On Sun, Mar 21, 2004 at 01:27:21PM -0800, Kayvan A. Sylvan spake thusly:
I have a lyx document with a branch called titlepage (for when I want
to turn a custom titlepage off or on).
When LyX reads it, I see:
no text in updateScrollbar
Buffer::Buffer()
Martin Vermeer wrote:
On Sun, Mar 21, 2004 at 01:27:21PM -0800, Kayvan A. Sylvan spake
thusly:
I have a lyx document with a branch called titlepage (for when I
want to turn a custom titlepage off or on).
When LyX reads it, I see:
no text in updateScrollbar
Buffer::Buffer()
John Levon wrote:
Or should the code in child_handler be truly trivial:
Yes.
John, I take it that you're (almost) happy with the existing signal
handling code in error_handler below. To my mind it needs only a
handling_error flag to ensure that it handles multiple signals
safely. (The
Angus Leeming [EMAIL PROTECTED] writes:
| John Levon wrote:
Or should the code in child_handler be truly trivial:
Yes.
| John, I take it that you're (almost) happy with the existing signal
| handling code in error_handler below. To my mind it needs only a
| handling_error flag to ensure that
Lars Gullik Bjønnes wrote:
Angus Leeming [EMAIL PROTECTED] writes:
| John Levon wrote:
Or should the code in child_handler be truly trivial:
Yes.
| John, I take it that you're (almost) happy with the existing
| signal handling code in error_handler below. To my mind it needs
| only a
Kayvan A. Sylvan wrote:
I don't know when this bug popped up, but I can no longer select
with the mouse (click-hold-drag) although the SHIFT-move method of
text selection seems to work.
The problem is that the relevant code has been commented out.
Unfortunately, it doesn't compile, so the
Alfredo Braunstein wrote:
There are two possibilities: either (snapshot6) we allow text after
the inset button:
blah blah blah blah [foot] blah blah
++
||
||
On Mon, Mar 22, 2004 at 09:28:01AM +, Angus Leeming wrote:
John Levon wrote:
Or should the code in child_handler be truly trivial:
Yes.
John, I take it that you're (almost) happy with the existing signal
handling code in error_handler below. To my mind it needs only a
Angus Leeming wrote:
Cute, but doesn't scale well if you have multiple insets on one line.
True.
Otherwise (snapshot7) ,
Yes please!
There are still a number of problems, mainly due to the fact that the inset
doesn't have his own row but just vertically enlarges the row it's in:
1) it
Don't appear to be used anymore.
$ grep -r noPop ../src
../src/cursor.h:void noPop();
$ grep -r nopop_ ../src
../src/cursor.C:nopop_ = false;
../src/cursor.C:if (nopop_ || !disp_.dispatched())
../src/cursor.h:bool nopop_;
Ok to remove?
--
Angus
Alfredo == Alfredo Braunstein [EMAIL PROTECTED] writes:
Alfredo which is more compact (and simpler to code), but it's a bit
Alfredo disconcerting that text that comes before (the footnote)
Alfredo shown actually after (weell, it not so uncommon for
Alfredo footnotes). This could be mitigated by
Gets rid of the unused nopop_ variable and rewrites Cursor::dispatch
to:
1. Get rid of the horrible 'operator=(safe)' call;
2. Make clear that the only member variable updated is Cursor::disp_:
DispatchResult LCursor::dispatch(FuncRequest const cmd)
{
LCursor temp_cursor(*this);
On Mon, Mar 22, 2004 at 10:37:30AM +, Angus Leeming wrote:
Not really happy actually. If we get a SIGPIPE and a SIGSEGV, we'd
lose the SIGSEGV afaics.
Hence my statement that we should remove the SIGPIPE. Ie, SIGPIPE
signals should not be handled by error_handler. In fact we don't
John Levon wrote:
Actually I've changed my mind, your code (modulo SIGPIPE issue) is
fine, with one change: we should do the emergencyCleanup *first* so
if the lyxerr *does* have problems, then we've at least made an
effort to save the files ASAP.
