| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
* acts as a safe-guard against uninterrupted infinite loops
or other operations that are costly to undo in interactive mode.
If we're out of memory, it is usually too late to react properly.
This implementation tries to avoid OOMs due to SciTECO behaviour.
We cannot fully exclude the chance of an OOM error.
* The undo stack size is only approximated using the
UndoToken::get_size() method.
Other ways to measure the exact amount of allocated heap
(including size fields in every heap object or using sbrk(0) and
similar) are either costly in terms of memory or platform-specific.
This implementation does not need any additional memory per heap
object or undo token but exploits the fact that undo tokens
are virtual already. The size of an undo token is determined
at compile time.
* Default memory limit of 500mb should be OK for most people.
* The current limit can be queried with "2EJ" and set with <x>,2EJ.
This also works interactively (a bit tricky!)
* Limiting can be disabled. In this case, undo token processing
is a bit faster.
* closes #3
|
|
without a corresponding loop start (<)
* assertions were introduced very early when there was no proper error handling
in SciTECO. However it points to a macro programming error instead of
a SciTECO programming error and should not crash the editor.
* Perhaps it would be best to check for this kind of "syntax" error also
in parse-only modes. This is not done currently.
* part of the solution to issue #3
|
|
|
|
|
|
* since SCI_SETDOCPOINTER resets character representations
(should probably be submitted as a bug to Scintilla)
we have to reset the representations each time we load
a q-register into the q-reg view.
* since the SCI_SETREPRESENTION call does not do any redrawing
(and it would be very slow if it did), the lines with control
characters were laid out wrong (too much spaces).
This happened when editing a q-reg or the command-line.
* Since it is not obvious how to fix Scintilla's behaviour here,
we work around the issue by temporarily disabling the layout
cache.
|
|
* also did some whitespace cleanup in SciTECO now that tabs are
displayed properly
|
|
|
|
* the ^I command was altered to insert indention characters
rather than plain tabs always.
* The <TAB> immediate editing command was added for all
insertion arguments (I, ^I but also FR and FS)
* documentation was extended for a discussion of indention
|
|
this is because ownership of the GError may be passed to GlibError()
|
|
* when throwing GlibError(), this is taken care of automatically.
* fixes a memleak since there may be resources associated with the
GError.
|
|
earlier
* Debian 7 is still at libglib v2.33 and since it should be
supported, I reimplemented the missing function (UNIX-only).
* This workaround can be removed once libglib v2.34 becomes common place.
Closes #2
|
|
|
|
* for non-existing directories, NULL was passed to g_dir_read_name().
This resulted in Glib errors being printed to stdout/stderr.
* this was broken in commit 427c9d
|
|
Windows
* this is actually UNTESTED on Windows
|
|
optimized screen refreshing/redrawing
* pressing e.g. TAB when the popup is showing a list of auto-completions
will show the next page, eventually beginning at the first one again.
* do not redraw curses windows in the UI methods directly. this resulted
in flickering during command-line editing macros and ordinary macro calls
because the physical screen was updated immediately.
Instead, window refreshing and updated is done centrally in
event_loop_iter() only after a key has been processed.
Also we use wnoutrefresh() and doupdate() to send as little to the
terminal (emulator) as possible.
|
|
* simplified code
* fixed spurious empty lines in the popup which truncated file names/tokens
that would otherwise be displayed
* fixed memleak when freeing the popup entry list
|
|
* added platform-dependant file_is_visible() function
|
|
|
|
$SCITECOPATH on Windows
* $SCITECOCONFIG has been introduced, so have a macro-accessible location
for the profile, buffer session etc.
This is set to the program dir on Windows. That way, the config files
will be found, regardless of the current working dir, but it may also
be set up for Unix-like environments on Windows.
* $SCITECOPATH defaults to the program dir + "/lib" now on Windows.
* The default profile is now always called ".teco_ini". Also on Windows.
Platform differences like this would need to be documented.
