| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
* This has been broken since the C conversion (432ad24e382681f1c13b07e8486e91063dd96e2e).
* Fixes getopt.tes, although no script actually checked the contents of
the getopt.X numeric Q-Registers.
grosciteco.tes also made use of this. It's unclear what has consequently been broken.
* NOTE: TECOC does not seem to support -Uq - this is a SciTECO extension.
|
|
* This has always been broken as Gtk will not hide the
window before suspending.
* It has been deemed to complicated to implement at the moment.
Even if we can catch SIGTSTP (not that trivial), it seems to be
impossible - at least without some lower level Xlib interaction -
to hide the program window before raising SIGTSTP.
* Even if everything worked, it is unclear whether it is actually
desirable to suspend a GUI application - ^Z may be pressed accidentally
and it will be inconvenient to resume the job.
So we would additionally have to check for the existence of
an attached console.
|
|
* Keeping a key pressed could still result in missing graphics updates and thus
visual feedback.
* Now try to process all Gdk events after thawing the window.
It no longer appears to happen.
* On the downside, key processing is much slower now which may result
in keys being queued up and processed some time even after releasing it.
There may be workarounds for that as well...
|
|
* Due to introducing another gtk_main_iteration_do(),
there could indeed be unforseen recursions of teco_interface_key_pressed_cb()
that resulted in additional teco_interface_handle_key_press() calls.
* This did not cause crashes, but we better prevent recursions altogether.
While emptying the key event queue, we only allow other events to be queued
by all possibly recursive invocations of teco_interface_key_pressed_cb().
|
|
* This was surprisingly easy to implement as Gtk+ 3 already
supports it via GtkPlug.
* Allows embedding SciTECO into other Xembed-aware applications.
* Unfortunately there are very few generic Xembed hosts.
tabbed (https://tools.suckless.org/tabbed/) would be one of them.
It could be used to add tabs to SciTECO even on non-tiling window managers:
$ tabbed sciteco --xembed
* Unfortunately, it does not seem to be possible to use this feature
to let SciTECO replace the contents of a terminal window even though
many terminal emulators provide $WINDOWID.
|
|
* Processing a queued list of key events with an idle timer turned out
to be tricky.
Since teco_interface_pop_key_idle_cb() would eventually drive the main
loop with gtk_main_iteration_do() which may result in a recursive
invocation of teco_interface_pop_key_idle_cb() which will eventually crash.
* We'd have to mask the idle watcher during the execution time of teco_interface_pop_key_idle_cb().
* Therefore it has been decided to use a tight loop again to process the
event queue.
After thawing the window, we now manually drive the event loop with gtk_main_iteration_do()
to make sure that the UI is updated.
This could result in a recursive invocation of teco_interface_key_pressed_cb() of course but
the callback is already secured against this.
|
|
* NOTE: Selections are currently only used to highlight search results.
* The default selection colors were not always visible well with default settings (--no-profile)
and they were not uniform across platforms.
On Curses, the selection would be reversed, while on Gtk it had a lighter foreground color.
They are now always reversed (black on white background).
The default styles do not assume any color support - they use only black and white.
* Since these defaults cannot possibly work on every color scheme,
color.selfore and color.selback has been added to color.tes.
All existing color schemes have been updated to configure selections as reversed
to the default colors.
This especially fixes selection colors on Gtk.
* On solarized.tes, the caret style was already distinct from inversed default colors.
On terminal.tes, the color of the caret is now bright white, so it stands out
from the selection colors.
* In Curses, the caret color is currently __not__ applied to the command line where
it is continued to be drawn reversed.
The command line drawing code is considered deprecated and will eventually be replaced
with a Scintilla minibuffer.
* In Gtk, we now apply the caret style to the commandline view as well.
* Fixed the comment color in solarized.light.
|
|
* The old implementation could apparently result in use-after-free
situations that are not related to unstopped watchers in Scintilla.
This would result in frequent crashes.
Possibly, this only now manifests after upgrading to Scintilla 5.
