emileclarkb/hammer
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
1109 commits 2 branches 0 tags 3.4 MiB
Commit graph

771 commits

Author SHA1 Message Date
Sven M. Hallberg
73e92f04df save kmax in HLLkTable 2015-09-06 14:48:41 +02:00
Sven M. Hallberg
8995097a1d let's just use ((void *)-1) as another invalid pointer result 2015-09-05 22:43:49 +02:00
Sven M. Hallberg
90b6f30fa1 add a test that needs to suspend on lookahead in the middle of a chunk 2015-09-04 21:06:50 +02:00
Sven M. Hallberg
1276004250 handle suspend on lookahead at the very end of the chunk 2015-09-04 21:05:56 +02:00
Sven M. Hallberg
010a1a36ff int_range test does not work for regex backend 2015-09-04 16:48:01 +02:00
Sven M. Hallberg
d4f933b2d3 move last_chunk flag into HInputStream 2015-09-04 12:55:37 +02:00
Sven M. Hallberg
28fa93d4cc make overrun flag a bool to match its usage 2015-09-04 12:54:00 +02:00
Sven M. Hallberg
d3fd0da7b7 don't skip tests on compile failure anymore 2015-09-04 12:48:10 +02:00
Sven M. Hallberg
932c9717d7 oops, forgot about the lookahead 2015-09-04 12:34:34 +02:00
Sven M. Hallberg
10fde548ad add a test for iterative parsing 2015-09-03 19:58:39 +02:00
Sven M. Hallberg
0231dc141e add iterative API to LL(k) backend 2015-09-03 19:18:07 +02:00
Sven M. Hallberg
f1d6d0bc5e split h_llk_parse into start/chunk/finish internally 2015-09-03 19:15:57 +02:00
Sven M. Hallberg
2845a9391e add API and backend hooks for iterative (chunked) parsing 2015-09-03 15:03:01 +02:00
Sven M. Hallberg
c5f2c5aa2b add a missing end-of-input check in the LL(k) backend 2015-08-25 14:14:48 +02:00
Nicolas Léveillé
9557448ae6 Introduce {v,}asprintf for windows 
We have now again two functions called h_platform_vasprintf and
h_platform_asprintf. On windows they are implemented in terms of vsnprint
and the like. On BSD/GNU libraries we use the supplied vasprintf and
asprintf.
 2015-08-16 16:46:52 +02:00
Nicolas Léveillé
124b4c381b regex_debug.c: Stop including dlfcn.h 
Since the block of code has been disabled anyway and is not very
portable, we introduce a define to control both the include
and the code.
 2015-08-16 16:21:21 +02:00
Nicolas Léveillé
e74e23d672 Change system_allocator to store size of blocks only for debugging 
Storing the size of the blocks is now only activated when DEBUG__MEMFILL
has been defined.
 2015-08-16 09:15:20 +02:00
Nicolas Léveillé
d9ae097533 Remove void pointer arithmetics in system_allocator.c 
We replace instances of void* w/ char* arithmetics.

This allows system_allocator.c to be compiled with MSVC.

Also we introduce a struct to represent the size_t header.
 2015-08-16 09:10:04 +02:00
Nicolas Léveillé
01333964bd Substitute spaces for tabs 2015-08-15 15:43:46 +02:00
Nicolas Léveillé
2346239887 Implement h_platform_stopwatch for Windows 
We use QueryPerformanceCounter which will return realtime, not user
time.
 2015-08-15 15:43:35 +02:00
Nicolas Léveillé
739f5d6d5c Split platform specific time measurement away 
We now have a `stopwatch` to measure time between two moments.
 2015-08-15 15:43:22 +02:00
Nicolas Léveillé
0c35525e30 Convert usages of errx to h_platform_errx 
This BSD call will have to be implemented on windows. Right now
we have a stub that just exits the process.
 2015-08-09 22:46:45 +02:00
Nicolas Léveillé
c8726244ae Wrap remaining __attribute__ 
We wrap all remaining references to __attribute__ into a macro call to
H_GCC_ATTRIBUTE, to allow compilation with MSVC CL.EXE
 2015-08-09 22:46:00 +02:00
Nicolas Léveillé
6b9ba6033e Allow windows users to include hammer.h 
We substitute H_GCC_ATTRIBUTE to any occurence of __attribute__ and
disable its function with non gcc, non clang compilers.

We use a H_ macro so as to not pollute users' namespaces.

