diff options
Diffstat (limited to 'source/luametatex/source/libraries')
8 files changed, 101 insertions, 97 deletions
diff --git a/source/luametatex/source/libraries/avl/readme.txt b/source/luametatex/source/libraries/avl/readme.txt index 90ff0949b..9b4708248 100644 --- a/source/luametatex/source/libraries/avl/readme.txt +++ b/source/luametatex/source/libraries/avl/readme.txt @@ -1,23 +1,23 @@ -Remark - -Usage of the avl library (irr) showed up in pdfTeX when Hartmut added some functionality. It therefore -also ended up in being used in LuaTeX. The two files avl.c and avl.h come from pyavl and are in the -public domain: - - license: this package, pyavl, is donated to the public domain - author : Richard McGraw - email : dasnar@fastmail.fm - -In the pdfTeX/LuaTeX the files were just there but I could track them down to - - https://github.com/pankajp/pyavl - -where the dates indicate that nothing has changed in the meantime. In the copies used here I added the -information mentioned above. The files had some (experimental) code as well as optional testing on NULL -values. As I don't expect updates (the code has been okay for quite a while) I made the tests mandate -and removed the experimental code. - -We can strip this library and save some 10K on the binary because we don't need that much of it. That -might happen at some point. - +Remark
+
+Usage of the avl library (irr) showed up in pdfTeX when Hartmut added some functionality. It therefore
+also ended up in being used in LuaTeX. The two files avl.c and avl.h come from pyavl and are in the
+public domain:
+
+ license: this package, pyavl, is donated to the public domain
+ author : Richard McGraw
+ email : dasnar@fastmail.fm
+
+In the pdfTeX/LuaTeX the files were just there but I could track them down to
+
+ https://github.com/pankajp/pyavl
+
+where the dates indicate that nothing has changed in the meantime. In the copies used here I added the
+information mentioned above. The files had some (experimental) code as well as optional testing on NULL
+values. As I don't expect updates (the code has been okay for quite a while) I made the tests mandate
+and removed the experimental code.
+
+We can strip this library and save some 10K on the binary because we don't need that much of it. That
+might happen at some point.
+
Hans Hagen
\ No newline at end of file diff --git a/source/luametatex/source/libraries/libcerf/readme-luametatex.txt b/source/luametatex/source/libraries/libcerf/readme-luametatex.txt index bb552f263..4ba0240e1 100644 --- a/source/luametatex/source/libraries/libcerf/readme-luametatex.txt +++ b/source/luametatex/source/libraries/libcerf/readme-luametatex.txt @@ -1,26 +1,26 @@ -LS, - -In the following files you can find the comment below. We don't want to bother or burden the -original authors with our problems. The cerf code is mostly used in MetaFun macros (by Alan -Braslau). The c.h and cpp.h files are gone. - - defs.h - cerf.h - ---------------------------------------------------------------------------------------------- -This file is patched by Mojca Miklavec and Hans Hagen for usage in LuaMetaTeX where we use -only C and also want to compile with the Microsoft compiler. So, when updating this library -one has to check for changes. Not that we expect many as this is a rather stable library. - -In the other files there are a few macros used that deal with the multiplication and addition -of complex and real nmbers. Of course the original code is kept as-is. ---------------------------------------------------------------------------------------------- - -So, when updating the library you need to diff for the changes that are needed in order to -compile the files with the Microsoft compiler. - -At some point I might patch the files so that we can intercept error messages in a way that -permits recovery and also plugs them into our normal message handlers. Maybe I should also -merge the code into just one file because it doesn't change. - -Hans +LS,
+
+In the following files you can find the comment below. We don't want to bother or burden the
+original authors with our problems. The cerf code is mostly used in MetaFun macros (by Alan
+Braslau). The c.h and cpp.h files are gone.
+
+ defs.h
+ cerf.h
+
+---------------------------------------------------------------------------------------------
+This file is patched by Mojca Miklavec and Hans Hagen for usage in LuaMetaTeX where we use
+only C and also want to compile with the Microsoft compiler. So, when updating this library
+one has to check for changes. Not that we expect many as this is a rather stable library.
+
+In the other files there are a few macros used that deal with the multiplication and addition
+of complex and real nmbers. Of course the original code is kept as-is.
