- Several problems pointed out by the Seed7 community have been resolved.
- HTML parsing has been improved.
- The search for the X11 Xrender extension has been improved.
- Seed7's approach to avoid double library includes has been improved.
- Now, unused system libraries are not linked to the executable.
- Additional optimizations have been introduced in the Seed7 compiler.
- Operations on enumerations are now safe to be in the range of defined values.
- The libraries pixmap_file.s7i and graph_file.s7i have been improved to draw to the correct graphic window. Many thanks to Anders Carstensen for pointing out that text was written into the wrong window and for providing a test program.
- Answers to three new questions have been added to the FAQ. Answers to existing questions have been improved. Many thanks to Zachary Menzies for pointing out a problem with two dimensional arrays and for providing a test program.
- An unclosed <i> tag has been removed from the manual. Thanks to ifethereal for pointing out this error. Several other places with illegal HTML in doc/manual.htm have also been fixed.
- The function getXmlTagHeadOrContent() has been improved in scanfile.s7i and scanstri.s7i. Now, tag names start with a letter or underscore (_), and the function can read XML contents that start with the character '<' (like < <= or <> ). In this case,
the character < is read as character reference (<).
- An explanation of the exception declaration has been added to the manual.
- The search for the X11 Xrender extension has been improved. In chkccomp.c the function determineX11Defines() has been improved. Declarations have been moved from x11_x.h to the new file x11_rend.h. Adjustements have been made in drw_x11.c and fwd_x11.c.
- Two code pages for APL symbol encoding have been added to charsets.s7i. The graphic representation of the character '\a' (bell) has been changed from '\16#25cf;' to '\16#2022;' in all code sets.
- In the vector font cronos16 the design of the characters 'J' '£' and '€' have been improved.
- In the vector font cronos27 the design of the characters '£' '¤' and '€' have been improved.
- In analyze.c, the function analyzeProg() has been improved to work correctly if an error occurs.
- In itflib.c, the function itf_create() has been improved to raise the exception ACTION_ERROR if the source struct is NULL.
- The prototypes of the functions conWrite() and stri_to_bstriw() have been changed to work with constant strings.
Changelog of the previous release (2021-10-09):
- Seed7's approach to avoid double library includes has been improved. Many thanks to Zachary Menzies for reporting the problem (a second library with the same name but in a different directory was not included) and for providing a test case to trigger
it. The new mechanism uses the absolute path of a library to determine if it already has been included. Now the map of included libraries is not part of the compiled executable anymore. Changes have been made in seed7_05.s7i, analyze.c, data.h, infile.c,
infile.h, libpath.c, libpath.h, prclib.c and prg_comp.c.
- The new library htmldom.s7i has been added. This library contains an improved HTML DOM parser. Many thanks to OddCitron1981 for suggesting to parse some of the wild HTML out there on the web. The functions readHtmlNode() and readHtml() have been moved
from xmldom.s7i to htmldom.s7i. The type htmlDocument and the function readHtmlContainerSubNodes() have been added. Improvements of HTML scanning functions were also made due to this suggestion. The new HTML parser considers several things special for
* Tag names and attribute names are converted to lower case.
* There are alternate end tags for tags with optional closing tag.
* Attributes without value get "" as value.
* The <!DOCTYPE data is not handled as xmlNode.
* Closing tags without opening tag are left in as is.
- The HTML scanning functions in scanfile.s7i and scanstri.s7i have been improved:
* HTML tag attributes without value are allowed now.
* Reading of HTML tag attributes now stops at the end of tag character (>).
* CDATA sections are considered now.
* The functions getXmlTagHeadOrContent(), getSymbolInXmlTag(), getHtmlAttributeValue() and getNextHtmlAttribute() have been improved.
* The function getXmlCdataContent() has been added.
- A chapter about for-until-loops has been added to the manual.
- The makefiles and the compiler (s7c.sd7) have been improved to avoid linking unused system libraries (e.g. with -lm). Changes have been done in cc_conf.s7i, comp/action.s7i, comp/flt_act.s7i, comp/library.s7i, s7c.sd7, cmd_rtl.c and in the makefiles.
- The bas7.sd7 (basic interpreter) example program has been improved.
* Now, it is possible to do a string multiplication with the * operator. E.g.: "ha"*3 results in "hahaha" and "ab"*2+"xy"*3 results in "ababxyxyxy".
* Now, the RPT$ function is checked for a negative factor.
- The wiz.sd7 example program has been refactored. The functions treasureNumber() and vendorDies() have been introduced.
- The bigfiles.sd7 example program has been improved to limit the length of the result list.
- The compiler has been improved:
* Now, unused system libraries are not linked to the executable.
* In comp/flt_act.s7i the implementation of FLT_DECOMPOSE has been improved and float comparisons set the flag mathLibraryUsed, if the implementation requires it.
* Two functions named appendLibrary() have been added to s7c.sd7. These functions avoid that a system library is linked twice.
* In comp/action.s7i calls of BIG_... actions now set the flag bigintLibraryUsed and calls of FLT_... actions (that need the math system library) now set the flag mathLibraryUsed.
* The flags bigintLibraryUsed and mathLibraryUsed have been added to comp/library.s7i.
- In xmldom.s7i the writeXml functions have been refactored. Unnecessary definitions of writeXml have been removed.
- Definitions of SYSTEM_BIGINT_LIBS and SYSTEM_MATH_LIBS have been added to cc_conf.s7i. The definition of ADDITIONAL_SYSTEM_LIBS has been removed. SYSTEM_BIGINT_LIBS and SYSTEM_MATH_LIBS are used in confval.sd7 and s7c.sd7.
