paweld/fpc
1
0
Fork 0
mirror of https://gitlab.com/freepascal.org/fpc/source.git synced 2025-04-29 12:23:46 +02:00
Code Issues Packages Projects Releases Wiki Activity
66,167 Commits 100 Branches 75 Tags 301 MiB
3c2703787c
Commit Graph

284 Commits

Author SHA1 Message Date
Jonas Maebe
a6c7ed5e61 vector support: use vector result types 
Set the result of a vector add node to a vector type.

Ensure that that these vector types are still asignment-compatible with
regular array types (may want to change this when we expose vector types),
and don't remove type conversions from vector types to array types (so the
code generator pass can still rely on the resultdef being a vector type)
 2022-09-16 11:21:04 +02:00
Sven/Sarah Barth
2372a06926 * fix #39740: consider a conversion of a procedure or procedure variable to a function reference as a bit more expensive than otherwise 
+ added test
 2022-05-31 22:50:14 +02:00
Sven/Sarah Barth
492754ecca * the check for nested procvars is now done during the conversion in capturer_add_procvar_or_proc so loosen the check in proc_to_funcref_equal_internal 2022-05-31 22:50:14 +02:00
Sven/Sarah Barth
4361e36ce0 * make proc_to_funcref_conv publicly available 2022-05-31 22:50:14 +02:00
Sven/Sarah Barth
5a680f0148 * fix #39745: the conversion from ordinals to pointer types in mode Delphi is not allowed for a void type (happens when using a typecast to convert a method without parameters to a function reference) 
+ added test
 2022-05-30 23:05:45 +02:00
Pierre Muller
bb39ec1610 Avoid invalid typecast to tabstractvarsym 2022-05-27 12:32:53 +00:00
Sven/Sarah Barth
7974f39522 * implement assignment of procedure and method variables and routines (global, instance and nested) to function references 2022-05-26 21:44:24 +02:00
Sven/Sarah Barth
76df7144ba * implement necessary conversion functionality for anonymous procdefs to function reference interfaces including capturing of variables 2022-05-26 21:44:23 +02:00
Sven/Sarah Barth
9b9ae2db1a * two function references are equal to each other as long as their signatures match 2022-05-26 21:44:23 +02:00
Sven/Sarah Barth
6fe9a15094 + add function to check whether a procdef can be converted to a function reference 2022-05-26 21:44:22 +02:00
Sven/Sarah Barth
47d7520b32 * consider invokables as related when they're equal 2022-05-26 21:43:39 +02:00
Sven/Sarah Barth
145b37425d + add function funcref_equal to check whether two function references are considered equal enough to be assignable 2022-05-26 21:43:39 +02:00
Sven/Sarah Barth
2be8f01efe * implement assignment of anonymous functions to procedure or method variables if they either capture nothing or (in case of method variables) at most the Self variable 2022-05-26 21:43:01 +02:00
Sven/Sarah Barth
17514ed5c0 * allow to ignore Self parameters; for this the parameter skipping needs to be repeated to correctly catch all cases 2022-05-26 21:43:01 +02:00
Sven/Sarah Barth
8e2478e632 * check whether an anonymous function can be assigned to a global, method or nested function variable 2022-05-26 21:43:01 +02:00
florian
6556d18bed * better fix for #39634 which avoids breaking existing code 2022-04-22 22:08:46 +02:00
Sven/Sarah Barth
094a353d87 + add ability to strictly compare defs with generic constraints (this is needed for declarations, while for normal code we're rather relaxed) 2022-02-18 17:53:35 +01:00
florian
aec18c2426 * weight currency->float conversions the same regardless if 
the currency type is handled by the integer unit or the x87 fpu,
    resolves #38309

git-svn-id: trunk@48089 -
 2021-01-05 22:39:16 +00:00
florian
1266afc0d0 * in equal_genfunc_paradefs take care of the fact that typesym might not be assigned for open array parameters, resolves #38012 
git-svn-id: trunk@47253 -
 2020-10-29 21:00:03 +00:00
svenbarth
2ff18e48a0 * fix for Mantis #36909: apply patch by Ryan Joseph so that array constructors can be used for static arrays as well. 
+ added test