* The old implementation also had the bug that freeing views
(e.g. via <EF>) would not release any memory in batch mode since the
main loop is not triggered.
* I don't pretend to understand why we need gtk_widget_destroy()
instead of g_object_unref().
|
|
* teco_interface_key_pressed_cb() could be called multiple times __before__
the idle timer (teco_interface_pop_key_idle_cb()) fires.
The recursion check would consequently not work and we started the
idle timer multiple times.
This would eventually crash.
* We now process the first queued key immediately.
The alternative would be to store the idle watcher id.
* The idle watcher's priority has been increased.
Since redrawing is guaranteed to take place at G_PRIORITY_HIGH_IDLE,
it is sufficient to process keys at G_PRIORITY_DEFAULT_IDLE.
* Should also reduce latency slightly.
* fixes up 71bf522231d2998f1fb183f343c2b1f9dbcd3b15
|
|
* Test case: sciteco -e '[a'
[aEX$$ in interactive mode would also crash.
* No longer use a destructor - it was executed after the Q-Reg view was
destroyed.
* Instead, we now explicitly call teco_qreg_stack_clear() in main().
* Added a regression test case.
|
|
updated regularily when holding down a key
* gdk_window_freeze_toplevel_updates_libgtk_only() is apparently no longer necessary
with Scintilla 5.
* When holding down a key constantly, it was not uncommon that the display would not be updated
until it is released.
This is now worked around by using a low priority idle timer for emptying the teco_interface.event_queue.
This ensures that Gtk can call other watchers after every keypress.
|
|
once per keypress
* Esp. costly since Scintilla 5.
* We now avoid any Scintilla message that automatically scrolls the caret (makes the
caret visible) and instead call SCI_SCROLLCARET only once after every keypress in the
interface implementation.
* From nowon, use
* SCI_SETEMPTYSELECTION instead of SCI_GOTOPOS
* SCI_SETEMPTYSELECTION(SCI_POSITIONFROMLINE(...)) instead of SCI_GOTOLINE
* SCI_SETSELECTIONSTART and SCI_SETSELECTIONEND instead of SCI_SETSEL
* With these optimizations we are significantly faster than before
the Scintilla upgrade (6e67f5a682ff46d69888fec61b94bf45cec46721).
It is now even safe to execute the Gtk test suite during CI.
|
|
* Esp. with the new Scintilla version, the representation
setting as part of every SCI_SETDOCPOINTER has turned out to
be a performance bottleneck.
* The new Scintilla has a custom tweak/patch that disables any
automatic representation setting in Scintilla itself.
It is now sufficient to initialize the SciTECO-style representations
only once in the lifetime of any view.
|
|
* Previous Scintilla version was 3.6.4 and Scinterm was 1.7 (with lots of custom patches).
All of the patches are now either irrelevant or have been merged upstream.
* Since Scintilla 5 requires C++17, this increases the minimum GCC version at least
to 5.0. We may actually require even newer versions.
* I could not upgrade the scintilla-mirror (which was imported from Mercurial),
so the old sciteco-dev branch was renamed to sciteco-dev-pre-v2.0.0,
master was deleted and I reimported the entire Scintilla repo using
git-remote-hg.
This means that scintilla-mirror now contains two entirely separate trees.
But it is still possible to clone old SciTECO repos.
* The strategy/workflow of maintaining hotfix branches on scintilla-mirror has been changed.
Instead of having one sciteco-dev branch that is rebased onto new Scintilla upstream
releases and tagging SciTECO releases in scintilla-mirror (to keep the commits referenced),
we now create a branch for every Scintilla version we are based on (eg. sciteco-rel-5-1-3).
This branch is never rebased or deleted. Therefore, we are guaranteed to be able to
clone arbitrary SciTECO repo commits - not only releases.
Releases no longer have to be tagged in scintilla-mirror.
On the downside, fixup commits may accumulate in these new branches.
They can only be squashed once a new branch for a new Scintilla release is created
(e.g. by cherry-picking followed by rebase).
* Scinterm does no longer have to reside in the Scintilla subdirectory,
so we added it as a regular submodule.