Makes sense.
I think it's fine to do such
On Mon, Mar 22, 2004 at 12:36:48PM +, Angus Leeming wrote:
Let's be careful here. Could you check the attached patch? If you're
happy, I'll commit it.
Looks magic
john
--
Spammers get STABBED by GOD. - Ron Echeverri
This is the patch, for if someone wants to test it and tell me how it feels
like.
Alfredo
? ChangeLog-old
? PosIterator.C-save
? PosIterator.h-save
? bfri.C
? safe
? textcursor.C-save
? textcursor.h-save
? insets/insetcollapsable-save.C
? insets/insettext-save.C
Index: cursor.C
Alfredo Braunstein wrote:
This is the patch, for if someone wants to test it and tell me how
it feels like.
Do things like math insets in display mode also return true for
breakAfter?
--
Angus
Angus Leeming wrote:
Alfredo Braunstein wrote:
This is the patch, for if someone wants to test it and tell me how
it feels like.
Do things like math insets in display mode also return true for
breakAfter?
Right now these two go in parallel.
display = break after + break before +
Alfredo Braunstein wrote:
[which reminds me: maybe it's time to talk about DEPM again?
[*ducking*]
I stayed out of the conversation last time(s) so suggest that you
start...
--
Angus
I've got an idéa of a new feature in the gui.
As it is now it doesn't mather how many enter
you press, you still cant get more then a new
part. Thats good but what if it transformed to
a vertical space if you pressed one more enter.
And if you pressed once more you got an vertical
fill.
Just a
Martin Forsgren wrote:
Hi, Martin.
I've got an idéa of a new feature in the gui.
As it is now it doesn't mather how many enter
you press, you still cant get more then a new
part. Thats good but what if it transformed to
a vertical space if you pressed one more enter.
And if you pressed
Martin Forsgren wrote:
I've got an ida of a new feature in the gui.
As it is now it doesn't mather how many enter
you press, you still cant get more then a new
part. Thats good but what if it transformed to
a vertical space if you pressed one more enter.
And if you pressed once more you
Martin == Martin Forsgren [EMAIL PROTECTED] writes:
Martin I've got an idéa of a new feature in the gui. As it is now it
Martin doesn't mather how many enter you press, you still cant get
Martin more then a new part. Thats good but what if it transformed to
Martin a vertical space if you
The xforms event loop in my tree now looks like:
while (!finished) {
ForkedcallsController::get().handleCompletedProcesses();
if (fl_check_forms() == FL_EVENT) {
...
}
}
where handleCompletedProcesses() simply checks whether a SIGCHLD
signal has been
On Mon, 22 Mar 2004, Angus Leeming wrote:
Things have changed in LyX 1.4.x. You'd use
Insert-Special Formatting-Vertical Space
which can be bound to the accelerator key of your choice.
Does any of the GUI (Qt, Xforms) allow you to define shortcuts that takes
the time into consideration, e.g
The following patches add proper detection of gsview and acroread on
windows (bug 1508).
Additionally, the 1.4.0 version gets rid of the \ps_command lyxrc
variable, which has not been used since the demise of the figinset.
I will apply them shortly.
JMarc
Index: lib/ChangeLog
On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake thusly:
Should this be made more clearly a warning or info message? Like
... Color \ lyxname '' added to color table endl ?
Yes. That is exactly what the code is doing, after all. Moreover, I
think that the warning
Martin Vermeer wrote:
On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake
thusly:
Should this be made more clearly a warning or info message? Like
... Color \ lyxname '' added to color table endl ?
Yes. That is exactly what the code is doing, after all. Moreover, I
Angus == Angus Leeming [EMAIL PROTECTED] writes:
Angus Martin Vermeer wrote:
On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake
thusly:
Should this be made more clearly a warning or info message?
Like ... Color \ lyxname '' added to color table
endl ?
