- 08 Apr, 2015 33 commits
-
-
Håvard Haugen authored
When decoding an invalid typeId the associated *decEngine was not removed from decoderMap. If the decoder was run again on the same input a nil *decEngine was found in the map and assumed to be initialized, resulting in a panic. Fixes #9649 Change-Id: I5bb51808362a21c09228c2705a658f073e5b59b3 Reviewed-on: https://go-review.googlesource.com/3509Reviewed-by:
Rob Pike <r@golang.org>
-
Robert Griesemer authored
There was no way to get to the error message before. Change-Id: I4aa9d3d9f468c33f9996295bafcbed097de0389f Reviewed-on: https://go-review.googlesource.com/8660Reviewed-by:
Alan Donovan <adonovan@google.com>
-
Rob Pike authored
Fixes #10379. Change-Id: Ia4cdda36ed57a06371f9ace7365ce9e215228487 Reviewed-on: https://go-review.googlesource.com/8654Reviewed-by:
-
Rob Pike authored
Just an oversight. Plus the code had an unnecessary call to os.Exit that now has a purpose. Fixes #10372. Change-Id: I456018f3a01ca05b4501c7f8a4961d48ab8c5e16 Reviewed-on: https://go-review.googlesource.com/8651Reviewed-by:
Minux Ma <minux@golang.org>
-
Shenghou Ma authored
Change-Id: Ied1582d8aabee2eb346e1c23bfd7781e4a091264 Reviewed-on: https://go-review.googlesource.com/8621Reviewed-by:
Robert Griesemer <gri@golang.org>
-
Robert Griesemer authored
Disable importer-dependent tests on platforms for which the respective builders don't have access to importable packages. Fixes #10368. Change-Id: I8072c59d2bbbc24a43d771fd04fd0b1a678d765a Reviewed-on: https://go-review.googlesource.com/8611Reviewed-by:
-
Ian Lance Taylor authored
Change-Id: Iaf2dba7d699a8d52f91ce10222ab0d1a0f1f21fc Reviewed-on: https://go-review.googlesource.com/8625Reviewed-by:
Josh Bleecher Snyder <josharian@gmail.com> Reviewed-by:
-
Michael Hudson-Doyle authored
This CL is quite conservative in some ways. It continues to define symbols that have no real purpose (e.g. epclntab). These could be deleted if there is no concern that external tools might look for them. It would also now be possible to make some changes to the pcln data but I get the impression that would definitely require some thought and discussion. Change-Id: Ib33cde07e4ec38ecc1d6c319a10138c9347933a3 Reviewed-on: https://go-review.googlesource.com/7616 Run-TryBot: Ian Lance Taylor <iant@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
Ian Lance Taylor <iant@golang.org>
-
Marko Tiikkaja authored
Previously if the connection pool was larger than maxBadConnRetries and there were a lot of bad connections in the pool (for example if the database server was restarted), a query might have failed with an ErrBadConn unnecessarily. Instead of trying to guess how many times to retry, try maxBadConnRetries times and then force a fresh connection to be used for the last attempt. At the same time, lower maxBadConnRetries to a smaller value now that it's not that important to retry so many times from the free connection list. Fixes #8834 Change-Id: I6542f151a766a658980fb396fa4880ecf5874e3d Reviewed-on: https://go-review.googlesource.com/2034 Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org> Reviewed-by:
Brad Fitzpatrick <bradfitz@golang.org>
-
Martin Möhrmann authored
Change function appendUint to appendInt with variable-width 0-padding. This allows the decimal for the year to be generated without extra code to handle the wider padding and directly handles negative numbers. Removes the special casing for numbers with one and two digits. The special case for 0 was unreachable. The new version is slightly slower. benchmark old ns/op new ns/op delta BenchmarkFormat 444 454 +2.25% BenchmarkFormatNow 398 415 +4.27% Change-Id: I4ddef96bf07ad35dca76053321d510441ec6d4f5 Reviewed-on: https://go-review.googlesource.com/2751Reviewed-by:
-
Dave Cheney authored