There are no more recursive submodules.
The Scinterm build system has not been improved at all, but we use
a trick based on VPATH to build Scinterm in scintilla/bin/.
* Scinterm is now in Git and we reference the upstream repo for the
time being.
We might mirror it and apply the same branching workflow as with Scintilla
if necessary.
The scinterm-mirror repository still exists but has not been touched.
We will also have to rewrite its master branch as it was a non-reproducible
Mercurial import.
* Scinterm now also comes with patches for Scintilla which we simply applied
on our sciteco-rel-5-1-3 branch.
* Scintilla 5 outsourced its lexers into the Lexilla project.
We added it as yet another submodule.
* All submodules have been moved into contrib/.
* The Scintilla API for setting lexers has consequently changed.
We now have to call SCI_SETILEXER(0, CreateLexer(name)).
As I did not want to introduce a separate command for setting lexers,
<ES> has been extended to allow setting lexers by name with the SCI_SETILEXER
message which effectively replaces SCI_SETLEXERLANGUAGE.
* The lexer macros (SCLEX_...) no longer serve any purpose - they weren't used
in the SciTECO standard library anyway - and have consequently been removed
from symbols-scilexer.c.
The style macros from SciLexer.h (SCE_...) are theoretically still useful - even
though they are not used by our current color schemes - and have therefore been
retained. They can be specified as wParam in <ES>.
* <ES> no longer allows symbolic constants for lParam.
This never made any sense since all supported symbols were always wParam.
* Scinterm supports new native cursor modes.
They are not used for the time being and the previous CARETSTYLE_BLOCK_AFTER
caret style is configured by default.
It makes no sense to enable native cursor modes now since the
command line should have a native cursor but is not yet a Scintilla view.
* The Scintilla upgrade performed much worse than before,
so some optimizations will be necessary.
|
|
|
|
* Was only broken on Gtk+ 3.24 for Windows with the builtin theme engine.
* The warning level was also not themed correctly because of a typo.
|
|
* This was an ancient bug apparently broken since d503c3b07c2157658f699294c44ad5be244727a5 (year 2014)
and was therefore broken even in v0.6.4.
|
|
* We don't need the PNG icons on Windows as the compiled-in ICO should suffice
* Ship the dependencies of the SVG pixbuf loader.
* The PNG pixbuf loader is still distributed, as we at least need it
for loading the icon theme.
* Install a loaders.cache - without it, the pixbuf loaders won't be found.
This file can be generated by gdk-pixbuf-query-loaders but apparently has
to be modified by hand.
* Regenerate the icon cache using gtk-update-icon-cache.
* Icon themes are found now.
Unfortunately, we have to distribute the entire Adwaita icon theme
as distributing only the scalable (SVG) icons does not work for some
strange reason (FIXME).
|
|
* Since we don't have an absolute and known installation directory,
we should look for these files in the same directory as sciteco.exe.
* For the time being, we look for them in $SCITECOCONFIG which defaults
to that directory.
* Cannot be a final solution as you may tweak $SCITECOCONFIG to
fit an Unix-like environment (eg. set SCITECOCONFIG=$HOME).
In such configurations it may also not be suitable to always
look in the directory of sciteco.exe since that may be some /bin dir.
Considering that GTK+ forces us to preserve some kind of UNIX-like
directory hierarchy even for portable builds, we should perhaps
install the icons into the hicolor icon theme.
This would also simplify Debian packaging.
|
|
* follow the current terminology:
* PDCurses/Win32a is now called PDCursesMod and includes all other PDCurses ports as well.
The Win32 GUI port is now called PDCurses/WinGUI.
* PDCurses/Win32 is now called PDCurses/WinCon.
* Since PDCursesMod supports WinCon as well, we use the PDCURSES_MOD macro only
to detect PDCursesMod API extensions.
GUIs (detached from system console) might be available both in classic PDCurses as well
as in PDCursesMod.
Only PDCursesMod allows detection of the port used *at runtime* using PDC_get_version().