- Several improvements in chkccomp.c have been done:
* Now SYSTEM_MATH_LIBS and LINKER_OPT_DYN_LINK_LIBS are considered. This helps to avoid linking unused libraries.
* ADDITIONAL_SYSTEM_LIBS has been renamed to SYSTEM_BIGINT_LIBS.
* Now, it checks if fileno() succeeds after a successful call of popen() (this fixes a problem with Emscripten).
* The function appendOption() has been improved.
* The type of several indices has been changed from int to unsigned int (this reduces the number of C warnings).
* The value LINKER_OPT_DYN_LINK_LIBS is now added to a corresponding list of system libraries if dynamic linking at run-time is necessary.
- In cmd_rtl.c the function doReadLink() has been improved to work also for symlinks in the Linux /proc filesystem (in /proc the stat() function reports a symlink size of 0).
- The macro environmenStrncmp has been renamed to environmentStrncmp.
- The function getProgramPath() has been moved from analyze.c to cmd_rtl.c. Additionally it has been improved and renamed to getAbsolutePath().
- In cmd_unx.c the function getExecutablePath() has been improved to use doReadLink() and to return a straightened absolute path (the special directories "." and ".." are interpreted according to their conventional meanings).
- The functions concatAndStraightenPath() and straightenAbsolutePath() have been added to str_rtl.c.
- In infile.c the functions open_infile(), close_infile(), open_string() and remove_prog_files() have been renamed to openInfile(), closeInfile(), openString() and removeProgFiles() respectively. Now openInfile() and openString() return a boolType result
to indicate the success.
- In libpath.c the functions find_include_file(), append_to_lib_path(), init_lib_path() and free_lib_path() have been renamed to findIncludeFile(), appendToLibPath(), initLibPath() and freeLibPath() respectively. The functions initIncludeFileHash(),
shutIncludeFileHash() and openIncludeFile() have been added. The added functions maintain a hashmap of already included files.
- In prclib.c the function prc_include() has been adjusted to call the new function that avoids double includes. Now the 2nd parameter ofthe action PRC_INCLUDE contains the file name to be included.
- In striutl.c the functions stri_to_os_utf8(), conv_to_os_stri() have been improved to return a boolType result that indicates success.
- Logging functions have been added to strlib.c.
- Documentation comments have been improved in cc_conf.s7i, html.s7i, osfiles.s7i, scanfile.s7i, scanstri.s7i, cmdlib.c, cmd_rtl.c, hshlib.c and hsh_rtl.c.
Changelog of the second to last release (2021-09-04):
- The linking of PostgreSQL has been improved. Many thanks go to SiliconWizard, for pointing out linking problems and for helping to investigate them. The function findPgTypeH() has been added to chkccomp.c. Now the search for pg_type.h and pg_type_d.h
does not include postgres.h. In sql_post.c the include of the file postgres.h has been removed.
- The wiz.sd7 example program has been refactored. Now it can be compiled. Many thanks go to Vasiliy Tereshkov, for reporting the compilation problem. Additionally, several improvements of wiz.sd7 have been done.
- The functions expm1() and log1p() have been added to the math.s7i library. Many thanks go to Sanjay Jain for pointing out that they were missing.
- In wrinum.s7i the functions str(ENGLISH, number) and str(GERMAN, number) have been improved to work correctly for zero.
- In forloop.s7i the definition of for-until-loops has been improved, such that the loop variable never gets a value outside of the range. The definition of for-loops has been changed to invoke the loop-body just at one place. Since the loop body is
inlined, this does shorten the generated code.
- Tests for for-loops have been added to chkprc.sd7.
- The compiler (s7c.sd7) has been improved to generate better code for the actions BLN_TERNARY, REF_ADDR, REF_SELECT and SET_ELEM (changes were done in bln_act.s7i, ref_act.s7i and set_act.s7i).
- The compiler has been improved (in comp/enu_act.s7i), to check for a possible RANGE_ERROR, if an integer is converted to an enumeration value (action ENU_ICONV2).
- The compiler has been improved to optimize expressions like ord(aBigExpression mod aPowerOfTwo).
- The function chkBigOrdWithBigMod has been added to chkbig.sd7. This function checks the optimizations done with expressions like ord(aBigExpression mod aPowerOfTwo).
- Tests for the ternary operator have been added to chkstr.sd7.
- Tests for the 'element in bitset' operator have been added to chkset.sd7. These tests check the compiler optimizations for SET_ELEM.
- Definitions of HAS_EXPM1 and HAS_LOG1P have been added to cc_conf.s7i.
- Interpreter and compiler have been improved, to support the actions HAS_EXPM1 and HAS_LOG1P.
- In comp/intrange.s7i the function getIntRange() has been improved to consider the actions INT_SUCC, INT_PRED, INT_ICONV1, INT_ICONV3 and SET_RAND. The handling of the actions INT_RAND, INT_ABS and INT_NEGATE has been improved. The functions
getIntAddRange() and getSetRandRange() have been added.
- The program chk_all.sd7 has been adjusted to the changes in the check programs.
- A spelling error in s7c.sd7 has been fixed.
- The program wrinum.sd7 has been changed to start with zero.
- Logging functions have been added to reflib.c.
Seed7 Homepage: http://seed7.sourceforge.net
Seed7 - The extensible programming language: User defined statements
and operators, abstract data types, templates without special
syntax, OO with interfaces and multiple dispatch, statically typed,
interpreted or compiled, portable, runs under linux/unix/windows.