* The sample teco.ini has been renamed to "sample.teco_ini" for clarity
|
|
minor change that fixes Doxygen-generated documentation
|
|
* in batch mode, Scintilla undo actions are simply leaked memory
* Since we have more than one Scintilla view now, we must empty
the undo buffer of all scintilla views when a command line is committed ($$)
|
|
* EW can save Q-Registers now
* the new E% may be used to save a q-register without making it
the current document
|
|
this will allow us to use the same algorithms for loading and saving
Q-Registers (from/to file).
* Saving with EW when a Q-Reg is edited has been fixed (was broken earlier)
* SciTECO save point files are now named .teco-X-BASENAME
When using IOView for Q-Regs, there will be no way to sensible count
the save points. Each write of a Q-Reg may be to another file.
Therefore, we number save-points globally.
If the sequence of writes has to be reconstructed manually,
one can still look at the save point files' modification dates
* give more informative error messages when saving a file fails
|
|
this is analoguous to EU as the string-build equivalent
of ^U.
|
|
it became apparent, that something like this is very useful,
when constructing the contents of a q-register without
editing it.
I have decided against introducing another modifier for toggling
string building. Most commands have string building enabled and it
doesn't hurt. For the few exceptions, an alternative variant can
be introduced.
|
|
this is more consistent with SciTECO's idea of abstract registers
and allows the currend buffer to be saved on the Q-Register stack.
This allows the idiom: [* ! ...change current buffer... ! ]*
|
|
* main motivation is to have a way of getting the number of buffers
in the ring. "EJ" or "1EJ" will do that.
This simplifies macros that will have to iterate all the buffers.
They no longer have to close the existing buffers to do that.
* "0EJ" will get the current user interface. This is useful to select
a different color scheme in the startup profile depending on the UI,
for instance.
|
|
* implements the same globbing as the EB command already did
* uses Globber helper class that behaves more like UNIX glob().
glib only has a glob-style pattern matcher.
* The Globber class may be extended later to provide more
UNIX-like globbing.
* lexer.tes has been updated to make use of globbing.
Now, lexers can be automatically loaded and registered at
startup. To install a new lexer, it's sufficient to copy
a file to the lexers/ directory.
|
|
|
|
|
|
the QUIT hook is actually not that trivial and required
some architectural changes.
First, the QUIT hook execution and any error that might
occurr cannot always be attached to an existing error stack
frame. Thereforce, to give a stack frame for QUIT hooks and
to improve the readability of error traces for ED hooks in general,
a special EDHookFrame is added to every ED hook execution error.
Secondly, since QUIT hooks can themselves throw errors, we cannot
run it from an atexit() handler. Instead it's always called manually
before __successful__ program termination. An error in a QUIT hook
will result in a failure return code nevertheless.
Thirdly, errors in QUIT hooks should not prevent program termination
(in interactive mode), therefore they are only invoked from main()
and always in batch mode. I.e. if the interactive mode is terminated
(EX$$), SciTECO will switch back to batch mode and run the QUIT
hook there. This is also symmetric to program startup, which is
always in batch mode.
This means that Interface::event_loop() no longer runs indefinitely.
If it returns, this signals that the interface shut down and
batch mode may be restored by SciTECO.
|
|
SciTECO commands usually only take parameters from the stack that
they need. Without protecting the ED hook execution with brace operators,
additional arguments not consumed by the hook-dispatching command
are passed into the ED hook invocation. Also an ED hook macro could
leave additional values on the expression stack (by accident).
All of this may lead to undefined behaviour if ED hooks
are involved.
|
|
introduced Error::set_coord()
|
|
This is only a problem if the macro created the local Q-Register table
(i.e. not when called with ":M") but resulted in segfaults.
Since we do not want to save in a Q-Reg whether it is local
(and that wouldn't suffice anyway), we do it in the Q-Register table
cleanup. The corresponding QRegisterTable::clear() must be called
explicitly, since the RBTree::clear() called on destruction does not
and cannot throw errors.
If QRegisterTable::clear() has been called successfully, the default object
destructor will not do much. If it has thrown an error, the destructor
will clean up the remaining Q-Registers.
|
|
it was not possible to inherit the parser state of the last command in
a macro. However, it was possible to exit the macro when its
last command was not properly terminated.