We therefore introduced a --with-interface=pdcurses-gui that must be given whenever
compiling for any kind of GUI port (including SDL on "classic" PDCurses).
* The PDCURSES macro is used to detect all PDCurses (whether classic or PDCursesMod) API extensions.
* __PDCURSES__ is used to detect PDCurses whenever API extensions are not required.
* Assume that A_UNDERLINE now works even on WinCon.
|
|
with Unix linebreaks
* when hosted on Windows, the default is DOS linebreaks
* Unix linebreaks are in many cases more consistent as all other sources use Unix linebreaks
* woman pages with Unix linebreaks are slightly faster to load due to EOL conversion
* especially Groff input must not contain CR as it will otherwise log lots of warnings
(affects htbl.tes and tedoc.tes).
|
|
teco-gtk-label.gob to plain C
* Using modern GObject idioms and macros greatly reduces the necessary boilerplate code.
* The plain C versions of our GObject classes are now "final" (cannot be derived)
This means we can hide the instance structures from the headers and avoid using
explicit private fields.
* Avoids some deprecation warnings when building the Gtk UI.
* GOB2 is apparently no longer maintained, so this seems like a good idea in the long run.
* The most important reason however is that there is no precompiled GOB2 for Windows
which prevents compilation on native Windows hosts, eg. during nightly builds.
This is even more important as Gtk+3 is distributed on Windows practically
exclusively via MSYS.
(ArchLinux contains MinGW gtk3 packages as well, so cross-compiling from ArchLinux
would have been an alternative.)
|
|
"optimized" code-path on UNIX
|
|
* Turned out to be useful in debugging the "Memory limiting during spawning" test case
on Windows.
* Use UNIX shell emulation (0,128ED) in all test cases.
Should be necessary in order to run the testsuite on Windows, but
it is currently broken anyway.
* avoid <EG> when preprocessing files - use GNU Make's $(shell) instead
* Fixes builds on MinGW where there are still problems with <EC> and <EG>
at least in the virtual build environment.
* Results in a another automake warning about non-POSIX Make constructs.
This is not critical since we depend on GNU Make anyway.
|
|
* Environment variables are case insensitive on Windows
while SciTECO variables are case sensitive.
We must therefore make sure that we first unset any $COMSPEC or $ComSpec
from the environment before resetting it, thereby fixing its case.
* Fixes command execution via <EC> on systems where the variable
was not called $ComSpec.
|
|
(<S>) for multiplication overflows
* Since the numbers come from "outside" (SciTECO scripts) this is easily possible,
resulting either in missed checks or even memory corruption.
* In particular, this fixes the "Searching with large counts" test case on 32-bit builds.
Perhaps at least one CI build should be 32-bit?
|
|
* this flag could negatively affect optimizations
* fixes builds on MinGW
|
|
don't build an empty libdlmalloc
* on some platforms (eg. Darwin/mac OS) we cannot apparently build empty
convenience libraries
* instead, we use conditional subdirectories and a conditional library dependency
|
|
* I could not get malloc replacement via dlmalloc to work.
This does not work like on Linux by overwriting weak malloc() functions.
It should theoretically be possible to overwrite the default malloc zone
but I could not properly debug this since I can only build for Mac OS
via CI.
* memory polling seems to work though - test suite runs through
and it includes memory limiting test cases.
|
|
would be exceeded
* Checking whether the allocation succeeded may not prevent exceeding the memory
limit excessively.
* Even if the memory limit is not exceeded, the allocation can fail theoretically
and the program would terminate abnormally.
This however is true for all allocations in SciTECO (via glib).
* teco_memory_check() therefore now supports checking whether an allocation would
exceed the memory limit which will be useful before very large or variable allocations
in addition to the regular checking in teco_machine_main_step().
* As a sideeffect, this fixes the "Searching with large counts" test case on Mac OS
where too large allocations were not detected as expected (apparently Mac OS
happily gives out ridiculously large chunks of memory).
Now, all platforms are guaranteed to have the same behaviour.
|
|
NOTE: Aliases and weak symbols must not be used for portability reasons!
|
|
file systems
* There is a "Scintilla.h" as well.
