Version 2.1 alpha-test release now available!
Euphoria 2.1 alpha is now available from:
http://members.aol.com/FilesEu/
Here are the changes from version 2.0:
* We've made a number of changes to the packaging, pricing, and
registration incentives for the Euphoria product:
* The Single-Platform (DOS32-only) package, formerly $32,
has been discontinued.
* The Dual-Platform (DOS32+WIN32) package, formerly $53
has been reduced to $39 U.S., effective immediately.
Anyone who registered or upgraded, to Dual-Platform
*or Single-Platform*, since July 15, 1998 (6 months ago)
gets a free upgrade to 2.1. All other registered users pay
just $24.
* The printed manual has been discontinued. Instead, there is now
an official HTML version of the manual, included with the
Public Domain .ZIP file.
* All useful 3rd-party include files, such as Win32Lib.ew and many
others, will be "stamped" by RDS with a code number that makes
them free, just like the files in euphoria\include. They will not
add to your statement count, provided you do not significantly
modify them. This will also allow 3rd-party developers
to get better diagnostic information from their users.
* Binding, shrouding and profiling will now be part of the
Complete Edition only. These are features that beginners
do not require, but serious users might find valuable.
* Short-form assignment operators += -= *= /= &= have been added.
For example, instead of saying:
count = count + 1
You can now say:
count += 1
Instead of saying:
matrix[row][column] = matrix[row][column] * 5.0
You can say:
matrix[row][column] *= 5.0
Instead of saying:
test_scores[start..finish] = test_scores[start..finish] / 100
You can say:
test_scores[start..finish] /= 100
* Euphoria now uses "short-circuit" evaluation of "and" and "or"
expressions in if/elsif/while conditions. e.g. in an "and" condition:
if A and B then ...
the interpreter will skip the evaluation of expression B whenever
expression A is 0 (FALSE), since it knows that the overall result
must be FALSE. In an "or" condition:
while A or B do ...
the interpreter will skip the evaluation of expression B whenever
expression A is non-zero (TRUE), since it knows that the overall result
must be TRUE.
Euphoria code written prior to version 2.1 may no longer work correctly
if expression B contains a function with side-effects such as setting a
global variable, doing I/O etc. In practice this kind of code is very
rare, but just in case, a warning will now be issued if a function with
side-effects might be short-circuited.
By skipping the evaluation of B, short-circuit evaluation is typically
faster, and will allow you to write statements such as:
if atom(x) or length(x)=1 then ...
that would generate an error on older versions of Euphoria whenever x
was an atom, since length() is not defined for atoms.
See REFMAN.DOC for the details.
* Several new routines were added.
Built-in to ex.exe/exw.exe:
profile() - turns profiling on/off so you can focus your profile
and profile_time runs on particular events within your
program.
system_exec() - gives you the exit code from calling a .exe or .com
file, or another Euphoria program
equal() - compare any 2 Euphoria objects for equality.
equivalent to: compare(a,b) = 0 but more readable
Added to various include files:
walk_dir() - recursively goes through a directory and
subdirectories, calling a routine that you supply
reverse() - returns a sequence in reverse order
sprint() - returns the string representation of any Euphoria
object
arcsin() - inverse trig function
arccos() - inverse trig function
get_bytes() - returns the next n bytes from a file
prompt_number() - prompts the user to enter a number
prompt_string() - prompts the user to enter a string
instance() - WIN32: returns the instance handle of the program
the constant PI - 3.14159... was added to misc.e
* The main Euphoria documentation can now be viewed locally with a Web
browser. The plain-text files REFMAN.DOC and LIBRARY.DOC are still
available in the DOC subdirectory, but we now have REFMAN.HTM and
LIBRARY.HTM in the new HTML subdirectory. We have developed a tool
(written in Euphoria) that lets us easily maintain both an up-to-date
HTML version, and an up-to-date plain-text version of REFMAN and
LIBRARY.
The documentation has also been clarified and expanded in many places.
* WIN32: you can create an unlimited number of Euphoria call-back
routines, as long as each routine is a function with 0 to 8 parameters.
See PLATFORM.DOC. In version 2.0 you could only have one call-back
routine and it had to have 4 parameters.
* The "xor" keyword has been added to complement: and/or/not and
xor_bits() e.g.
if a xor b then...
"xor" works on sequences too. It's similar to "or".
* The dir(path) library routine now officially supports the use of
wildcards * and ? in the path that you supply. This feature was always
available, but wasn't documented until now. e.g.
info = dir("mydata\\*.d?t")
* OPTIMIZATION: subroutine call+return overhead was reduced by an
average of 30%. The speed-up occurs for all normal
function/procedure/type calls, user-defined type-checks,
call_proc/call_func calls using a routine id, and Windows call-backs.
Only recursive calls cost the same as before. Programs with a
reasonably-high frequency of calls can easily be 10% faster
overall because of this.