git-svn-id: trunk@46891 -
 2020-09-18 15:00:15 +00:00
Jonas Maebe
23956e9f10 * prevent quadratic complexity in the size of the inheritance chain when 
comparing objectdefs for equality

git-svn-id: trunk@46774 -
 2020-09-05 12:25:13 +00:00
svenbarth
fa0c9adbf4 * extract checking whether two parameter defs of two procdefs are equal into a separate function 
git-svn-id: trunk@45972 -
 2020-07-31 15:55:31 +00:00
yury
72dc2145e9 * In Delphi mode throw a conversion error when trying to typecast a pointer to a procdef when they use different memory models. 
git-svn-id: trunk@45856 -
 2020-07-25 17:13:29 +00:00
svenbarth
52678562e3 * correctly handle selection between Single and Double overloads for Currency 
+ added test

git-svn-id: trunk@45707 -
 2020-06-27 22:33:36 +00:00
svenbarth
eaa2a2006c * fix for Mantis #37272: also allow the empty set for dynamic array parameters; Delphi probably started to allow that together with array constructors. 
+ added tests

git-svn-id: trunk@45706 -
 2020-06-27 22:33:31 +00:00
yury
2808873d1b * Reworked the optimization of unused $parentfp for nested routines. 
- Do not remove the $parentfp parameter as was done in the previous optimization approach. Instead when $parentfp is unused to the following:
      - On the caller side: Omit passing the value for $parentfp for targets where tcgcallparanode.push_zero_sized_value_para=false (classic CPU targets). 
          Pass 0/nil as $parentfp for targets where tcgcallparanode.push_zero_sized_value_para=true;
      - On the callee side: Prevent allocation of registers/resources for $parentfp.
  - When possible keep $parentfp in a register.
  - Set the pio_nested_access flag in tprocinfo.set_needs_parentfp() to properly handle deep nesting levels;

git-svn-id: trunk@45436 -
 2020-05-19 13:17:47 +00:00
yury
40504a6f9d * If a nested procedure does not access its parent's frame pointer, optimize it by removing the hidden $parentfp parameter. 
* Improved the tisogoto1.pp test.

git-svn-id: trunk@45292 -
 2020-05-06 14:43:49 +00:00
svenbarth
60345366f2 * fix for Mantis #35140: apply patch by Ryan Joseph together with some further changes by me to add support for constant parameters in generics 
+ added tests

git-svn-id: trunk@45080 -
 2020-04-25 22:12:35 +00:00
florian
fd20dc1086 * support unique copies of currencies which are handles by torddef, resolves #36013 
git-svn-id: trunk@42837 -
 2019-08-26 19:31:48 +00:00
Jonas Maebe
956aab3be0 * implement support for the Objective-C "related result type" convention 
as described on
    http://releases.llvm.org/8.0.0/tools/clang/docs/LanguageExtensions.html#objective-c-features
    (rest of mantis #35994)

git-svn-id: trunk@42816 -
 2019-08-25 15:23:53 +00:00
svenbarth
44bfa98a30 * fix for Mantis #35955: when the element of an array constructer requires an operator for conversion we don't let the caller of compare_defs_ext know that, instead we simply say that some conversion is required and let the typecheck handler deal with the element wise conversion 
+ added test

git-svn-id: trunk@42700 -
 2019-08-15 14:33:00 +00:00
Jonas Maebe
1da43f67d4 * (non-ancient) Delphi versions that accept explicit ordinal -> floating 
point typecasts behave the same as FPC (convert the value) rather than
    as Turbo Pascal (reinterpret the bit pattern) (mantis #35886)

git-svn-id: trunk@42507 -
 2019-07-28 11:14:54 +00:00
Jonas Maebe
c038e4c3f2 * relaxation of r42272: again only show warnings rather than errors for 
out-of-range constants, because of the comments mentioned in #35753
    (except for enums, as apparently Delphi does the same)
  * added range check warnings about explicit type casts that throw away
    bits (e.g. byte($fff)), without giving warnings for most common cases
    (like cardinal(-1))
  * fixed masking/sign exting constant array indices (must be based on index
    range type size/signedness rather than on array size/"signedness")

