aboutsummaryrefslogtreecommitdiff
path: root/src/analyzer.js
AgeCommit message (Expand)Author
2012-12-07properly count pointer sizes in structural typesAlon Zakai
2012-12-07legalize illegal parts of structural typesAlon Zakai
2012-12-07properly legalize returnAlon Zakai
2012-12-07properly decide when to legalize a call return valueAlon Zakai
2012-12-07standardize tempRet to start at 0Alon Zakai
2012-12-07generalize handling of zeroinitializer and undef in legalizer codeAlon Zakai
2012-12-07legalize insertvalueAlon Zakai
2012-12-07properly legalize storing an entire structure literalAlon Zakai
2012-12-07legalize loads and stores of entire structuresAlon Zakai
2012-12-07do not legalize landingpad, for nowAlon Zakai
2012-12-07legalize select structuralsAlon Zakai
2012-12-07fix bug when we legalize call params and assume return value is also in need ...Alon Zakai
2012-12-07start to legalize structural return values into one direct return plus the re...Alon Zakai
2012-12-07start to legalize structural typesAlon Zakai
2012-12-07shorten temp var nameAlon Zakai
2012-12-0764-bit dynamic shifts in asmAlon Zakai
2012-11-20commentAlon Zakai
2012-11-20warn on very large structural types that slow us downAlon Zakai
2012-11-12handle as best we can surprising types for legalized bitcasts. fixes #701Alon Zakai
2012-11-09remove dead codeAlon Zakai
2012-11-08initial hackingAlon Zakai
2012-11-06split js compiler into three passes, to facilitate future parallizationAlon Zakai
2012-10-26do not do signalyzer pass if sign correction is on - it can only hurt thereAlon Zakai
2012-10-12remove some phi warnings by defaultAlon Zakai
2012-08-29fix extension of non-i32s to >i32sAlon Zakai
2012-08-24check for additional stack allocations and varargs in the whole function, not...Alon Zakai
2012-07-16handle i64 icmp using processMathOp codeAlon Zakai
2012-07-16handle illegal icmps; fixes #514Alon Zakai
2012-05-14allow nonexistent phis in another location + add testAlon Zakai
2012-05-14allow nonexistend phi labels, llvm generates unneeded ones sometimesAlon Zakai
2012-05-12simply compiler by always pushing back phisAlon Zakai
2012-05-11fix bug with missing entry labels being misused between original labels and n...Alon Zakai
2012-04-19zext sign correction for bitcasting from <32 bits to i64Alon Zakai
2012-03-02commentAlon Zakai
2012-03-02minify label ids to numbers early in the analysisAlon Zakai
2012-03-01emit smaller labels for loopsAlon Zakai
2012-02-29fix i64 invoke args, and parsing of very large i64 (and other illegal) constantsAlon Zakai
2012-02-28handle invoke that returns i64Alon Zakai
2012-02-27properly handle illegal phi literalsAlon Zakai
2012-02-27legalize switch i64Alon Zakai
2012-02-26remove I64_MODE, it is equal to 1 in USE_TYPED_ARRAYS 2 and only thereAlon Zakai
2012-02-25support more mathops in legalizerAlon Zakai
2012-02-25refactor out paramX in mathopsAlon Zakai
2012-02-20generate signed legalized 32-bit chunksAlon Zakai
2012-02-18basic support for setjmp/longjmpAlon Zakai
2012-02-14handle the case of the initial alloca being a variable oneAlon Zakai
2012-02-11do not try to legalize bitcasts between doubles and floatsAlon Zakai
2012-02-11legalize invokeAlon Zakai
2012-02-11fix typo in select legalizationAlon Zakai
2012-02-11fix bug with i64 return valuesAlon Zakai