* should fix macOS and builds on native Windows hosts
* It wasn't practical to refer to the Scintilla includes using paths since
the Scintilla location is configurable (--with-scintilla).
So we'd have to write something like #include <include/Scintilla.h>.
For Scinterm we cannot avoid collisions neither as its path is also
configurable (--with-scinterm).
Effectively, we must prevent name clashes across SciTECO and all
of Scintilla and Scinterm.
|
|
This is a total conversion of SciTECO to plain C (GNU C11).
The chance was taken to improve a lot of internal datastructures,
fix fundamental bugs and lay the foundations of future features.
The GTK user interface is now in an useable state!
All changes have been squashed together.
The language itself has almost not changed at all, except for:
* Detection of string terminators (usually Escape) now takes
the string building characters into account.
A string is only terminated outside of string building characters.
In other words, you can now for instance write
I^EQ[Hello$world]$
This removes one of the last bits of shellisms which is out of
place in SciTECO where no tokenization/lexing is performed.
Consequently, the current termination character can also be
escaped using ^Q/^R.
This is used by auto completions to make sure that strings
are inserted verbatim and without unwanted sideeffects.
* All strings can now safely contain null-characters
(see also: 8-bit cleanliness).
The null-character itself (^@) is not (yet) a valid SciTECO
command, though.
An incomplete list of changes:
* We got rid of the BSD headers for RB trees and lists/queues.
The problem with them was that they used a form of metaprogramming
only to gain a bit of type safety. It also resulted in less
readble code. This was a C++ desease.
The new code avoids metaprogramming only to gain type safety.
The BSD tree.h has been replaced by rb3ptr by Jens Stimpfle
(https://github.com/jstimpfle/rb3ptr).
This implementation is also more memory efficient than BSD's.
The BSD list.h and queue.h has been replaced with a custom
src/list.h.
* Fixed crashes, performance issues and compatibility issues with
the Gtk 3 User Interface.
It is now more or less ready for general use.
The GDK lock is no longer used to avoid using deprecated functions.
On the downside, the new implementation (driving the Gtk event loop
stepwise) is even slower than the old one.
A few glitches remain (see TODO), but it is hoped that they will
be resolved by the Scintilla update which will be performed soon.
* A lot of program units have been split up, so they are shorter
and easier to maintain: core-commands.c, qreg-commands.c,
goto-commands.c, file-utils.h.
* Parser states are simply structs of callbacks now.
They still use a kind of polymorphy using a preprocessor trick.
TECO_DEFINE_STATE() takes an initializer list that will be
merged with the default list of field initializers.
To "subclass" states, you can simply define new macros that add
initializers to existing macros.
* Parsers no longer have a "transitions" table but the input_cb()
may use switch-case statements.
There are also teco_machine_main_transition_t now which can
be used to implement simple transitions. Additionally, you
can specify functions to execute during transitions.
This largely avoids long switch-case-statements.
* Parsers are embeddable/reusable now, at least in parse-only mode.
This does not currently bring any advantages but may later
be used to write a Scintilla lexer for TECO syntax highlighting.
Once parsers are fully embeddable, it will also be possible
to run TECO macros in a kind of coroutine which would allow
them to process string arguments in real time.
* undo.[ch] still uses metaprogramming extensively but via
the C preprocessor of course. On the downside, most undo
token generators must be initiated explicitly (theoretically
we could have used embedded functions / trampolines to
instantiate automatically but this has turned out to be
dangereous).
There is a TECO_DEFINE_UNDO_CALL() to generate closures for
arbitrary functions now (ie. to call an arbitrary function
at undo-time). This simplified a lot of code and is much
shorter than manually pushing undo tokens in many cases.
* Instead of the ridiculous C++ Curiously Recurring Template
Pattern to achieve static polymorphy for user interface
implementations, we now simply declare all functions to
implement in interface.h and link in the implementations.
This is possible since we no longer hace to define
interface subclasses (all state is static variables in
the interface's *.c files).