Yes. That is
The xforms event loop in my tree now looks like:
while (!finished) {
ForkedcallsController::get().handleCompletedProcesses();
if (fl_check_forms() == FL_EVENT) {
...
}
}
where handleCompletedProcesses() simply checks whether a SIGCHLD
signal
Kayvan == Kayvan A Sylvan [EMAIL PROTECTED] writes:
Kayvan Here is a cleanup fix
Applied.
JMarc
Kayvan == Kayvan A Sylvan [EMAIL PROTECTED] writes:
Kayvan I want to Insert-File-ASCII as Lines, but this option is
Kayvan grayed out along with LyX Document and Ascii as
Kayvan paragraphs.
This is cured by the following patch.
Andre', this patch also cures the handling of read-only buffers. I
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg LaTeXFeatures::buffer_ is in fact the master buffer, because it
Georg is a reference that is set in the constructor.
Yes.
Georg The problem now is that loadIfNeeded() is called with that
Georg buffer, but it really needs the buffer of the
Andre == Andre Poenitz [EMAIL PROTECTED] writes:
Andre When splitting the initial paragraph in UserGuide20 (i.e. 20
Andre copies of the UserGuide concatenated) we spend 76% of the time
Andre in
Andre [17] 76.0 0.06 30.41 1659902
Andre ShareContainerParameterStruct::get(ParameterStruct const)
Jean-Marc Lasgouttes wrote:
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg The problem now is that loadIfNeeded() is called with that
Georg buffer, but it really needs the buffer of the file that
Georg contains the inset.
Why? From the code, I do not see what problem it can cause.
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg No, I did not explain it in detail. The problem is that
Georg includedFilename(), that gets called by loadIfNeeded() with the
Georg buffer that originally came from LaTeXFeatures, returns in some
Georg cases a wrong filename for relative
Jean-Marc Lasgouttes wrote:
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg No, I did not explain it in detail. The problem is that
Georg includedFilename(), that gets called by loadIfNeeded() with the
Georg buffer that originally came from LaTeXFeatures, returns in some
Georg cases a
Kuba Ober wrote:
The xforms event loop in my tree now looks like:
while (!finished) {
ForkedcallsController::get().handleCompletedProcesses();
if (fl_check_forms() == FL_EVENT) {
...
}
}
where handleCompletedProcesses() simply checks whether a
On Mon, Mar 22, 2004 at 06:54:15PM +, Angus Leeming wrote:
It turns out that QEventLoop exists only for Qt = 3.1. I'm running
RH8 here (Qt 3.0.5). For Qt 3.1, the trick seems to be to use a
QTimer with 0 timeout. That *does* work.
Excuse me if I'm misunderstanding, is this going to
John Levon wrote:
On Mon, Mar 22, 2004 at 06:54:15PM +, Angus Leeming wrote:
It turns out that QEventLoop exists only for Qt = 3.1. I'm running
RH8 here (Qt 3.0.5). For Qt 3.1, the trick seems to be to use a
QTimer with 0 timeout. That *does* work.
Excuse me if I'm misunderstanding,
Ok, this code should be able to handle multiple simultaneaous SIGCHLD
signals safely and no cycles are wasted in a busy GUI loop. In fact
things will just work automagically even if there is *no* GUI loop.
Of course, Win32 users are stuffed ;-)
--
Angus? src/support/lyxlib-orig.C
?
On Mon, Mar 22, 2004 at 10:00:22PM +, Angus Leeming wrote:
We only need to reap children every 10 minutes or whatever
Then we go back to the timer and the whole thing has been a waste of
time?
Why do you say that? It's completely normal and expected practice.
It is trivially easy to
On Sun, Mar 21, 2004 at 01:27:21PM -0800, Kayvan A. Sylvan spake thusly:
> I have a lyx document with a branch called "titlepage" (for when I want
> to turn a custom titlepage off or on).
