-This file documents bugs which were "re-introduced" in various pl's
-pl69
- * slurs.ly
+This documents serious bugs
-pl68:
- * opening rests wohltemperirt
+********
+
+[Linux Intel, EGCS-1.1b]
+
+Compiling with -O2 -finline-functions will cause crashes during
+parsing stage.
+
+
+[All platforms]
+
+When dealing with beaming that is not correct (eg quarter notes in
+beams.), you can get the following assert. This is a serious bug, but
+a good solution is quite a lot of work.
+
+ \score{
+ \melodic{
+ [c2 c]
+ }
+ }
+
+results in
+
+ lilypond: ../flower/include/varray.hh:141: struct Rhythmic_grouping *& Array<Rhythmic_grouping *>::elem(int) const: Assertion `i >=0&&i<size_' failed.
+
+And this
+
+ \score{
+ \melodic{
+ [c]
+ }
+ }
+
+in
+
+ lilypond: ../flower/include/cursor.tcc:104: int Cursor<void *>::operator -(class Cursor<void *>) const: Assertion `c.ok()' failed.
+ Aborted (core dumped)
+
+
+[Linux libg++ 2.7]
+
+LilyPond occasionally crashes while parsing the initialisation files.
+This is a very obscure bug, and usually entering the commandline
+differently "fixes" it.
+
+ lilypond input.ly
+
+and
+
+ lilypond -I. ./input.ly
+
+makes a difference
+
+Typical stacktrace:
+
+ SIGSEGV
+ __libc_malloc (bytes=16384)
+ ?? ()
+ yyFlexLexer::yy_create_buffer ()
+ Includable_lexer::new_input (this=0x8209a00, s={strh_ = {
+ :
+
+This behaviour has been observed with machines that have old libg++
+versions (LinuxPPC feb '98, RedHat 4.x).
+
+
+
+[Linux Intel]
+
+A problem resembling the previous: usage of libg++.2.8.x with the
+wrong version of libc results in a coredump from the scanner while
+reading the init files. Stacktrace:
+
+ ios::eof (this=0x0)
- * lyrics broken,
+ yyFlexLexer::LexerInput (this=0x8294848, buf=0x82955f0 "", max_size=8192)
+ yyFlexLexer::yy_get_next_buffer (this=0x8294848)
+ My_lily_lexer::yylex (this=0x8294848)
- * midi broken
+Fix: follow the install instructions of libg++: match the right
+library versions.