728115 - UINT32_MAX not defined

Status: RESOLVED INVALID

(Core :: General, defect)

Description

[lkcl]

User Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.19) Gecko/20110430 Iceweasel/3.5.9 (like Firefox/3.5.9)
Build ID: 20100414091236

Steps to reproduce:

jshashtable.h contains UINT32_MAX usage whereas all other header files (this is /usr/include/xulrunner-9/*.h on a debian system) use PR_UINT32_MAX


Actual results:

        JS_STATIC_ASSERT(((sMaxInit * sInvMaxAlpha) >> 7) < sMaxCapacity);
        JS_STATIC_ASSERT((sMaxCapacity * sInvMaxAlpha) <= UINT32_MAX);
        JS_STATIC_ASSERT((sMaxCapacity * sizeof(Entry)) <= UINT32_MAX);

Comment 1

[Boris Zbarsky [:bzbarsky]]

UINT32_MAX is defined in the system standard stdint.h header.  What exact Gecko version did you get your Spidermonkey from, and on what OS are you compiling?

Comment 2

[lkcl]

hi boris, thanks for responding so quickly.  this is xulrunner 9.0 as installed by
debian/testing (amd64)

(In reply to Boris Zbarsky (:bz) from comment #1)
> UINT32_MAX is defined in the system standard stdint.h header.

 ah that's strange because the code (hulahop) used to compile perfectly.

>  What exact
> Gecko version did you get your Spidermonkey from, and on what OS are you
> compiling?

 debian/amd64, and it's not spidermonkey, it's hulahop.  i've made some rather
 dreadful hacks:
 http://lkcl.net/hulahop/

 there appears to be a bit more involved than just putting in stdint.h because
 there are also other objects / defines missing: i've commented out the
 js_pop and js_push_context functions for now in the interests of "getting something
 working".

Comment 3

[Boris Zbarsky [:bzbarsky]]

> ah that's strange because the code (hulahop) used to compile perfectly.

Used to until what happened?

> debian/amd64, and it's not spidermonkey, it's hulahop.

If it's including Gecko it's certainly including SpiderMonkey (the JS engine).

> there appears to be a bit more involved than just putting in stdint.h

OK.  What files are actually failing to compile?  jshashtable back then appears not to include stdint (which is certainly a bug), but all the code that includes it seems to include stdint.  Do you have your own code which includes private JS headers?

Comment 4

[lkcl]

(In reply to Boris Zbarsky (:bz) from comment #3)
> > ah that's strange because the code (hulahop) used to compile perfectly.
> 
> Used to until what happened?

 i don't know, because there were so many other dependencies involved, we had
 to wait until todd finished porting python-xpcom to cope with the xulrunner 2
 changes.

 to define at least _something_ for you: this code used to work perfectly with
 xulrunner-1.9.1 (ok and i've just seen a comment that it was ok with 1.9.2 as well)


 
> > debian/amd64, and it's not spidermonkey, it's hulahop.
> 
> If it's including Gecko it's certainly including SpiderMonkey (the JS
> engine).

 ok.

> > there appears to be a bit more involved than just putting in stdint.h
> 
> OK.  What files are actually failing to compile?

 hulahop-web-view.c

 the relevant includes are:

#if 0
#include <jscntxt.h>
#endif

#if 0
void
hulahop_web_view_push_js_context(HulahopWebView *web_view)
{
    nsCOMPtr<nsIJSContextStack> stack(do_GetService("@mozilla.org/js/xpc/ContextStack;1"));
    nsresult rv = stack->Push(nsnull);
    g_assert(NS_SUCCEEDED(rv));
}

void
hulahop_web_view_pop_js_context(HulahopWebView *web_view)
{
    nsCOMPtr<nsIJSContextStack> stack(do_GetService("@mozilla.org/js/xpc/ContextStack;1"));
    nsresult rv = stack->Pop(nsnull);
    g_assert(NS_SUCCEEDED(rv));
}
#endif


>  jshashtable back then
> appears not to include stdint (which is certainly a bug), but all the code
> that includes it seems to include stdint.  Do you have your own code which
> includes private JS headers?

ah... scratch that: i've just recompiled, it's fine.

hmh.  wonder how that didn't work before? bizarre.

ok, stdint.h it is - thank you for your help, boris.