* Headers are now significantly shorter than in C++ since
we can often hide more of our "class" implementations.
* Memory counting is based on dlmalloc for most platforms now.
Unfortunately, there is no malloc implementation that
provides an efficient constant-time memory counter that
is guaranteed to decrease when freeing memory.
But since we use a defined malloc implementation now,
malloc_usable_size() can be used safely for tracking memory use.
malloc() replacement is very tricky on Windows, so we
use a poll thread on Windows. This can also be enabled
on other supported platforms using --disable-malloc-replacement.
All in all, I'm still not pleased with the state of memory
limiting. It is a mess.
* Error handling uses GError now. This has the advantage that
the GError codes can be reused once we support error catching
in the SciTECO language.
* Added a few more test suite cases.
* Haiku is no longer supported as builds are instable and
I did not manage to debug them - quite possibly Haiku bugs
were responsible.
* Glib v2.44 or later are now required.
The GTK UI requires Gtk+ v3.12 or later now.
The GtkFlowBox fallback and sciteco-wrapper workaround are
no longer required.
* We now extensively use the GCC/Clang-specific g_auto
feature (automatic deallocations when leaving the current
code block).
* Updated copyright to 2021.
SciTECO has been in continuous development, even though there
have been no commits since 2018.
* Since these changes are so significant, the target release has
been set to v2.0.
It is planned that beginning with v3.0, the language will be
kept stable.
|
|
detect EMCurses
* Emscripten can be used (theoretically) to build a host-only platform-independant version
of SciTECO (running under node.js instead of the browser).
* I ported netbsd-curses with Emscripten for that purpose. Therefore, adaptions for running
in the browser are restricted to EMcurses now.
|
|
* Array allocations were not properly accounted since the compiler
would call the replacement new() which assumes that it would
always be called along with the replacement sized-deletion.
This is not true for array new[] allocations resulting in
a constant increase of memory_usage and unrecoverable situations.
This problem however could be fixed in principle by avoiding
memory counting for arrays or falling back to malloc_usable_size.
* The bigger problem was that some STLs (new_allocator) are broken, calling the
non-sized delete for regular new() calls which could in principle
be matched by sized-delete.
This is also the reason why I had to provide a non-sized
delete replacement, which in reality intoduced memory leaks.
* Since adding checks for the broken compiler versions or a configure-time
check that tries to detect these broken systems seems tedious,
I simply removed that optimization.
* This means we always have to rely on malloc_usable_size() now
for non-SciTECO-object memory measurement.
* Perhaps in the future, there should be an option for allowing
portable measurement at the cost of memory usage, by prefixing
each memory chunk with the chunk size.
Maintainers could then decide to optimize their build for "speed"
at the cost of memory overhead.
* Another solution to this non-ending odyssey might be to introduce
our own allocator, replacing malloc(), and allowing our own
precise measurements.
|
|
* it turned out to be possible to provoke memory_usage
overflows or underruns, resulting in unrecoverable states
* a possible reason can be that at least with G++ 5.4.0,
the compiler would sometimes call the (default) non-sized
delete followed by our custom sized delete/deallocator.
* This was true even after compiling Scintilla with -fsized-deallocation.
* therefore we provide an empty non-sized delete now.
* memory_usage counting can now be debugged by uncommenting
DEBUG_MAGIC in memory.cpp. This uses a magic value to detect
instrumented allocations being mixed with non-instrumented
allocations.
* simplified the global sized-deallocation functions
(they are identical to the Object-class allocators).
|
|
Automakefiles could be simplified by updating CXXFLAGS
in configure.ac instead.
|
|
* avoid warnings
* make sure Doxygen finds RBEntryOwnString
* it would be nice to strip the top level `SciTECO` namespace
but this is not supported without some macro magic that
ommit the namespace declaration when processing with
Doxygen.
|
|
* when enabled, it will automatically upper-case all
one or two letter commands (which are case insensitive).
* also affects the up-carret control commands, so they when inserted
they look more like real control commands.
* specifically does not affect case-insensitive Q-Register specifications
* the result are command lines that are better readable and conform
to the coding style used in SciTECO's standard library.
