mirrors/ecl
1
Fork 0
mirror of https://gitlab.com/embeddable-common-lisp/ecl.git synced 2026-02-14 19:30:57 -08:00
Code Issues Projects Releases Packages Wiki Activity
8413 commits 16 branches 53 tags 56 MiB
Commit graph

8413 commits

This branch
This branch
All branches
Author SHA1 Message Date
Marius Gerbershagen
77bf1a7c90 Merge branch 'cmp-cleanups' into 'develop' 
cmp: various cleanups

See merge request embeddable-common-lisp/ecl!289
 2023-06-11 13:23:49 +00:00
Daniel Kochmański
4ca79933d3 Merge branch 'thread-timeouts' into 'develop' 
Various changes related to time issues

See merge request embeddable-common-lisp/ecl!290
 2023-06-11 12:58:59 +00:00
Daniel Kochmański
e10c2010d4 cosmetic: update headers 2023-06-11 10:40:21 +02:00
Daniel Kochmański
95e7bdd7d7 cmpc: change compiler-macro cl:float to c-inliner macro 
The compiler macro expanded float to c-inline.
 2023-06-11 10:40:21 +02:00
Daniel Kochmański
4a1902658c cmpc: move sysfun to the cxx backend 
sysfun declarations revolve strictly around c function inlining that is specific
to the C backend.

Moreover be more explicit about symbol packages and check feature-conditioned
inlines at runtime (not at readtime) in case that we construct the inline
information for a cross-compiled target. This should be further improved.
 2023-06-11 10:40:18 +02:00
Daniel Kochmański
8bb0b99499 cmpc: move c-inline expansions of PRINC to the backend 
Instead of using DEFINE-COMPILER-MACRO we implement them with DEFINE-C-INLINER
because these optimizations are backend-specific.
 2023-06-06 22:34:08 +02:00
Daniel Kochmański
c4b32c5905 cmpc: move produce-inline-loc and co to cmpc-inliner 
These operators are used across different optimizations and clearly do not
belong to a specific pass.
 2023-06-06 15:02:51 +02:00
Daniel Kochmański
ff67ebfa1a cmpc: move wt "call" routines to cmpc-wt 2023-06-06 14:48:58 +02:00
Daniel Kochmański
ae6614ebba cmpc: move variables related to code generation to the backend 2023-06-06 14:48:58 +02:00
Daniel Kochmański
220b7c7363 cmpc: move constant and inlining c/c++ optimizations 
Move appropriate files to the C/C++ backend.
 2023-06-06 14:48:58 +02:00
Daniel Kochmański
03be475fb9 cmpc: clean up numeric optimizers and propagators 
- merge bits and numeric optimizers in a single file
- move c/c++ optimizers to the backend
- move bits and numeric propagators to a separate file
 2023-06-06 14:48:58 +02:00
Daniel Kochmański
1472bb18e6 cmpc: move c/c++ backend to a separate directory 2023-06-06 14:48:58 +02:00
Daniel Kochmański
e9f05ac85e cmp: c1constant-value: remove defunct parameter ONLY-SMALL-VALUES 
After inclusion of the externalizable NAN values the parameter
':ONLY-SMALL-VALUES T' is the same as ':ALWAYS NIL' (the default).

That makes the the operator slightly easier to understand.
 2023-06-06 14:48:58 +02:00
Daniel Kochmański
d62cf434c6 cmp: enable externalizing NAN 2023-06-06 14:48:58 +02:00
Daniel Kochmański
3565c89ca9 cmp: cosmetic: write (function ,foo) instead of #',foo 2023-06-06 14:48:58 +02:00
Marius Gerbershagen
b049695ec0 ecl_musleep: get rid of unused alertable argument 
Was always set to true.
 2023-06-04 19:19:16 +02:00
Marius Gerbershagen
3f7581e884 time.d: change the internal time units to microseconds 
There is really no reason not to use microseconds. The operating
system APIs we are calling usually report time in microseconds anyway,
so as a bonus there is even less time spent rounding.
 2023-06-03 19:25:12 +02:00
Marius Gerbershagen
51627c7e9b get-internal-real-time: use monotonic clock if available 
Prevents jumps in the returned value if the system time is reset.
 2023-06-03 19:04:53 +02:00
Marius Gerbershagen
c300f2ee3a threads.h: use clock_monotonic for timeouts on posix if available 
Prevents threads from timing out too early or too late if the system
clock is adjusted while waiting.
 2023-05-27 18:36:16 +02:00
Marius Gerbershagen
c1078cda7d threads.h: add fallback if gettimeofday is not defined 
Probably unnecessary but since we have the same fallback in
get-internal-real-time we should add it here too for consistency.
 2023-05-27 16:41:57 +02:00
Daniel Kochmański
0321023572 cmp: propagate types in functions too 
The propagator for FSET did not descend to the function object while the
propagator for LOCALS did (that was inconsistent). Also cmpmain called
P1PROPAGATE on *LOCAL-FUNS* before the second pass so that was no-op.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
1e9786fd0a cmp: simplify the type propagation pass 
Previously we've carries the "assumption" list through all calls, but said
assumption list was never used to make any decisions (and we had numerous
assertions that it must be null at various places).

