on my > > > desktop (as icon backgrounds and title bars as it happens) as well as > > > black and gold? > > > > Let's just call it an educated guess based on the evidence available. AKA, > > a shot in the dark ;-) > > You hited the gold spot on the darkness. Ok, I'm starting to see a pattern > here. ;-) Please tell all. I might have got a colour wrong somewhere. Allan. (ARRae)
Re: tex2lyx
On Tue, 11 Feb 2003, Angus Leeming wrote: > On Tuesday 11 February 2003 3:54 pm, Andre Poenitz wrote: > > On Tue, Feb 11, 2003 at 03:52:31PM +, Angus Leeming wrote: > > > André, here's a patch to get it to work on this little test case. Else it > > > get's lost in an infinite loop in handle_opts (There are none to > > > documentclass...). > > > > You are big enough to apply this by yourself, aren't you? ;-) > > Nope! There's a girl I know who thinks I regress with each passing day ;-) Better bring that wedding day forward a few months then. ;-) Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Fri, 7 Feb 2003, Andre Poenitz wrote: > Currently it looks like neither Asger, Angus and John could come if we > insisted June 20th. > > Currently I have (5 - "ok", 0 - "not ok"): I think I'm going to be a 0 for all dates this year unless I win the lottery. I am however making good progress on my thesis -- but I have a cunning plan to get me a new distraction this Friday. Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Wed, 12 Feb 2003, Andre Poenitz wrote: > On Wed, Feb 12, 2003 at 02:14:32PM +1000, Allan Rae wrote: > > I think I'm going to be a 0 for all dates this year unless I win the > > lottery. > > Is that the price of spending the whole life upside down? No the price for spending my life upside down is that it takes about 7 hours of flying in a 747 to leave the Australian land-mass behind and be flying over water (toward Singapore or Hong Kong) -- and no, the wheels are not in contact with terra firma during those 7 hours. About 7.5hrs to Singapore or 8.5hrs to Hong Kong (I think -- never been that way before). Then 12.5hrs Singapore to Milan or maybe Munich a little longer to Frankfurt. Actually the main reason for the "0" rating is that I've quit my job and am surviving on what little savings I have in an effort to get this thesis finished without too many distractions. > I wonder whether RyanAir will go beyond Europe some time. Yesterday I've > seen Frankfurt-Glasgow for 7 EUR... How much extra for a seat? And how much on top of that for a seat belt? And how much more on top of that for a pressurised cabin? Allan. (ARRae)
Re: Angus, Asger, Jean-Marc...! (was: Re: [patch]: pass $...$ and$$...$$ through reLyX unchanged)
On Wed, 12 Feb 2003, Andre Poenitz wrote: > My last trip a few weeks ago was Berlin-London for ~50 EUR (one way, > including taxes) with a company called "buzz" and it was ok. No food on > board (well, there was, but you would have to pay extra) but I really don't > care about that for a two hours journey... That's about what we pay here for a similar length domestic flight (between capital cities -- Brisbane to Melbourne (usually 2hrs in a 767)). That's flying about half the length of the eastern seaboard of Australia -- just so you get some feel for the scale. Flying anywhere other than between capital cities the prices get much more hideous. Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, John Levon wrote: On Fri, Feb 07, 2003 at 12:39:32PM +1000, Allan Rae wrote: The chosen scheme doesn't seem to be properly persistent. Cookie problem have even worse problems when using the current LyX website -- try I do not have any problems with the current LyX website. My browser is mozilla trunk of a few days ago. Mozilla is one of the few browsers that doesn't have the caching problem or the redraw on change of css problem so of you don't see any problems. Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, [utf-8] José Matos wrote: On Wednesday 05 February 2003 18:12, Angus Leeming wrote: Don't worry. His default colour scheme is _really_ nasty ;-) It gives a new meaning to dark ages. ;-) So you don't like black and gold? Or at least black background with yellow writing? I use that as my default for all my xterms and for LyX. Easy to read. At least I thought so anyway... Oh well, you don't have to use it there are other options. You could even submit your own puce and tangarine colour scheme if you really prefer such a combination. Allan. (ARRae)
Re: [PATCH] changes: LColor