>
> When LyX reads it, I see:
>
> no text in updateScrollbar
> Buffer::Buffer()
>
Martin Vermeer wrote:
> On Sun, Mar 21, 2004 at 01:27:21PM -0800, Kayvan A. Sylvan spake
> thusly:
>
>> I have a lyx document with a branch called "titlepage" (for when I
>> want to turn a custom titlepage off or on).
>>
>> When LyX reads it, I see:
>>
>> no text in updateScrollbar
>>
John Levon wrote:
>> Or should the code in child_handler be truly trivial:
> Yes.
John, I take it that you're (almost) happy with the existing signal
handling code in error_handler below. To my mind it needs only a
handling_error flag to ensure that it handles multiple signals
safely. (The
Angus Leeming <[EMAIL PROTECTED]> writes:
| John Levon wrote:
>>> Or should the code in child_handler be truly trivial:
>> Yes.
>
| John, I take it that you're (almost) happy with the existing signal
| handling code in error_handler below. To my mind it needs only a
| handling_error flag to
Lars Gullik Bjønnes wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> | John Levon wrote:
Or should the code in child_handler be truly trivial:
>>> Yes.
>>
> | John, I take it that you're (almost) happy with the existing
> | signal handling code in error_handler below. To my mind it
Kayvan A. Sylvan wrote:
> I don't know when this bug popped up, but I can no longer select
> with the mouse (click-hold-drag) although the SHIFT-move method of
> text selection seems to work.
The problem is that the relevant code has been commented out.
Unfortunately, it doesn't compile, so the
Alfredo Braunstein wrote:
> There are two possibilities: either (snapshot6) we allow text after
> the inset button:
>
> blah blah blah blah [foot] blah blah
> ++
> ||
> ||
>
On Mon, Mar 22, 2004 at 09:28:01AM +, Angus Leeming wrote:
> John Levon wrote:
> >> Or should the code in child_handler be truly trivial:
> > Yes.
>
> John, I take it that you're (almost) happy with the existing signal
> handling code in error_handler below. To my mind it needs only a
>
Angus Leeming wrote:
> Cute, but doesn't scale well if you have multiple insets on one line.
True.
>> Otherwise (snapshot7) ,
>
> Yes please!
There are still a number of problems, mainly due to the fact that the inset
doesn't have his own row but just vertically enlarges the row it's in:
1)
Don't appear to be used anymore.
$ grep -r noPop ../src
../src/cursor.h:void noPop();
$ grep -r nopop_ ../src
../src/cursor.C:nopop_ = false;
../src/cursor.C:if (nopop_ || !disp_.dispatched())
../src/cursor.h:bool nopop_;
Ok to remove?
--
Angus
> "Alfredo" == Alfredo Braunstein <[EMAIL PROTECTED]> writes:
Alfredo> which is more compact (and simpler to code), but it's a bit
Alfredo> disconcerting that text that comes before (the footnote)
Alfredo> shown actually after (weell, it not so uncommon for
Alfredo> footnotes). This could be
Gets rid of the unused nopop_ variable and rewrites Cursor::dispatch
to:
1. Get rid of the horrible 'operator=(safe)' call;
2. Make clear that the only member variable updated is Cursor::disp_:
DispatchResult LCursor::dispatch(FuncRequest const & cmd)
{
LCursor temp_cursor(*this);
On Mon, Mar 22, 2004 at 10:37:30AM +, Angus Leeming wrote:
> > Not really happy actually. If we get a SIGPIPE and a SIGSEGV, we'd
> > lose the SIGSEGV afaics.
>
> Hence my statement that we should remove the SIGPIPE. Ie, SIGPIPE
> signals should not be handled by error_handler. In fact we
John Levon wrote:
> Actually I've changed my mind, your code (modulo SIGPIPE issue) is
> fine, with one change: we should do the emergencyCleanup *first* so
> if the lyxerr *does* have problems, then we've at least made an
> effort to save the files ASAP.
Makes sense.