The assumption list made the code less readable because it intorduced numerous
loops and multiple-value-bind calls just to maintain it. Removal introduces no
known regressions.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
18030bf1b4 cmp: rearrange the order in +all-c1-forms+ to hint top-levelness 2023-05-24 15:49:06 +02:00
Daniel Kochmański
0ccb877b8a cmp: remove a kludge from the function object-type 
The function OBJECT-TYPE special-cased NIL and returned SYMBOL as its
type. git-blame shown that this kludge was part of the initial release so I
could not find any reason why it does not return NULL instead. Removing the
kludge did not yield any noticeable regressions.

My guess is that the type system was deficient at that time and type-of did not
work on the type NIL. Returning SYMBOL instead of NULL lead to a sloppy type
propagation of locations containing NIL.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
e23c82b90d cmp: cosmetic: don't use nconc in ctop-write 
We simply iterate over both lists without nconcing them.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
aae21dc9ae cmp: return-from: fix a problem where block-name was never used 
The intention of c2return-from was to print the block name in the error
message when the block was referenced outside of its dynamic extent, but CCB
was never passed as the return type. The code is simplified to always use the
same code path for non-local returns.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
b8aa49a053 cmp: add a helper function GET-OBJECT for the data segment 
This function may be used to get the object from one of the segments or fail
-- most notably it does not add the object to the data segment.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
244f4e0485 cmp: separate cmpbackend from cmpmain (2) 
Introduce compiler-pass/assemble-cxx.
 2023-05-24 15:49:06 +02:00
Daniel Kochmański
3f3c89ddb2 cmp: separate cmpbackend from cmpmain (1) 
Introduce compiler-pass/generate-cxx.
 2023-05-24 15:49:03 +02:00
Daniel Kochmański
1a5d1c6ca4 .gitignore: add the directory /local as ignored 2023-05-22 10:16:39 +02:00
Marius Gerbershagen
fceb54c237 replace some deprecated macro names 2023-05-21 19:04:43 +02:00
Jeronimo Pellegrini
1c1db8637c Make (log x 0) return 0 (as per the spec) 
The CL spec says that "If base is zero, log returns zero."  But ECL
was not checking for zero, so a divide by zero error was generated.
This patch makes it return zero when the base is zero:

> (log  1 0)

0
> (log 1/2 0)

0
> (log #C(2 1) 0)

0
> (log 2 0.0)

0.0
 2023-04-15 15:30:04 +02:00
Marius Gerbershagen
8cffaf858c numbers/log.d: fix typo 2023-03-25 13:46:43 +01:00
Marius Gerbershagen
8586876051 tests: multiprocessing: increase timeout value in semaphore smoke tests 
This makes the test run more reliably on slow computers.
 2023-03-18 16:41:41 +01:00
Marius Gerbershagen
4705967061 cmp: remove superfluous encode-string function 2023-03-18 15:18:20 +01:00
Marius Gerbershagen
acd81e16cd allow non-ascii characters in debug info 2023-03-18 15:18:10 +01:00
Marius Gerbershagen
ae19006cb8 Merge branch 'cmpwerk' into 'develop' 
Separate COMPILER and EXT,FFI,MP packages

See merge request embeddable-common-lisp/ecl!285
 2023-03-12 15:31:36 +00:00
Marius Gerbershagen
0c09ded44d doc: be more precise about error cases for ext:terminate-process 2023-03-12 16:07:42 +01:00
Marius Gerbershagen
6845a9a480 Merge branch 'external-processes-encoding' into 'develop' 
Some improvements for ext:run-program

Closes #607

See merge request embeddable-common-lisp/ecl!287
 2023-03-12 14:31:39 +00:00
Daniel Kochmański
0a3c4f53be cosmetic: fix a typo 2023-03-09 13:56:48 +01:00
Daniel Kochmański
eb580f4084 doc: fix a spelling mistake s/:signalled/:signaled/ for the status 2023-03-07 17:13:03 +01:00
Daniel Kochmański
f160a6e07e tests: run-program: don't call terminate-process excessively 
Calling terminate-process on a process that has already finished has undefined
consequences.
 2023-03-07 17:10:16 +01:00
Daniel Kochmański
08c48b5346 doc: describe calling terminate-process on a finished process 
Consequences of such action are undefined. This is prompted by the fact that
windows will signal an error in this case.
 2023-03-07 17:01:14 +01:00
Marius Gerbershagen
c637d9d41a ext:string-to-octets: avoid allocating unnecessarily large output array 
If start and end are given we may only need a smaller array. Also add
1 for possible null terminator.
 2023-02-26 17:37:48 +01:00
Marius Gerbershagen
dd29e35966 tests: 2am: if (finishes) fails due to a signal, print the error 2023-02-26 17:05:12 +01:00
Marius Gerbershagen
115b52f9f5 terminate-process: improve error message 2023-02-26 17:05:12 +01:00
Marius Gerbershagen
828b2b482a tests: run-program: fix terminate-process test on Windows 2023-02-26 17:05:12 +01:00
Marius Gerbershagen
c52ec9fe9e tests: implement process-environ test on windows 2023-02-26 17:03:09 +01:00
Marius Gerbershagen
02ef05479c ext:run-program: support unicode characters for the process name, arguments and environment 
The encoding is determined by ext:*default-external-format* as usual.
 2023-02-26 17:03:09 +01:00
Marius Gerbershagen
53388175f3 unixsys: better error messages if we fail to spawn a subprocess 2023-02-26 16:56:43 +01:00