On Sat, 8 Feb 2003, Angus Leeming wrote: John Levon wrote: On Sat, Feb 08, 2003 at 12:28:42AM +, Angus Leeming wrote: Anyway, they're configurable aren't they, so if Martin wants gold with purple linings he can have that too. Hey, how about you leave Allan alone now ! Thank you John for sticking up for me. (Comes over all innocent) what dod I say? What I'd like to know is how did Angus know I have purple on my desktop (as icon backgrounds and title bars as it happens) as well as black and gold? Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, John Levon wrote: On Mon, Feb 10, 2003 at 11:27:22PM +0100, Lars Gullik Bj?nnes wrote: Have I said that I really do not like the dialogs at all? I'd prefere an application completely without dialogs... Wouldn't we all ? However, this is not possible. Menues, toolbar(s) and a minibuffer. What more could we need? Rendering some configuration pages in a buffer -- like (x)emacs' customization settings. Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, John Levon wrote: > On Fri, Feb 07, 2003 at 12:39:32PM +1000, Allan Rae wrote: > > > > The chosen scheme doesn't seem to be properly persistent. Cookie problem > > > > have even worse problems when using the current LyX website -- try > > I do not have any problems with the current LyX website. My browser is > mozilla trunk of a few days ago. Mozilla is one of the few browsers that doesn't have the caching problem or the redraw on change of css problem so of you don't see any problems. Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, [utf-8] José Matos wrote: > On Wednesday 05 February 2003 18:12, Angus Leeming wrote: > > > > Don't worry. His default colour scheme is _really_ nasty ;-) > > It gives a new meaning to dark ages. ;-) So you don't like black and gold? Or at least black background with yellow writing? I use that as my default for all my xterms and for LyX. Easy to read. At least I thought so anyway... Oh well, you don't have to use it there are other options. You could even submit your own puce and tangarine colour scheme if you really prefer such a combination. Allan. (ARRae)
Re: [PATCH] changes: LColor
On Sat, 8 Feb 2003, Angus Leeming wrote: > John Levon wrote: > > > On Sat, Feb 08, 2003 at 12:28:42AM +, Angus Leeming wrote: > > > >> Anyway, they're configurable aren't they, so if Martin wants gold with > >> purple linings he can have that too. > > > > Hey, how about you leave Allan alone now ! Thank you John for sticking up for me. > (Comes over all innocent) what dod I say? What I'd like to know is how did Angus know I have purple on my desktop (as icon backgrounds and title bars as it happens) as well as black and gold? Allan. (ARRae)
Re: shorcut automatization
On Tue, 11 Feb 2003, John Levon wrote: > On Mon, Feb 10, 2003 at 11:27:22PM +0100, Lars Gullik Bj?nnes wrote: > > > Have I said that I really do not like the dialogs at all? > > I'd prefere an application completely without dialogs... > > Wouldn't we all ? However, this is not possible. Menues, toolbar(s) and a minibuffer. What more could we need? Rendering some configuration pages in a buffer -- like (x)emacs' customization settings. Allan. (ARRae)
Re: I've been experimenting with the web site
On Wed, 5 Feb 2003, John Levon wrote: On Wed, Feb 05, 2003 at 05:58:31PM +, Jos? Matos wrote: http://www.itee.uq.edu.au/~rae/www-user/ Comments : The chosen scheme doesn't seem to be properly persistent. Cookie problem ? Browser problem. As best as I can see from my testing several browsers refuse to redraw the web pages when the css style sheets change¹ (I echoed the cookies and css file names to the output during testing and these definitely changed). A reload is required to force a redraw in these dodgy browsers. These were the same browsers that have even worse problems when using the current LyX website -- try changing to short/long labels then complex/simple and then one/two columns and you'll see what I mean. ¹ By change I don't just mean when I've made mods to blah.css but when blah.css is replaced by bizzo.css I assume the default you have (Bowser's castle ?) isn't going to be the default. Light blue theme instead ? Most likely. The fancy-black scheme (black marble and gold fittings -- inspired by a very opulent bathroom) was what started me fiddling with the website (it doesn't come close to what I dreamed of BTW). As I said in my orginal email one of the simple schemes should be the default -- at this time I'm inclinded to go with the blue one with no italics (ie. the same simple blue scheme currently available on the main site). BTW, in answer to Angus' stinging remark about the current default: I happen to like to like it a