There is no useful conscious application of this behaviour.
* If something like an uncomplete E-Command was last in the
macro, there was absolutely no effect. However this points
to a macro programming error.
* If the last command started a string parameter but did not
terminate it, the command might be (partially) executed.
However the State's done() method wasn't called, which means
that some commands will not execute at all. Again there's
no useful application of this.
When on the other hand, the last command was not terminated
by accident, this was not reported to the user.
|
|
this eases handling of the "*" register
|
|
* does not change ./configure parameters
You still have to specifiy --with-interface=ncurses for
the Curses interface with default settings
* the "NCurses" UI was used for many different Curses
variants, so plain "Curses" is a better name.
|
|
* it must be initialized after the UI (Interface::main), so I added
a View::initialize() function
* the old initialize() method was renamed to setup()
* use a global instance of QRegister::view so it is guaranteed to
be destroyed only after any QRegisters that could still need it
* Document API adapted to work with ViewCurrent references
|
|
Both UIs have a current_view, so this field, the
ssm(), undo_ssm() and get_current_view() methods can be
refactored into the Interface base class
|
|
it now works as good (or bad) as it did before.
* fixed entering of the Escape character
|
|
* allowed me to remove some obscure global functions and methods like
QRegister::update_string().
* Document updating is concentrated in qregisters.cpp now
* also fixes some bugs introduced earlier, like undo tokens being
generated for non-undo registers (resulting in segfaults on rubout)
|
|
Template Pattern.
* without the one-view-per-buffer designs, many Scintilla send message (SSM)
calls could be inlined
* with the new design, this was no longer possible using the abstract
base classes. the CRT pattern allows inlining again but introduces
a strange level of code obscurity.
* tests suggest that at high optimization levels, the one-view-per-buffer
design and the CRT pattern reduces typical macro runtimes by 30%
(e.g. for symbols-extract.tes).
* only updated the NCurses UI for the time being
|
|
run() method later in interface.h
|
|
The user interface provides a Scintilla view abstraction and
every buffer is based on a view. All Q-Register strings use
a single dedicated view to save memory and initialization time
when using many string registers.
* this means we can finally implement a working lexer configuration
and it only has to be done once when the buffer is first added
to the ring. It is unnecessary to magically restore the lexer
styles upon rubout of EB (very hard to implement anyway). It
is also not necessary to rerun the lexer configuration macro
upon rubout which would be hard to reconcile with SciTECO's
basic design since every side-effect should be attached to a
character.
* this means that opening buffers is slightly slower now
because of the view initialization
* on the other hand, macros with many string q-reg operations
are faster now, since the document must no longer be changed
on the buffer's view and restored later on.
* also now we can make a difference between editing a document
in a view and changing the current view, which reduces UI calls
* the Document class has been retained as an abstraction about
Scintilla documents, used by QRegister Strings.
It had to be made virtual, so the view on which the document
is created can be specified by a virtual function.
There is no additional space overhead for Documents.
|
|
also changed precedence of + operator (higher than minus).
the effects of this should be minimal
|
|
this fixes the "\" command and ^E\ string building characters
|
|
* the GError expection has been renamed to GlibError, to avoid
nameclashes when working from the SciTECO namespace
|
|
normally, since SciTECO is not a library, this is not strictly
necessary since every library should use proper name prefixes
or namespaces for all global declarations to avoid name clashes.
However
* you cannot always rely on that
* Scintilla does violate the practice of using prefixes or namespaces.
The public APIs are OK, but it does define global functions/methods,
e.g. for "Document" that clashed with SciTECO's "TECODocument" class at
link-time.
Scintilla can put its definitions in a namespace, but this feature
cannot be easily enabled without patching Scintilla.
* a "SciTECO" namespace will be necessary if "SciTECO" is ever to be
turned into a library. Even if this library will have only a C-linkage
API, it must ensure it doesn't clutter the global namespace.
So the old "TECODocument" class was renamed back to "Document"
(SciTECO::Document).
|
|
the expression stack
now it's more like standard TECO's ^U command
|
|
|
