Commit Graph

2920 Commits

Author SHA1 Message Date
peter
65a721de66 * test for wrong first match of operator overloading
git-svn-id: trunk@7703 -
2007-06-17 14:17:30 +00:00
daniel
013107d245 + Test for automated types.
git-svn-id: trunk@7693 -
2007-06-16 20:16:05 +00:00
Jonas Maebe
ce94d89bd9 + added cwstring unit for unix
git-svn-id: trunk@7679 -
2007-06-16 09:13:01 +00:00
Jonas Maebe
0c79628c8a * ifdef linux -> ifdef unix for cthreads
git-svn-id: trunk@7678 -
2007-06-16 08:48:13 +00:00
Jonas Maebe
62b3b307e0 * also process all subnodes of try-except and try-finally in
foreachnode(static) (fixes first example of mantis #9076)

git-svn-id: trunk@7671 -
2007-06-15 17:36:09 +00:00
Jonas Maebe
ba95cc22ee * only allow automatic type conversions of array constructors of
char to pchar/array of char, rather than of arbitrary array
    constructors (mantis #9085)

git-svn-id: trunk@7670 -
2007-06-15 17:16:44 +00:00
Jonas Maebe
2496946c23 * put tests in procedure so that an unbalanced stack will cause a crash
when exiting it

git-svn-id: trunk@7662 -
2007-06-14 16:50:08 +00:00
micha
2bcbebc438 * theapthread test: put the sleep in the right place to increase odds of freeing after other thread exit
git-svn-id: trunk@7651 -
2007-06-13 20:33:32 +00:00
micha
b667ea06ba * extend theapthread test to explicitly test freeing of block after owned-by thread exits
git-svn-id: trunk@7650 -
2007-06-13 20:29:08 +00:00
Jonas Maebe
ceaaa0e552 * fixed writing of widestrings
git-svn-id: trunk@7645 -
2007-06-13 13:00:25 +00:00
joost
4d9c3244fe * If a variant contains a datetime value and is converted to a string, it should return a string in a date/time format. (+test)
git-svn-id: trunk@7643 -
2007-06-13 12:24:16 +00:00
daniel
02732bfb55 * Fix tests for constexprint rework.
git-svn-id: trunk@7638 -
2007-06-13 07:30:09 +00:00
daniel
a0d601cfe7 * 12345678901234567890>high(int64), therefore int64var=12345678901234567890 always
evaluates to false. Remove 1 digit to get the value below high(int64).

git-svn-id: trunk@7636 -
2007-06-12 22:20:24 +00:00
Jonas Maebe
af97f3350b * fixed bitsize of qwords in bitpacked records (mantis #9054)
git-svn-id: trunk@7632 -
2007-06-11 09:11:42 +00:00
Jonas Maebe
b998cb1208 * changed so it still fails now that you can take the address of some
fields of bitpacked records

git-svn-id: trunk@7621 -
2007-06-10 11:55:05 +00:00
Jonas Maebe
c77d74127b * fixed some record size checks
(some are still broken, but can't be fixed currently because
    def.size/sym.getsize return an aint, which means low(aint)
    in case of structures with a size = high(aint)+1)

git-svn-id: trunk@7614 -
2007-06-09 19:48:14 +00:00
Jonas Maebe
10341eabaa * don't allow taking the address of ordinal bitpacked array elements
whose size is not a power of two
  + some more packed array/record tests

git-svn-id: trunk@7610 -
2007-06-09 18:13:04 +00:00
Jonas Maebe
2af8ca7a19 * allow taking the address of fields of packed records iff
* their offset is a multiple of 8 bits; and
   * their size is a multiple of 8 bits; and
   * if it's a range type, the size is also a power of 2

git-svn-id: trunk@7609 -
2007-06-09 17:52:20 +00:00
Jonas Maebe
20b99dade3 - reverted r7600, see additional comments in mantis #9027
git-svn-id: trunk@7606 -
2007-06-09 17:14:28 +00:00
Jonas Maebe
044e616bea * added packrecords statement so the test compiles on all systems regardless
of the default packrecords value

git-svn-id: trunk@7605 -
2007-06-09 17:05:35 +00:00
Jonas Maebe
5a8679131f + support for new/dispose with extra size parameter in macpas mode
(mantis #9027)

git-svn-id: trunk@7600 -
2007-06-08 18:02:31 +00:00
Jonas Maebe
de75b34dba * fixed another IE with indexing bitpacked arrays of composite types,
this time with constant rather than variable indices

git-svn-id: trunk@7599 -
2007-06-08 13:53:01 +00:00
Jonas Maebe
445d116b78 * added missing {$mode macpas}
git-svn-id: trunk@7598 -
2007-06-08 13:44:57 +00:00
Jonas Maebe
65e35ac401 * fixed internal error with bitpacked arrays of composite types whose
size is not a power of two

git-svn-id: trunk@7596 -
2007-06-08 09:11:08 +00:00
daniel
d6e6e39d16 * Minor changes.
git-svn-id: trunk@7580 -
2007-06-05 07:53:55 +00:00
daniel
023be72367 + Pascal conversion of Stream benchmark.
git-svn-id: trunk@7579 -
2007-06-05 07:06:01 +00:00
Jonas Maebe
8bcb6e689d * only check validity of assembler instructions in pass_2, after all
local operands have been resolved (mantis #8950)