> I think it's fine to do
On Mon, Mar 22, 2004 at 12:36:48PM +, Angus Leeming wrote:
> Let's be careful here. Could you check the attached patch? If you're
> happy, I'll commit it.
Looks magic
john
--
"Spammers get STABBED by GOD." - Ron Echeverri
This is the patch, for if someone wants to test it and tell me how it feels
like.
Alfredo
? ChangeLog-old
? PosIterator.C-save
? PosIterator.h-save
? bfri.C
? safe
? textcursor.C-save
? textcursor.h-save
? insets/insetcollapsable-save.C
? insets/insettext-save.C
Index: cursor.C
Alfredo Braunstein wrote:
> This is the patch, for if someone wants to test it and tell me how
> it feels like.
Do things like math insets in display mode also return true for
breakAfter?
--
Angus
Angus Leeming wrote:
> Alfredo Braunstein wrote:
>
>> This is the patch, for if someone wants to test it and tell me how
>> it feels like.
>
> Do things like math insets in display mode also return true for
> breakAfter?
Right now these two go in parallel.
display = break after + break
Alfredo Braunstein wrote:
> [which reminds me: maybe it's time to talk about DEPM again?
> [*ducking*]
I stayed out of the conversation last time(s) so suggest that you
start...
--
Angus
I've got an idéa of a new feature in the gui.
As it is now it doesn't mather how many enter
you press, you still cant get more then a new
part. Thats good but what if it transformed to
a vertical space if you pressed one more enter.
And if you pressed once more you got an vertical
fill.
Just a
Martin Forsgren wrote:
Hi, Martin.
> I've got an idéa of a new feature in the gui.
> As it is now it doesn't mather how many enter
> you press, you still cant get more then a new
> part. Thats good but what if it transformed to
> a vertical space if you pressed one more enter.
> And if you
Martin Forsgren wrote:
> I've got an idÃa of a new feature in the gui.
> As it is now it doesn't mather how many enter
> you press, you still cant get more then a new
> part. Thats good but what if it transformed to
> a vertical space if you pressed one more enter.
> And if you pressed once more
> "Martin" == Martin Forsgren <[EMAIL PROTECTED]> writes:
Martin> I've got an idéa of a new feature in the gui. As it is now it
Martin> doesn't mather how many enter you press, you still cant get
Martin> more then a new part. Thats good but what if it transformed to
Martin> a vertical space
The xforms event loop in my tree now looks like:
while (!finished) {
ForkedcallsController::get().handleCompletedProcesses();
if (fl_check_forms() == FL_EVENT) {
...
}
}
where handleCompletedProcesses() simply checks whether a SIGCHLD
signal has been
On Mon, 22 Mar 2004, Angus Leeming wrote:
> Things have changed in LyX 1.4.x. You'd use
> Insert->Special Formatting->Vertical Space
> which can be bound to the accelerator key of your choice.
Does any of the GUI (Qt, Xforms) allow you to define shortcuts that takes
the time into consideration,
The following patches add proper detection of gsview and acroread on
windows (bug 1508).
Additionally, the 1.4.0 version gets rid of the \ps_command lyxrc
variable, which has not been used since the demise of the figinset.
I will apply them shortly.
JMarc
Index: lib/ChangeLog
On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake thusly:
> >
> > Should this be made more clearly a warning or info message? Like
> > "... Color \"" << lyxname << '"' added to color table" << endl ?
>
> Yes. That is exactly what the code is doing, after all. Moreover, I
> think
Martin Vermeer wrote:
> On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake
> thusly:
>
>> >
>> > Should this be made more clearly a warning or info message? Like
>> > "... Color \"" << lyxname << '"' added to color table" << endl ?
>>
>> Yes. That is exactly what the code is doing,
> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes:
Angus> Martin Vermeer wrote:
>> On Mon, Mar 22, 2004 at 09:12:56AM +, Angus Leeming spake
>> thusly:
>>
>>> > > Should this be made more clearly a warning or info message?