lot and you've hurt my feelings you big bully. I don't like the italic headings, it looks too busy. I'd prefer normal text for the headings (maybe keep italic for the dates in news). As I said the headings were experiments on the day I announced it to you. I think the LHS menu should be reworked. That's one of the things remaining on my todo list and one of the reasons I wasn't really ready to unleash it yet. Insult someone today, Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, Allan Rae wrote: On Wed, 5 Feb 2003, John Levon wrote: On Wed, Feb 05, 2003 at 05:58:31PM +, Jos? Matos wrote: http://www.itee.uq.edu.au/~rae/www-user/ Comments : The chosen scheme doesn't seem to be properly persistent. Cookie problem ? Browser problem. As best as I can see from my testing several browsers refuse to redraw the web pages when the css style sheets change¹ (I echoed the cookies and css file names to the output during testing and these definitely changed). A reload is required to force a redraw in these dodgy browsers. Hmmm... It seems that maybe it could be more complicated than that and the following more elaborate method might be better: http://www.alistapart.com/stories/alternate/ cookies + javascript + alternate stylesheet links. Allan. (ARRae)
Re: I've been experimenting with the web site
On Wed, 5 Feb 2003, John Levon wrote: > On Wed, Feb 05, 2003 at 05:58:31PM +, Jos? Matos wrote: > > > > http://www.itee.uq.edu.au/~rae/www-user/ > > Comments : > > The chosen scheme doesn't seem to be properly persistent. Cookie problem > ? Browser problem. As best as I can see from my testing several browsers refuse to redraw the web pages when the css style sheets change¹ (I echoed the cookies and css file names to the output during testing and these definitely changed). A reload is required to force a redraw in these dodgy browsers. These were the same browsers that have even worse problems when using the current LyX website -- try changing to short/long labels then complex/simple and then one/two columns and you'll see what I mean. ¹ By change I don't just mean when I've made mods to blah.css but when blah.css is replaced by bizzo.css > I assume the default you have (Bowser's castle ?) isn't going to be the > default. Light blue theme instead ? Most likely. The fancy-black scheme (black marble and gold fittings -- inspired by a very opulent bathroom) was what started me fiddling with the website (it doesn't come close to what I dreamed of BTW). As I said in my orginal email one of the simple schemes should be the default -- at this time I'm inclinded to go with the blue one with no italics (ie. the same simple blue scheme currently available on the main site). BTW, in answer to Angus' stinging remark about the current default: I happen to like to like it a lot and you've hurt my feelings you big bully. > I don't like the italic headings, it looks too busy. I'd prefer normal > text for the headings (maybe keep italic for the dates in news). As I said the headings were experiments on the day I announced it to you. > I think the LHS menu should be reworked. That's one of the things remaining on my todo list and one of the reasons I wasn't really ready to unleash it yet. Insult someone today, Allan. (ARRae)
Re: I've been experimenting with the web site
On Fri, 7 Feb 2003, Allan Rae wrote: > On Wed, 5 Feb 2003, John Levon wrote: > > > On Wed, Feb 05, 2003 at 05:58:31PM +, Jos? Matos wrote: > > > > > > http://www.itee.uq.edu.au/~rae/www-user/ > > > > Comments : > > > > The chosen scheme doesn't seem to be properly persistent. Cookie problem > > ? > > Browser problem. As best as I can see from my testing several > browsers refuse to redraw the web pages when the css style sheets > change¹ (I echoed the cookies and css file names to the output during > testing and these definitely changed). A reload is required to force > a redraw in these dodgy browsers. Hmmm... It seems that maybe it could be more complicated than that and the following more elaborate method might be better: http://www.alistapart.com/stories/alternate/ cookies + javascript + "alternate stylesheet" links. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Wed, 5 Feb 2003, John Levon wrote: On Wed, Feb 05, 2003 at 03:37:19PM +1000, Allan Rae wrote: Does someone want to comment on this bug fix? For example, may I commit it or not? The bug isn't in Bugzilla as far as I can see but that doesn't make it any less annoying. It's too late for 1.3.0 IMHO. We have plenty of unfixed bugs and this is a rare case I reckon. Righto. It can wait for 1.3.1 then. Allan. (ARRae)
