When using any optimisation flag that wasn't O1 (optimize for size), the UI would be malformed. After about two weeks straight of doing this, Bug 1197316 - remove PR_snprintf calls in xpcom/ was found to be the first reason Waterfox would quit straight after being launched. So I went to Mozilla's mercurial repository and started downloading each changelog between 42->43 and compiling each one until it would break.
#Download firefox 43.0.1 Patch
I did the only thing I found plausible at the time.and that was to rummage through every patch between Firefox 42 -> Firefox 43 ( changelog here).Īs you can see.no clear indicator of what would be the contributions of Firefox 42 compiling fine with IC++ but Firefox 43 to not. After applying the relevant patches to get IC++ to build Firefox, the application would exit straight away. So.lets get Waterfox 43 done and out the gate, we've got a week until release I'm sure I can get it done since I've gotten 42 to build! Oh boy was I wrong. After a week of messing around I realised I could just use the std::tuple to replace Mozilla's Tuple calls and it worked! But alas, Firefox 43 was finalised and I though to myself if I released Waterfox 42 this late.what would be the point? Almost impossible to follow where this was occurring or why.Īs mentioned in a previous blog post Intel has a compiler bug causing an internal error with the use of Mozilla's own Tuple header. Debugging wasn't fruitful as almost every variable would read or.
So with the latest Firefox 43 codebase, Intel's C++ has been causing havoc, creating a difficult time for me.įor the last two months since the release of 40.1.0, I've been working every single day (I wish I was exaggerating.) waking up and going to bed doing nothing but trying to get Intel's C++ compiler to build Firefox. Just a delay on the installers until the bugs get fixed.
#Download firefox 43.0.1 portable
Waterfox 43 for Mac, Portable Edition and existing Waterfox/Firefox users unaffected.
0 kommentar(er)