git-svn-id: trunk@42275 -
 2019-06-23 14:12:33 +00:00
Jonas Maebe
77658b925b * disable regular array -> dynamic array type coversion support unless 
{$modeswitch arraytodynarray} is active (mantis #35576)
   o changed compiler to compile without this modeswitch
   o added the modeswitch to a test that depends on it

git-svn-id: trunk@42118 -
 2019-05-25 12:31:32 +00:00
Jonas Maebe
281b3ad276 * fix case completeness and unreachable code warnings in compiler that would 
be introduced by the next commit

git-svn-id: trunk@42046 -
 2019-05-12 14:29:03 +00:00
florian
925531b99c * constrained generics are also compatible with formaldefs, resolves #34893 
git-svn-id: trunk@40921 -
 2019-01-19 21:23:36 +00:00
Jonas Maebe
50ab607676 + customint torddef type to create arbitraty bit-width integers 
o use this to handle non-power-of-two-sized parameters for llvm
   o no general support in the parser/code generator, so don't expose

git-svn-id: trunk@40398 -
 2018-11-29 20:57:08 +00:00
Jonas Maebe
9bd931e931 * make "formal const/var" parameters the least preferred conversion 
(mantis #32179)
  - removed code to handle conversion to formaldef parameters that are not by
    reference (so the default conversion preference is kept)

git-svn-id: trunk@40012 -
 2018-10-21 21:02:17 +00:00
Jonas Maebe
2c7c0d1144 * handle widechar constants directly in the scanner, instead of in the 
overload selection (where you can't even know whether the string is
    a valid widechar constant) (mantis #33875)

git-svn-id: trunk@40009 -
 2018-10-21 17:34:00 +00:00
Jonas Maebe
0b246f3dbd * converted Boolean8 to an internal type, and mapped Boolean to the 
new internal pasbool1(type) (part of mantis #34411)
   o apply the _Bool x86-64 parameter passing rules only to pasbool1

git-svn-id: trunk@39949 -
 2018-10-16 21:14:18 +00:00
florian
eea088c5f5 * allow nil to be assigned to generic types, resolves #34037 
* niln is also a constant node

git-svn-id: trunk@39934 -
 2018-10-14 07:38:13 +00:00
Jonas Maebe
1f9d518c57 * support for non-ASCII widechar constants (mantis #31605) 
* also improved type checking for converting constant strings to integers
    in MacPas mode

git-svn-id: trunk@38730 -
 2018-04-10 19:24:15 +00:00
nickysn
85ca504f4a * make near and far procedures/procvars incompatible with each other 
git-svn-id: trunk@38645 -
 2018-03-29 15:39:54 +00:00
svenbarth
5971e1327e * fix for Mantis #31756: have array constructors prefer open array parameters instead of dynamic array parameters for backwards compatibility 
+ added test

git-svn-id: trunk@36175 -
 2017-05-10 21:01:23 +00:00
svenbarth
bdef19fd79 * indeed allow the conversion of array constructors to dynamic arrays 
git-svn-id: trunk@36100 -
 2017-05-04 21:54:27 +00:00
svenbarth
df893b4a30 * implement a type conversion from an array constructor to a dynamic array 
git-svn-id: trunk@36095 -
 2017-05-04 21:34:38 +00:00
maciej-izak
ff28d5c85d * Fix usage of class operators in generics types. If we have generic type declaration/implementation then constraint in comparison to not specialized generic is not "exact" nor "incompatible". Mantis #30534. 
git-svn-id: trunk@35740 -
 2017-04-05 13:47:21 +00:00
svenbarth
5fa181b5d9 * second fix for Mantis #31033: also check for df_specialization instead of is_specialization for parameters to correctly handle nested types that aren't really true specializations themselves 
+ added additional test

git-svn-id: trunk@35103 -
 2016-12-10 19:47:09 +00:00
Jonas Maebe
94d7a7274b * compare floating point default parameter values bytewise instead of as 
floating point values, so that NaNs can also be compared (mantis #30299)

git-svn-id: trunk@34597 -
 2016-10-02 12:56:49 +00:00
svenbarth
0500c678e5 * fix for Mantis #30534: don't consider all types with generic constraints as equal, but at least ensure that they have the same basetype (Note: this still needs a bit more improvement) 
+ added test

git-svn-id: trunk@34526 -
 2016-09-16 13:24:28 +00:00