I've been experimenting with the web site
and you can see what a mess I've made here: http://www.itee.uq.edu.au/~rae/www-user/ There is still a bit of cleaning up to do in the php code and in the css but it should all look good in links, lynx, IE, Konq and Mozilla (and its derivatives). Don't tell me it's broken in Netscape 4 because I already know this. It will remain broken (as far as 2 or 3 columns is concerned) and I'll just force one column for it but will try to get the rendering (colour + spacing) working. I'm not really ready for releasing this to you guys but with 1.3.0 around the corner (right next door?) and some of you already overhauling the LGT I figured I'd better give you a heads up on what I'm doing. The h2 and h3 borders or solid backgrounds replace the hr's that used to be used and are a result of some playing today. If you have a few minutes to spare you will find that the choice of one, two or three columns should all work nicely. Depite the fact that there are 6 colour scheme choices I only have the first three implemented at present. Don't worry, the traditional green graphic website will be included although it won't be the default -- I'd rather one of the simple colour-schemes be the default so initial contact with the site gets a mimimum payload and fast rendering. Anyway, there are a few other things I'd like to try but my main aims have been to maximise the amount of a page that is static so I can later just write the header and which css files to use and append a static copy of the generated page -- ie. do a simple caching scheme to speedup the website. I also wanted to try to avoid the cookie/cache conflicts certain browsers have with our current customization scheme -- hence the separate page with a form. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Wed, 5 Feb 2003, John Levon wrote: > On Wed, Feb 05, 2003 at 03:37:19PM +1000, Allan Rae wrote: > > > Does someone want to comment on this bug fix? For example, may I > > commit it or not? The bug isn't in Bugzilla as far as I can see > > but that doesn't make it any less annoying. > > It's too late for 1.3.0 IMHO. We have plenty of unfixed bugs and this is > a rare case I reckon. Righto. It can wait for 1.3.1 then. Allan. (ARRae)
I've been experimenting with the web site
and you can see what a mess I've made here: http://www.itee.uq.edu.au/~rae/www-user/ There is still a bit of cleaning up to do in the php code and in the css but it should all look good in links, lynx, IE, Konq and Mozilla (and its derivatives). Don't tell me it's broken in Netscape 4 because I already know this. It will remain broken (as far as 2 or 3 columns is concerned) and I'll just force one column for it but will try to get the rendering (colour + spacing) working. I'm not really ready for releasing this to you guys but with 1.3.0 around the corner (right next door?) and some of you already overhauling the LGT I figured I'd better give you a heads up on what I'm doing. The h2 and h3 borders or solid backgrounds replace the hr's that used to be used and are a result of some playing today. If you have a few minutes to spare you will find that the choice of one, two or three columns should all work nicely. Depite the fact that there are 6 colour scheme choices I only have the first three implemented at present. Don't worry, the traditional green graphic website will be included although it won't be the default -- I'd rather one of the simple colour-schemes be the default so initial contact with the site gets a mimimum payload and fast rendering. Anyway, there are a few other things I'd like to try but my main aims have been to maximise the amount of a page that is static so I can later just write the header and which css files to use and append a static copy of the generated page -- ie. do a simple caching scheme to speedup the website. I also wanted to try to avoid the cookie/cache conflicts certain browsers have with our current customization scheme -- hence the separate page with a form. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Thu, 30 Jan 2003, Allan Rae wrote: An annoying bug in multipart document support occurs when you use a VCS like RCS on the documents. Load the master, update DVI to get all the other parts loaded, now try to edit one of the auto-loaded docs. You can't because the lyxvc for that document wasn't setup. A simple one line fix is attached. I considered moving the lyxvc code from loadLyXFile() into readFile() instead but decided that would be too dramatic and probably not the right thing to do anyway. Besides, I want this change in 1.3.0 if it isn't already out the door. Does someone want to comment on this bug fix? For example, may I commit it or not? The bug isn't in Bugzilla as far as I can see but that doesn't make it any less annoying. Allan. (ARRae)