On windows build without c4201 so as to allow anonymous struct. This
shows hammer is not C99 compliant but rather C11.
 2015-08-09 22:46:00 +02:00
Meredith L. Patterson
0fb9d77e40 remove error from HSVMContext, make svm_stack_ensure_cap return bool instead 2015-08-02 23:42:38 +02:00
Meredith L. Patterson
759fbe77e5 C++ bindings needed to use system_allocator's free() too 2015-08-02 22:03:30 +02:00
Meredith L. Patterson
6bffacefaf fix memory leak introduced by putting setjmp in the wrong place 2015-08-02 21:46:57 +02:00
Meredith L. Patterson
1ffd8d9276 Fix undefined behaviour around system_allocator, issue #133. 2015-08-02 21:32:47 +02:00
Meredith L. Patterson
0653a9e48a add explicit null checks to system_allocator malloc/realloc 2015-08-02 14:34:45 +02:00
Nicolas Léveillé
2a6233b74f Print user defined token type while pretty printing 
The pretty printing format chosen is USER:<token name>

We use the registry function `h_get_token_type_name` which returns 0
when no name has been defined. This will print (null) as a token name.
 2015-08-01 13:42:43 +02:00
Meredith L. Patterson
39c9096fd6 changes from tutorial: add h_literal, make h_make_bytes more practical 2015-05-04 19:00:37 -04:00
Meredith L. Patterson
26373bf25f Merge pull request #129 from pesco/h_bind 
Add monadic bind combinator h_bind
 2015-03-10 21:27:24 +01:00
Meredith L. Patterson
af260bdd58 Merge pull request #122 from mrdomino/fix-res_unamb-leak 
Free res_unamb
 2015-03-10 18:06:01 +01:00
Sven M. Hallberg
aa2dc7bcf5 put the message into the assert in aa_realloc 2015-03-04 14:54:42 +01:00
Sven M. Hallberg
be80fdedfc make arena-allocator implementation functions static 2015-03-04 14:52:28 +01:00
Sven M. Hallberg
9a8dca2c04 Merge branch 'master' of git://github.com/UpstandingHackers/hammer into h_bind 2015-03-04 14:47:49 +01:00
Sven M. Hallberg
1ce0ef9b8a pass a temporary arena allocator to the continuation in h_bind 2015-02-23 14:10:29 +01:00
Sven M. Hallberg
2eaf5d9052 add some missing type conversions to make h_read_bits work with count>32 2015-02-12 01:40:45 +01:00
Sven M. Hallberg
7ba5931095 Merge branch 'master' of git://github.com/UpstandingHackers/hammer into h_bind 2015-01-30 20:48:49 +01:00
Sven M. Hallberg
12771a5655 oh, there's t_regression.c now, neat. 2015-01-30 20:42:11 +01:00
Sven M. Hallberg
3f197ea1b0 Merge branch 'master' of git://github.com/UpstandingHackers/hammer into fix-h_seq_index_path 2015-01-30 20:37:26 +01:00
Sven M. Hallberg
341123e2ea descend properly in h_seq_index_path 2015-01-30 20:36:37 +01:00
Sven M. Hallberg
cbc7358b66 add a test that shows a bug in h_seq_index_path 2015-01-30 20:35:35 +01:00
Sven M. Hallberg
c9b2957832 add g_check_cmp_int 2015-01-30 20:35:14 +01:00
Sven M. Hallberg
78266af453 allow h_bind's function argument to return NULL as a shortcut for h_nothing_p() 2015-01-23 22:06:43 +01:00
Sven M. Hallberg
42d51ed479 add monadic bind combinator h_bind 2015-01-23 22:00:59 +01:00
Steven Dee
a0104776ed Free res_unamb 2015-01-04 21:38:04 -05:00
Meredith L. Patterson
5efaba87e8 Merge pull request #120 from thequux/fix-bug118 
Fix #118
 2015-01-04 04:18:11 +01:00
TQ Hirsch
af73181cf4 Fix #118 
NEWS:
* Switching endianness mid-byte no longer potentially re-reads bytes.
* bit_offset now consistently refers to the number of bits already
  read.
* HParsedTokens now have a bit_length field; this is a size_t.  This
  may be removed for memory reasons.

The bit writer has not yet been updated to match; the result of
switching bit writer endianness in the middle of a byte remains
undefined.
 2015-01-04 04:08:01 +01:00
Steven Dee
2dad0c48b4 H_VALIDATE_APPLY macro 
I've found this especially useful in combination with my own
_attr_uint_const for things like flags and type specifiers.

It's possible that its usefulness might be diminished significantly if
there were a built-in bitfield constant parser -- that certainly would
eliminate all of my current uses of it -- but it still seems nicely
symmetric with H_ACT_APPLY.
 2015-01-03 16:44:36 -05:00