+---------------------------------------------------------------------------------------------
+
+So, when updating the library you need to diff for the changes that are needed in order to
+compile the files with the Microsoft compiler.
+
+At some point I might patch the files so that we can intercept error messages in a way that
+permits recovery and also plugs them into our normal message handlers. Maybe I should also
+merge the code into just one file because it doesn't change.
+
+Hans
diff --git a/source/luametatex/source/libraries/miniz/readme.txt b/source/luametatex/source/libraries/miniz/readme.txt index 8a5e6979e..4527133d7 100644 --- a/source/luametatex/source/libraries/miniz/readme.txt +++ b/source/luametatex/source/libraries/miniz/readme.txt @@ -1,8 +1,8 @@ -Remark - -Conform the recommendation we use the official merged files (release) not the github files. Also, we -only use part of that single file because we do all file handling ourselves because we operate within -the file name regime of LuaMetaTeX that is aware of operating system specifics like wide filenames on -MSWindows). We don't drop in updates without careful checking them first for potential clashes.\\ - +Remark
+
+Conform the recommendation we use the official merged files (release) not the github files. Also, we
+only use part of that single file because we do all file handling ourselves because we operate within
+the file name regime of LuaMetaTeX that is aware of operating system specifics like wide filenames on
+MSWindows). We don't drop in updates without careful checking them first for potential clashes.\\
+
release url: https://github.com/richgel999/miniz/releases
\ No newline at end of file diff --git a/source/luametatex/source/libraries/pplib/ppapi.h b/source/luametatex/source/libraries/pplib/ppapi.h index e9ced5718..56137f8f2 100644 --- a/source/luametatex/source/libraries/pplib/ppapi.h +++ b/source/luametatex/source/libraries/pplib/ppapi.h @@ -8,7 +8,7 @@ #include "ppconf.h" -#define pplib_version "v2.1" +#define pplib_version "v2.2" #define pplib_author "p.jackowski@gust.org.pl" /* types */ diff --git a/source/luametatex/source/libraries/pplib/ppcrypt.c b/source/luametatex/source/libraries/pplib/ppcrypt.c index ce63e7cab..832e8c327 100644 --- a/source/luametatex/source/libraries/pplib/ppcrypt.c +++ b/source/luametatex/source/libraries/pplib/ppcrypt.c @@ -137,7 +137,7 @@ static int ppcrypt_password_encoding (uint8_t *password, size_t *passwordlength) { uint8_t *p, newpassword[PPCRYPT_MAX_PASSWORD], *n; const uint8_t *e; - uint32_t unicode; + uint32_t unicode = 0; for (n = &newpassword[0], p = &password[0], e = p + *passwordlength; p < e; ++n) { @@ -398,7 +398,10 @@ static ppcrypt_status ppcrypt_authenticate_permissions (ppcrypt *crypt, ppstring aes_decode_data(perms->data, perms->size, permsdata, crypt->filekey, crypt->filekeylength, nulliv, AES_NULL_PADDING); if (permsdata[9] != 'a' || permsdata[10] != 'd' || permsdata[11] != 'b') - return PPCRYPT_FAIL; + { /* if we get here, the password hash is correct, we don't need to fail because of unreadable perms (found such docs) */ + crypt->flags |= PPCRYPT_UNREADABLE_PERMISSIONS; + return PPCRYPT_DONE; + } /* do not check/update permissions flags here; they might be different inside crypt string */ if (0) diff --git a/source/luametatex/source/libraries/pplib/ppcrypt.h b/source/luametatex/source/libraries/pplib/ppcrypt.h index 9fa52d878..a7131adbb 100644 --- a/source/luametatex/source/libraries/pplib/ppcrypt.h +++ b/source/luametatex/source/libraries/pplib/ppcrypt.h @@ -46,6 +46,7 @@ typedef struct { #define PPCRYPT_STRING_RC4 (1<<4) #define PPCRYPT_STREAM_AES (1<<5) #define PPCRYPT_STRING_AES (1<<6) +#define PPCRYPT_UNREADABLE_PERMISSIONS (1<<7) #define PPCRYPT_STREAM (PPCRYPT_STREAM_AES|PPCRYPT_STREAM_RC4) #define PPCRYPT_STRING (PPCRYPT_STRING_AES|PPCRYPT_STRING_RC4) diff --git a/source/luametatex/source/libraries/pplib/readme.txt b/source/luametatex/source/libraries/pplib/readme.txt index ee5d141dc..550367a28 100644 --- a/source/luametatex/source/libraries/pplib/readme.txt +++ b/source/luametatex/source/libraries/pplib/readme.txt @@ -1,3 +1,3 @@ -This is (to be) added to util/utilflate.c: - -# include "../../utilities/auxzlib.h" +This is (to be) added to util/utilflate.c:
