Mercurial > hg > xemacs-beta
annotate src/README @ 5911:48386fd60fd0
GMP functions that take doubles choke on non-finite values, avoid that.
src/ChangeLog addition:
2015-05-10 Aidan Kehoe <kehoea@parhasard.net>
* floatfns.c (double_to_integer):
Rename this from float_to_int to fit our newer, bignum-compatible
terminology.
GMP can signal SIGFPE when asked to turn NaN or infinity into a
bignum, and we're not prepared to handle that signal if the OS float
library routines don't do that, so check for those values
explicitly.
* floatfns.c (ceiling_two_float):
* floatfns.c (ceiling_one_float):
* floatfns.c (floor_two_float):
* floatfns.c (floor_one_float):
* floatfns.c (round_two_float):
* floatfns.c (round_one_float):
* floatfns.c (truncate_two_float):
* floatfns.c (truncate_one_float):
Call double_to_integer() with its new name.
* number.c:
Don't use the {bignum,ratio,bigfloat}_set_double functions
directly here, with GMP they can choke when handed non-finite C
doubles, call Ftruncate() and the new float_to_bigfloat() from
floatfns.c. Maybe we should extend number-gmp.c with GMP-specific
implementations that check for non-finite values.
tests/ChangeLog addition:
2015-05-10 Aidan Kehoe <kehoea@parhasard.net>
* automated/lisp-tests.el:
Backslash a few parentheses in the first column for the sake of
fontification.
* automated/lisp-tests.el:
Check that the rounding functions signal Lisp errors correctly
when handed positive and negative infinity and NaN.
| author | Aidan Kehoe <kehoea@parhasard.net> |
|---|---|
| date | Sun, 10 May 2015 19:07:09 +0100 |
| parents | 56144c8593a8 |
| children |
| rev | line source |
|---|---|
| 428 | 1 This directory contains the source files for the C component of XEmacs. |
| 2 Nothing in this directory is needed for using XEmacs once it is built | |
| 3 and installed, if the dumped Emacs is copied elsewhere. | |
| 4 | |
| 5 See the files ../README and then ../INSTALL for installation instructions. | |
| 6 | |
| 7 Under Unix, the file `Makefile.in.in' is used as a template by the script | |
| 8 `../configure' to produce `Makefile.in'. The same script then uses `cpp' | |
| 9 to produce the machine-dependent `Makefile' from `Makefile.in'; | |
| 10 `Makefile' is the file which actually controls the compilation of | |
| 11 Emacs. Most of this should work transparently to the user; you should | |
| 12 only need to run `../configure', and then type `make'. | |
| 13 | |
| 14 General changes for XEmacs: | |
| 15 --------------------------- | |
| 16 1. Lisp objects. | |
| 17 | |
| 18 -- XFASTINT has been eliminated. Use of this expression as an lvalue | |
| 19 is incompatible with the union form of Lisp objects, and use as | |
| 20 an rvalue is likely to lead to errors and doesn't really save much | |
| 21 time. Expressions of the form `XFASTINT (obj) = num;' get replaced | |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4953
diff
changeset
|
22 by `obj = make_fixnum (num);' or `XSETINT (obj, num);' and |
| 428 | 23 expressions of the form `num = XFASTINT (obj);' get replaced by |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4953
diff
changeset
|
24 `num = XFIXNUM (obj);'. Use Qzero in place of `make_fixnum (0)'. |
| 428 | 25 |
| 26 -- Use of XTYPE gets replaced by the appropriate predicate. Using | |
| 27 XTYPE only works for the small number of types that are not stored | |
| 28 using the Lisp_Record type (int, cons, string, and vector). For | |
| 29 example, `(XTYPE (foo) == Lisp_Buffer)' gets replaced by | |
| 30 `(BUFFERP (foo))'. | |
| 31 | |
| 32 -- `XSET (obj, Lisp_Int, num)' gets replaced by `XSETINT (obj, num)', | |
| 33 for consistency. | |
| 34 | |
| 35 -- Some occurrences of XSET need to get replaced by XSETR -- | |
| 36 specifically, those where the type is not a primitive type | |
| 37 (primitive types are int, cons, string, and vector). | |
| 38 | |
| 39 -- References to `XSTRING (obj)->size' get replaced with | |
| 40 `XSTRING_LENGTH (obj)'. This is currently for cosmetic reasons | |
| 41 but there may be other reasons in the future. (This change is | |
| 42 currently incomplete in the source files.) | |
| 43 | |
| 44 | |
| 45 2. Storage classes: | |
| 46 | |
| 47 -- All occurrences of `register' should be replaced by `REGISTER'. | |
| 48 It interferes with backtraces so we disable it if DEBUG_XEMACS | |
| 49 is defined. | |
| 50 | |
| 51 | |
| 52 3. Errors, messages, I18N3 snarfing: | |
| 53 | |
| 54 -- Errors are continuable in XEmacs but are not in FSF Emacs. | |
| 55 Therefore, it's important that functions do something reasonable | |
| 56 if an error gets continued. If you want to signal a non- | |
| 57 continuable error, the call to Fsignal() gets put inside a | |
| 58 `while (1)' loop. To facilitate this, and also for proper I18N3 | |
| 59 message snarfing, most calls to Fsignal() have been replaced by | |
| 60 calls to signal_error(), signal_simple_error(), etc. Look at | |
| 61 eval.c for a classification of various error functions. | |
| 62 | |
| 63 -- Constant strings occurring in source files need to get wrapped | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
442
diff
changeset
|
64 in a call to GETTEXT (or if inside of a call to `build_ascstring', |
| 428 | 65 change that function to `build_translated_string') if they don't |
| 66 occur in certain places where the I18N3 message snarfer will see | |
| 67 them. For a complete discussion of this, see the file | |
| 68 lib-src/make-msgfile.lex. | |
| 69 | |
| 70 NOTE: I18N3 support is not currently working, so the above may | |
| 71 or may not apply. Thus it is not a good idea to add random | |
| 72 GETTEXTs, unless you really know what you are doing. | |
| 73 | |
| 74 -- Calls to `fprintf (stderr, ...)' and `printf (...)' get replaced | |
| 75 with calls to `stderr_out' and `stdout_out'. This is for I18N3 | |
| 76 message snarfing. | |
| 77 | |
| 78 4. Initialization: | |
| 79 | |
| 80 -- FSF constructs like `obj = intern ("string"); staticpro (&obj);' | |
| 81 get replaced by `defsymbol (&obj);'. This is for code cleanness | |
| 82 and better purespace usage. | |
| 83 -- FSF constructs like | |
| 84 obj = intern ("error"); | |
| 85 Fput (obj, Qerror_message, "message"); | |
| 86 Fput (obj, Qerror_conditions, some list); | |
| 87 get replaced by calls to deferror(). See the definition of | |
| 88 deferror() for how the correct arguments to pass. This is for | |
| 89 code cleanness and I18N3 message snarfing. | |
| 90 -- Code in keys_of_foo() functions has been moved into Lisp. |