Re: [PATCH] InsetInclude and version control bug
On Thu, 30 Jan 2003, Allan Rae wrote: > An annoying bug in multipart document support occurs when you use a > VCS like RCS on the documents. Load the master, update DVI to get all > the other parts loaded, now try to edit one of the auto-loaded docs. > You can't because the lyxvc for that document wasn't setup. > > A simple one line fix is attached. > > I considered moving the lyxvc code from loadLyXFile() into readFile() > instead but decided that would be too dramatic and probably not the > right thing to do anyway. Besides, I want this change in 1.3.0 if it > isn't already out the door. Does someone want to comment on this bug fix? For example, may I commit it or not? The bug isn't in Bugzilla as far as I can see but that doesn't make it any less annoying. Allan. (ARRae)
Re: [Bug 848] IEEEtran reference should be updated
On Tue, 28 Jan 2003 [EMAIL PROTECTED] wrote: http://bugzilla.lyx.org/show_bug.cgi?id=848 I get Error 400 attempting to get to bugzilla.lyx.org [EMAIL PROTECTED] changed: What|Removed |Added CC||[EMAIL PROTECTED] Status|UNCONFIRMED |NEW everconfirmed|0 |1 --- Additional Comments From [EMAIL PROTECTED] 2003-01-28 11:16 --- It appears that we do not have any documentation for IEEEtran in Extended.lyx, but rather some documentation for latex8. So far so good. Allan, am I right that latex8 should be dead now? No. I don't think so. Latex8 is/was for conference papers. The very latest IEEEtran has a conference option but I'm not sure if all conferences are using that. Can we just use IEEEtran instead? Also I see that authors and affiliations handling require special code that lyx cannot do now, like \author(\authorblockN{John Doe}\authorblockA{address\on several\\lines}} I did a quick test with one of my old docs yesterday and found that it still formatted the same as before so I don't think it is too much to worry about yet. Although it would be nice to figure out how to support it. Finally, IEEEtran has two modes, namely conference and journal (journal is default). Should we provide a different layout for conferences? I've only have a quick glimpse of the latest code and haven't tested any of the options yet. Plenty of conference papers get turned into journal papers so it would be simpler if they were the same class -- mind you if IEEEtran(conference) consisted of just: input IEEEtran_journal extra_option conference without any environment name changes (although font etc. info might change) then it wouldn't hurt but would also be overkill methinks. Allan. (ARRae) (not subscribed yet but will be soon)
[PATCH] InsetInclude and version control bug
An annoying bug in multipart document support occurs when you use a VCS like RCS on the documents. Load the master, update DVI to get all the other parts loaded, now try to edit one of the auto-loaded docs. You can't because the lyxvc for that document wasn't setup. A simple one line fix is attached. I considered moving the lyxvc code from loadLyXFile() into readFile() instead but decided that would be too dramatic and probably not the right thing to do anyway. Besides, I want this change in 1.3.0 if it isn't already out the door. I keep getting Error 400 when trying to access bugzilla.lyx.org. Any ideas? Allan. (ARRae) Index: src/insets/ChangeLog === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/insets/ChangeLog,v retrieving revision 1.552 diff -u -p -r1.552 ChangeLog --- src/insets/ChangeLog23 Jan 2003 16:23:41 - 1.552 +++ src/insets/ChangeLog30 Jan 2003 04:05:54 - @@ -1,3 +1,8 @@ +2003-01-27 Allan Rae [EMAIL PROTECTED] + + * insetinclude.C (loadIfNeeded): included files might be under + VCS control so we need loadLyXFile() not readFile() for that. + 2003-01-20 Michael Schmitt [EMAIL PROTECTED] * insetert.C: Index: src/insets/insetinclude.C === RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/insets/insetinclude.C,v retrieving revision 1.91 diff -u -p -r1.91 insetinclude.C --- src/insets/insetinclude.C 4 Dec 2002 12:07:48 - 1.91 +++ src/insets/insetinclude.C 30 Jan 2003 04:05:54 - @@ -274,7 +274,7 @@ bool InsetInclude::loadIfNeeded() const if (!finfo.isOK()) return false; - return bufferlist.readFile(getFileName(), !finfo.writable()) != 0; + return bufferlist.loadLyXFile(getFileName(), false) != 0; }