* OPTIMIZATION: Branch "straightening" has been implemented. The
compiler will optimize branches in the internal code such that a branch
from A->B where location B contains a branch to location C, will be
optimized to a direct branch from A->C. Even something like A->B->C->D
can be straightened to A->D. This often occurs in while-loops that
contain if-statements.
* OPTIMIZATION: In many cases, variable initialization checks are
now replaced by "no-ops" after the first check is performed. Euphoria
was already optimizing out many checks at compile-time.
* OPTIMIZATION: get() and value() are now much faster in most cases
thanks to Jiri Babor and some further optimizations by RDS.
The new v2.1 ex.exe with the new v2.1 get.e is:
1.45x faster reading a sequence of f.p. numbers from a file and
2.25x faster when reading sequence of integers from a file
* OPTIMIZATION: power(x,2) is converted internally to x*x which is faster
in all cases, especially when x is a large integer or a f.p. number.
* OPTIMIZATION: thanks to Jiri Babor, int_to_bits() is at least 15% faster
in most cases.
* OPTIMIZATION: Plotting a long sequence of pixels in 16-color graphics
modes is about 3% faster.
* OPTIMIZATION: draw_line() has been sped up by a few percent.
* Language War has had a major face-lift. It now runs in pixel-graphics
mode 18 (640 x 480 x 16 colors) instead of text mode. It also has
"fine-grain" parallelism, i.e. virtually anything can happen in parallel
with anything else. Multiple torpedos, phasors etc can be drawn on the
screen simultaneously, while ships are moving, commands are being
entered, things are exploding etc. Even the timing needed for the PC
speaker sound effects is handled by the task scheduler. There are no
time-delay "busy" loops executed during the game. The galaxy scan
now shows you a scaled picture of the whole galaxy, rather than just a
bunch of numbers.
* The default print format for atoms was changed from %g to %.10g
This format is used by print(), ?, the trace facility, and ex.err
dumps. This allows large integers -9,999,999,999 to +9,999,999,999
to be printed as integers, rather than as scientific notation. It also
provides about 10 digits of accuracy to be displayed on fractional
numbers, rather than just 6. Art Adamson and others made it clear that
more digits should be displayed.
* The state of all with/without settings is saved upon entering an
included file, and restored at the end of the included file.
An included file can change the settings, but they will be restored at
the end of the included file. e.g. warnings might be turned off just
within the included file (and any files it includes). As a result some
programs now display warnings where none were seen before.
* Warnings are now displayed *after* your program finishes execution,
so they won't be erased by clear_screen(), graphics_mode() etc.
Some programs now show warnings where none were seen before.
* The security of scrambled code and bound code has been improved
thanks to ideas contributed by Rusty Davis. When a bound program
starts executing, a quick integrity check will be made to detect any
corruption or tampering. It's still ok to add data to the end of a bound
.exe file, as long as your last line is "abort(x)".
* The ed editor now lets you view and edit beyond column 80.
* ed has a new command: Esc m (modifications). It will show the
differences between the original file on disk and the current edit
buffer. This can be very useful when you've forgotten what changes
you've made, and you are wondering if it's safe to save them.
* The trace window now provides an upper case Q command which
lets the program run to completion, ignoring any trace(1) commands.
Lower case q lets it run to the next trace(1).
* safe.e (debug version of machine.e) has been enhanced. It will now
automatically catch additional cases where data is illegally written just
before, or just after, the boundaries of an allocated block of memory.
This can be particularly useful in WIN32 where Windows might overwrite
one of your under-sized blocks. Without a tool such as safe.e, this type
of bug could take hours or even days to track down.
* The euphoria\tutorial directory was created to hold several small
tutorial programs.
* The limit on the number of open files was raised to 25 from 15.
Three of these files are 0,1,2: standard-input, standard-output
and standard-error, so you can now have up to 22 of your own files
open simultaneously. (As far as we know, no one ever exceeded the
old limit, but it seemed wise to raise it.)
* When the user simply types "ex" or "exw" and is prompted for the
name of the Euphoria .ex or .exw file to run, command_line() will now
be updated to include the filename as the second command-line
argument, just as if the user had originally typed: ex filename
Thanks to Mathew Hounsell for suggesting this.
* mset.ex now saves pictures in .bmp format. Previously it was using a
non-standard, compressed format.
* lines.ex (lines.bat) now reports non-blank/non-comment lines as well.
This is *not* the same as the "statement count" used by Euphoria for
the diagnostic limit, but it's usually within +/- 10%, assuming you write
one statement per line.
* numeric literals greater than 1e308 (roughly) are now set to +/- "inf".
They used to cause a compile-time error.
A bunch of fairly minor bugs were fixed. They are listed in
RELNOTES.DOC in EUPHOR21.ZIP.
Regards,
Rob Craig
Rapid Deployment Software
http://members.aol.com/FilesEu/
|
Not Categorized, Please Help
|
|