This eases reusing command lines as well.
* Consequently, string-building and pattern match characters should
be case-folded as well, but they aren't currently since
State::process_edit_cmd() does not have sufficient insight
into the MicroStateMachines. Also, it could not be delegated
to the MicroStateMachines.
Perhaps they should be abandoned in favour of embeddedable
regular state machines; or regular state machines with a stack
of return states?
|
|
* this resulted in assertions (crashes!) for harmless typos like "+23="
* a test case has been added
|
|
* test case: HECcat$ on a large buffer (>= 64kb)
truncates the buffer or repeats its beginning
* it turns out that the incremental writing to the process' stdin
was broken. We were always writing data from the beginning of the buffer
which fails if the stdin watcher must be activated more than once.
* Also, EOLWriter::convert() can validly return 0, even if bytes have
been written on the data sink, so this value cannot be used to
check whether the process has closed its stdin.
We now make sure that the entire buffer range is written to stdin.
* Piping large buffers no longer removes the buffer gap.
This makes little difference when filtering via EC since
it will change the buffer gap anyway.
Can make a huge difference when not touching the buffer, though
(e.g. HEGAcat$).
* I did not add a test suite case since that requires
a very large test file and it cannot be easily generated automatically.
|
|
* StateQueryQReg is now derived from StateExpectQReg
whose semantics have been changed slightly.
* The alternative would have been another common base class for both
StateQueryQReg and StateExpectQReg.
|
|
support
* Since netbsd-curses can act as a drop-in replacement to ncurses,
SciTECO builds with --with-interface=ncurses as well.
However, it is unintuitive for users to build with ncurses support
when actually linking against netbsd-curses; so another option has been added.
* The UNIX/TTY specific code (which works with both ncurses and netbsd-curses)
was selected when NCURSES was detected at build-time.
This does not work for netbsd-curses, so we define a new symbol
NETBSD_CURSES. At build-time, a CURSES_TTY macro may now be defined.
* This effectively fixes the stdio in interactive mode, window titles
and the XTerm clipboard support for netbsd-curses.
Some minor features like the reduced ESCDELAY are still broken.
|
|
been shown to be unacceptably broken, so the fallback implementation has been improved
* mallinfo() is not only broken on 64-bit systems but slows things
down linearilly to the memory size of the process.
E.g. after 500000<%A>, SciTECO will act sluggish! Shutting down
afterwards can take minutes...
mallinfo() was thus finally discarded as a memory measurement
technique.
* Evaluating /proc/self/statm? has also been evaluated and discarded
because doing this frequently is even slower.
* Instead, the fallback implementation has been drastically improved:
* If possible use C++14 global sized deallocators, allowing memory measurements
across the entire C++ code base with minimal runtime overhead.
Since we only depend on C++11, a lengthy Autoconf check had to be introduced.
* Use malloc_usable_size() with global non-sized deallocators to
measure the approx. memory usage of the entire process (at least
the ones done via C++).
The cheaper C++11 sized deallocators implemented via SciTECO::Object still
have precedence, so this affects Scintilla code only.
* With both improvements the test case
sciteco -e '<@EU[X^E\a]"^E\a"%a>'
is handled sufficiently well now on glibc and performance is much better
now.
* The jemalloc-specific technique has been removed since it no longer
brings any benefits compared to the improved fallback technique.
Even the case of using malloc_usable_size() in strict C++ mode is
up to 3 times faster.
* The new fallback implementation might actually be good enough for
Windows as well if some MSVCRT-specific support is added, like
using _msize() instead of malloc_usable_size().
This must be tested and benchmarked, so we keep the Windows-specific
implementation for the time being.
|
|
* in a flat list of undo tokens, we need to store
the program counter (ie. command line position)
that the undo token corresponds to.
Since in general there is more than one undo token per
input character, this stored PCs redundantly.
* For input characters with no undo tokens
(only applies to NOPs like space in the command line
macro), this needs one more pointer than before.
* In case of 1 undo token per input character,
the new implementation uses approx. the same memory.
