Remove --int-literals option #9597
Conversation
|
Is it mentioned on the man page or in the docs? getting-started and I think one or two other places mention (some of the) command-line args but have gotten out of date. |
|
Was in the first section of manual - thanks. |
|
also Lines 123 to 125 in 96d97ca and malformed tree failure on travis, oh boy |
| } | ||
| else { | ||
| return (jl_value_t*)jl_box_int64(i64); | ||
| } |
There was a problem hiding this comment.
Whoa, deleted way too much here. All this code should remain, just with different platform ifdefs.
|
Also need to change the |
| jl_compileropts.int_literals!=64 | ||
| #endif | ||
| ) { | ||
| if (i64 > (int64_t)S32_MAX || i64 < (int64_t)S32_MIN) |
There was a problem hiding this comment.
This should always be used on 32-bit.
|
Also the option help text in repl.c. |
| { "compile", required_argument, 0, 303 }, | ||
| { "depwarn", required_argument, 0, 304 }, | ||
| { "inline", required_argument, 0, 305 }, | ||
| { "dump-bitcode", required_argument, 0, 301 }, |
There was a problem hiding this comment.
Not sure what happened.. I rebased with --whitespace=fix and apparently
something was misconfigured. In any case, I will wait to rebase this until
after Jake's compileropts PR is merged.
On Mon, Jan 5, 2015 at 3:04 PM, Tony Kelman [email protected]
wrote:
In ui/repl.c
#9597 (diff):@@ -92,11 +92,10 @@ void parse_opts(int argcp, char **argvp)
{ "track-allocation",required_argument, 0, 'm' },
{ "check-bounds", required_argument, 0, 300 },
{ "optimize", no_argument, 0, 'O' },
{ "int-literals", required_argument, 0, 301 },{ "dump-bitcode", required_argument, 0, 302 },{ "compile", required_argument, 0, 303 },{ "depwarn", required_argument, 0, 304 },{ "inline", required_argument, 0, 305 },{ "dump-bitcode", required_argument, 0, 301 },no tabs!
—
Reply to this email directly or view it on GitHub
https://github.com/JuliaLang/julia/pull/9597/files#r22485714.
|
Rebased. |
|
If there are no objections I'll merge this later today. |
|
I kind of liked having this - for testing - mostly just saw it was broken.. 32-bit VM should do.. and then Julia code should work for sure (as in physical). And another thing: My switchable literal type idea was shot down. For integer (or possibly for other types, say, strings - but not Decimal Float-literals..), should switching on runtime be possible as an alternative? |
Did I miss anything? ref #9567