>>> Like > "... Color \"" << lyxname << '"' added to color
> The xforms event loop in my tree now looks like:
>
> while (!finished) {
> ForkedcallsController::get().handleCompletedProcesses();
> if (fl_check_forms() == FL_EVENT) {
> ...
> }
> }
>
> where handleCompletedProcesses() simply checks whether a SIGCHLD
> "Kayvan" == Kayvan A Sylvan <[EMAIL PROTECTED]> writes:
Kayvan> Here is a cleanup fix
Applied.
JMarc
> "Kayvan" == Kayvan A Sylvan <[EMAIL PROTECTED]> writes:
Kayvan> I want to Insert->File->ASCII as Lines, but this option is
Kayvan> grayed out along with "LyX Document" and "Ascii as
Kayvan> paragraphs".
This is cured by the following patch.
Andre', this patch also cures the handling of
> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
Georg> LaTeXFeatures::buffer_ is in fact the master buffer, because it
Georg> is a reference that is set in the constructor.
Yes.
Georg> The problem now is that loadIfNeeded() is called with that
Georg> buffer, but it really needs the
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> When splitting the initial paragraph in UserGuide20 (i.e. 20
Andre> copies of the UserGuide concatenated) we spend 76% of the time
Andre> in
Andre> [17] 76.0 0.06 30.41 1659902
Andre> ShareContainer::get(ParameterStruct const&)
Jean-Marc Lasgouttes wrote:
>> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
> Georg> The problem now is that loadIfNeeded() is called with that
> Georg> buffer, but it really needs the buffer of the file that
> Georg> contains the inset.
>
> Why? From the code, I do not see what problem
> "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
Georg> No, I did not explain it in detail. The problem is that
Georg> includedFilename(), that gets called by loadIfNeeded() with the
Georg> buffer that originally came from LaTeXFeatures, returns in some
Georg> cases a wrong filename for
Jean-Marc Lasgouttes wrote:
> > "Georg" == Georg Baum <[EMAIL PROTECTED]> writes:
>
> Georg> No, I did not explain it in detail. The problem is that
> Georg> includedFilename(), that gets called by loadIfNeeded() with the
> Georg> buffer that originally came from LaTeXFeatures, returns in
Kuba Ober wrote:
>> The xforms event loop in my tree now looks like:
>>
>> while (!finished) {
>> ForkedcallsController::get().handleCompletedProcesses();
>> if (fl_check_forms() == FL_EVENT) {
>> ...
>> }
>> }
>>
>> where handleCompletedProcesses()
On Mon, Mar 22, 2004 at 06:54:15PM +, Angus Leeming wrote:
> It turns out that QEventLoop exists only for Qt >= 3.1. I'm running
> RH8 here (Qt 3.0.5). For Qt < 3.1, the trick seems to be to use a
> QTimer with 0 timeout. That *does* work.
Excuse me if I'm misunderstanding, is this going
John Levon wrote:
> On Mon, Mar 22, 2004 at 06:54:15PM +, Angus Leeming wrote:
>
>> It turns out that QEventLoop exists only for Qt >= 3.1. I'm running
>> RH8 here (Qt 3.0.5). For Qt < 3.1, the trick seems to be to use a
>> QTimer with 0 timeout. That *does* work.
>
> Excuse me if I'm
Ok, this code should be able to handle multiple simultaneaous SIGCHLD
signals safely and no cycles are wasted in a busy GUI loop. In fact
things will just work automagically even if there is *no* GUI loop.
Of course, Win32 users are stuffed ;-)
--
Angus? src/support/lyxlib-orig.C
?
On Mon, Mar 22, 2004 at 10:00:22PM +, Angus Leeming wrote:
> > We only need to reap children every 10 minutes or whatever
>
> Then we go back to the timer and the whole thing has been a waste of
> time?
Why do you say that? It's completely normal and expected practice.
> It is trivially
84 matches
Mail list logo