+
+# include "../../utilities/auxzlib.h"
diff --git a/source/luametatex/source/libraries/readme.txt b/source/luametatex/source/libraries/readme.txt index f249eae99..a4ccbfc0b 100644 --- a/source/luametatex/source/libraries/readme.txt +++ b/source/luametatex/source/libraries/readme.txt @@ -1,37 +1,37 @@ -Nota bene, - -The currently embedded libcerf library might become an optional one as soon as we decide to provide -it as such. It doesn't put a dent in filesize but as it's used rarely (and mostly as complement to -the complex math support) that makes sense. The library was added because some users wanted it as -companion the other math libraries and because TeX is often about math it sort of feels okay. But -it looks like there will never be support for the MSVC compiler. Mojca and I (Hans) adapted the -sources included here to compile out of the box, but that didn't make it back into the original. - -The pplib library has a few patches with respect to memory allocation and zip compression so that -we can hook in the minizip and mimalloc alternatives. - -The avl and hnj libraries are adapted to Lua(Meta)TeX and might get some more adaptations depending -on our needs. The decnumber library that is also used in mplib is unchanged. - -In mimalloc we need to patch init.c: #if defined(_M_X64) || defined(_M_ARM64) to get rid of a link -error as well as in options.c some snprint issue with the mingw64 cross compiler: - -/* HH */ snprintf(tprefix, sizeof(tprefix), "%sthread 0x%x: ", prefix, (unsigned) _mi_thread_id()); /* HH: %z is unknown */ - -In decNumber.c this got added: - -# include "../../utilities/auxmemory.h" -# define malloc lmt_memory_malloc -# define free lmt_memory_free - -In softposit/source/include/softposit_types.h we have to comment the initializations in the unions -bcause the compiler complains about it (we're not using c++). So: - -uint32_t ui; // =0; // patched by HH because the compilers don't like this -uint64_t ui[2]; // ={0,0}; // idem -uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idme -uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idem -uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idem - - +Nota bene,
+
+The currently embedded libcerf library might become an optional one as soon as we decide to provide
+it as such. It doesn't put a dent in filesize but as it's used rarely (and mostly as complement to
+the complex math support) that makes sense. The library was added because some users wanted it as
+companion the other math libraries and because TeX is often about math it sort of feels okay. But
+it looks like there will never be support for the MSVC compiler. Mojca and I (Hans) adapted the
+sources included here to compile out of the box, but that didn't make it back into the original.
+
+The pplib library has a few patches with respect to memory allocation and zip compression so that
+we can hook in the minizip and mimalloc alternatives.
+
+The avl and hnj libraries are adapted to Lua(Meta)TeX and might get some more adaptations depending
+on our needs. The decnumber library that is also used in mplib is unchanged.
+
+In mimalloc we need to patch init.c: #if defined(_M_X64) || defined(_M_ARM64) to get rid of a link
+error as well as in options.c some snprint issue with the mingw64 cross compiler:
+
+/* HH */ snprintf(tprefix, sizeof(tprefix), "%sthread 0x%x: ", prefix, (unsigned) _mi_thread_id()); /* HH: %z is unknown */
+
+In decNumber.c this got added:
+
+# include "../../utilities/auxmemory.h"
+# define malloc lmt_memory_malloc
+# define free lmt_memory_free
+
+In softposit/source/include/softposit_types.h we have to comment the initializations in the unions
+bcause the compiler complains about it (we're not using c++). So:
+
+uint32_t ui; // =0; // patched by HH because the compilers don't like this
+uint64_t ui[2]; // ={0,0}; // idem
+uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idme
+uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idem
+uint64_t ui[8]; // ={0,0,0,0, 0,0,0,0}; // idem
+
+
Hans
\ No newline at end of file |