git-svn-id: trunk@7516 -
2007-05-29 16:53:28 +00:00
Jonas Maebe
4ce4742bc8 * mark invisible high parameters of open arrays/arrays of const
of parent procedures which are implicitly accessed from within
    nested procedures for range checking purposes as non-regable in
    the resulttype pass of the nested procedure (as opposed to in its
    pass1/pass2, because by then the regvar assignment of the parent
    procedure is already finished) (mantis #8975)
  * related fix regarding checking whether the high parameter actually
    exists (must check calling convention of the procedure to which the
    high parameter belongs, which is not the same as checking that of
    the current procedure in case of nested procedures)

git-svn-id: trunk@7512 -
2007-05-29 13:42:49 +00:00
peter
0605a67968 * new implementation
git-svn-id: trunk@7500 -
2007-05-28 16:10:15 +00:00
Jonas Maebe
e1b7ce5aa9 * fixed dynarr_property:=nil + test
git-svn-id: trunk@7495 -
2007-05-28 13:48:32 +00:00
peter
b9a384ba08 * multiple fpc version support for shootout
git-svn-id: trunk@7494 -
2007-05-28 11:59:10 +00:00
Jonas Maebe
f12428506e + some more property assignment tests (which already/still work)
git-svn-id: trunk@7487 -
2007-05-27 10:57:24 +00:00
Jonas Maebe
fa5e232055 * fixed assigning an interface to a property after better property
checks

git-svn-id: trunk@7484 -
2007-05-27 10:25:46 +00:00
Jonas Maebe
17bc78aa65 + added
git-svn-id: trunk@7483 -
2007-05-27 10:05:03 +00:00
daniel
86ae9d5475 * Make it work under Linux.
git-svn-id: trunk@7482 -
2007-05-26 14:56:09 +00:00
micha
4b73505019 * make bansi1 bench multi threaded, in bansi1mt
git-svn-id: trunk@7459 -
2007-05-24 20:16:49 +00:00
micha
b51f0d5e20 * make blists1 bench more detailed
git-svn-id: trunk@7458 -
2007-05-24 20:16:14 +00:00
Jonas Maebe
03fe8d622a * skip test for cpus with alignment constraints
git-svn-id: trunk@7454 -
2007-05-24 12:06:35 +00:00
yury
cf19c0993a * fixed bug #8919.
+ test.

git-svn-id: trunk@7448 -
2007-05-24 08:15:28 +00:00
peter
cc5b22c0e1 * new test
git-svn-id: trunk@7423 -
2007-05-22 20:55:28 +00:00
micha
7d109a1a33 * prevent freeing in both threads at the same time in theapthread test
git-svn-id: trunk@7415 -
2007-05-21 20:39:33 +00:00
Jonas Maebe
df7ad9f68d * renamed so it gets tested by the makefile (all test programs
must start with t)

git-svn-id: trunk@7410 -
2007-05-21 12:06:50 +00:00
peter
b029477b8e * new tests
git-svn-id: trunk@7409 -
2007-05-21 07:54:18 +00:00
micha
7f2a257102 + heap manager now per thread, reduce heap lock contention
+ heap threading test

git-svn-id: trunk@7407 -
2007-05-20 20:58:12 +00:00
Jonas Maebe
d375ae6f7f * test now fails under FPC
git-svn-id: trunk@7401 -
2007-05-20 08:54:10 +00:00
Jonas Maebe
497df2bb37 * also check for properties passed to read(ln) when reading integers
different from the native size, or when reading enums (because those
    are handled via a temp internally -> regular var parameter checks
    were not automatically performed)

git-svn-id: trunk@7398 -
2007-05-19 18:53:24 +00:00
Jonas Maebe
f50eb8c9f6 * fixed 64 bit compilation
git-svn-id: trunk@7396 -
2007-05-19 17:24:11 +00:00
Jonas Maebe
a0b57eddb5 * new internal set format for big endian systems. Advantages:
* varsets ({$packset x}) are now supported on big endian targets
    * gdb now displays sets properly on big endian systems
    * cleanup of generic set code (in, include/exclude, helpers), all
      based on "bitpacked array[] of 0..1" now
  * there are no helpers available yet to convert sets from the old to
    the new format, because the set format will change again slightly
    in the near future (so that e.g. a set of 24..31 will be stored in
    1 byte), and creating two classes of set conversion helpers would
    confuse things (i.e., it's not recommended to use trunk currently for
    programs  which load sets stored to disk by big endian programs compiled
    by previous FPC versions)
  * cross-endian compiling has been tested and still works, but one case
    is not supported: compiling a compiler for a different endianess
    using a starting compiler from before the current revision (so first
    cycle natively, and then use the newly created compiler to create a
    cross-compiler)

git-svn-id: trunk@7395 -
2007-05-19 17:15:15 +00:00
pierre
320e5d7315 * improve Run Comparison
git-svn-id: trunk@7390 -
2007-05-18 14:18:44 +00:00
peter
55f79a37b6 * new test
git-svn-id: trunk@7382 -
2007-05-17 20:28:24 +00:00