* In the most common case of more than one undo token
per input character, this saves at least 4 bytes per
undo token.
* In large macros and long loops the effect is especially
pronounced. E.g. 500000<%A> will use 8MB less memory
with the new implementation.
|
|
editing key
* StateEscape should return the same fnmacro mask as StateStart
* When rubbing out a command, we should stop at StateEscape as well.
Therefore we reintroduced States::is_start().
RTTI is still not used.
|
|
as State::process_edit_cmd() virtual methods
* Cmdline::process_edit_cmd() was much too long and deeply nested.
It used RTTI excessively to implement the state-specific behaviour.
It became apparent that the behaviour is largely state-specific and could be
modelled much more elegantly as virtual methods of State.
* Basically, a state can now implement a method to customize its
commandline behaviour.
In the case that the state does not define custom behaviour for
the key pressed, it can "chain" to the parent class' process_edit_cmd().
This can be optimized to tail calls by the compiler.
* The State::process_edit_cmd() implementations are still isolated in
cmdline.cpp. This is not strictly necessary but allows us keep the
already large compilations units like parser.cpp small.
Also, the edit command processing has little to do with the rest of
a state's functionality and is only used in interactive mode.
* As a result, we have many small functions now which are much easier to
maintain.
This makes adding new and more complex context sensitive editing behaviour
easier.
* State-specific function key masking has been refactored by introducing
State::get_fnmacro_mask().
* This allowed us to remove the States::is_*() functions which have
always been a crutch to support context-sensitive key handling.
* RTTI is almost completely eradicated, except for exception handling
and StdError(). Both remaining cases can probably be avoided in the
future, allowing us to compile smaller binaries.
|
|
and added a FreeBSD/jemalloc-specific implementation
* largely reverts 39cfc573, but leaves in minor and documentation
changes.
* further experimentation of memory limiting using malloc() wrapping
has shown additional problems, like dlsym() calling malloc-functions,
further reducing the implementation to glibc-specific means.
This means there had been no implementation for FreeBSD and checks
would have to rely on undocumented internal implementation details
of different libcs, which is not a good thing.
* Other problems have been identified, like having to wrap calloc(),
guarding against underruns and multi-thread safety had been identified
but could be worked around.
* A technique by calculating the memory usage as sbrk(0) - &end
has been shown to be effective enough, at least on glibc.
However even on glibc it has shortcomings since malloc() will
somtimes use mmap() for allocations and the technique
relies on implementation details of the libc.
Furthermore another malloc_trim(0) had to be added to the error
recovery in interactive mode, since glibc does not adjust the program break
automatically (to avoid syscalls I presume).
* On FreeBSD/jemalloc, the sbrk(0) method totally fails because jemalloc
exclusively allocates via mmap() -> that solution was discarded as well.
* Since all evaluated techniques turn out to be highly platform
specific, I reverted to the simple and stable platform-specific
mallinfo() API on Linux.
* On FreeBSD/jemalloc, it's possible to use mallctl("stats.allocated")
for the same purpose - so it works there, too now.
It's slower than the other techniques, though.
* A lengthy discussion has been added to memory.cpp, so that we
do not repeat the previous mistakes.
|
|
* shouldn't make much of a difference, since we're in deep trouble
when they return NULL, but the wrappers should be transparent
instead of crashing in malloc_usable_size().
|
|
portable and faster hack
* Works by "hooking" into malloc() and friends and counting the
usable heap object sizes with malloc_usable_size().
Thus, it has no memory-overhead.
* Will work at least on Linux and (Free)BSD.
Other UNIXoid systems may work as well - this is tested by ./configure.
* Usually faster than even the fallback implementation since the
memory limit is hit earlier.
* A similar approach could be tried on Windows (TODO).
* A proper memory-limiting counting all malloc()s in the system can make
a huge difference as this test case shows:
sciteco -e '<@EU[X^E\a]"^E\a"%a>'
It will allocate gigabytes before hitting the 500MB memory limit...
* Fixed the UNIX-function checks on BSDs.
|