Update #10203 When the portable componentgen was introduced in b9602632 it produced broken code on arm64 and was deactivated. In the month since it looks like the underlying issues have been fixed so componentgen produces working binaries that are slightly smaller, ~3kb reduction in size for godoc binary. Benchmarks are underwhelming, but where visible, trending towards an improvement (this is with Minux's peep optimiser CL). benchmark old ns/op new ns/op delta BenchmarkBinaryTree17 15336842000 15002766000 -2.18% BenchmarkFannkuch11 10848984000 10896931000 +0.44% BenchmarkFmtFprintfEmpty 203 188 -7.39% BenchmarkFmtFprintfString 753 720 -4.38% BenchmarkFmtFprintfInt 667 670 +0.45% BenchmarkFmtFprintfIntInt 1103 1102 -0.09% BenchmarkFmtFprintfPrefixedInt 981 969 -1.22% BenchmarkFmtFprintfFloat 1396 1348 -3.44% BenchmarkFmtManyArgs 4151 4102 -1.18% BenchmarkGobDecode 34202360 32933020 -3.71% BenchmarkGobEncode 27579180 27438820 -0.51% BenchmarkGzip 1296119000 1285096000 -0.85% BenchmarkGunzip 291099800 289727200 -0.47% BenchmarkHTTPClientServer 169476 169803 +0.19% BenchmarkJSONEncode 70313600 69973400 -0.48% BenchmarkJSONDecode 227811800 232875200 +2.22% BenchmarkMandelbrot200 12985600 12996430 +0.08% BenchmarkGoParse 14708930 14507320 -1.37% BenchmarkRegexpMatchEasy0_32 464 460 -0.86% BenchmarkRegexpMatchEasy0_1K 4516 4517 +0.02% BenchmarkRegexpMatchEasy1_32 452 454 +0.44% BenchmarkRegexpMatchEasy1_1K 4664 4669 +0.11% BenchmarkRegexpMatchMedium_32 602 602 +0.00% BenchmarkRegexpMatchMedium_1K 172939 172494 -0.26% BenchmarkRegexpMatchHard_32 9733 9577 -1.60% BenchmarkRegexpMatchHard_1K 301356 298143 -1.07% BenchmarkRevcomp 2754334000 2753874000 -0.02% BenchmarkTemplate 315664000 311810800 -1.22% BenchmarkTimeParse 1034 989 -4.35% BenchmarkTimeFormat 1118 1137 +1.70% benchmark old MB/s new MB/s speedup BenchmarkGobDecode 22.44 23.31 1.04x BenchmarkGobEncode 27.83 27.97 1.01x BenchmarkGzip 14.97 15.10 1.01x BenchmarkGunzip 66.66 66.98 1.00x BenchmarkJSONEncode 27.60 27.73 1.00x BenchmarkJSONDecode 8.52 8.33 0.98x BenchmarkGoParse 3.94 3.99 1.01x BenchmarkRegexpMatchEasy0_32 68.92 69.47 1.01x BenchmarkRegexpMatchEasy0_1K 226.71 226.65 1.00x BenchmarkRegexpMatchEasy1_32 70.75 70.42 1.00x BenchmarkRegexpMatchEasy1_1K 219.55 219.28 1.00x BenchmarkRegexpMatchMedium_32 1.66 1.66 1.00x BenchmarkRegexpMatchMedium_1K 5.92 5.94 1.00x BenchmarkRegexpMatchHard_32 3.29 3.34 1.02x BenchmarkRegexpMatchHard_1K 3.40 3.43 1.01x BenchmarkRevcomp 92.28 92.29 1.00x BenchmarkTemplate 6.15 6.22 1.01x Change-Id: I0b7d95388d6920fcbd7fe305df0c7c630a407726 Reviewed-on: https://go-review.googlesource.com/8636Reviewed-by:
-
Dave Cheney authored
Add end to end tests for arm64 to support CL 8405. There are several instruction forms commented out at the moment they will be addressed in CL 8405 or later followups. Change-Id: I6eeeb810c1e03cd49bb3c881bc46a29cdb817822 Reviewed-on: https://go-review.googlesource.com/8631Reviewed-by:
Aram Hăvărneanu <aram@mgk.ro> Reviewed-by:
-
David Crawshaw authored
The tests for go/types depend on reading gc export data from the $GOROOT/pkg directory. This is the first use of these files as testdata, so previously they were not copied to the android device. Now they are used, copy them. Fixes android/arm build. Change-Id: If13bbe603ce0aff697a73a97ae9a7d6b3ea800f9 Reviewed-on: https://go-review.googlesource.com/8624Reviewed-by:
-
Michael Matloob authored
The return type for bsdthread_register is int32. See runtime/os_darwin.go. This change also rewrites declaration comments for go functions to use go syntax and fixes vet errors in sys_darwin_amd64.s. Change-Id: I7482105f7562929e0ede30099efac9e76babd8a3 Reviewed-on: https://go-review.googlesource.com/3260Reviewed-by:
-
Péter Surányi authored
"returns ... the first error" was misleading or at least confusing: in case a Read results in an error with non-zero bytes read, and the subsequent Write also results in an error, the error from Write is returned, which is the second one (in the temporal dimension). Fixes #9744 Change-Id: If8925a701e4fae820cd9df7446503403fc0785d4 Reviewed-on: https://go-review.googlesource.com/3686Reviewed-by:
-
Shenghou Ma authored
Change-Id: I1dcca264d9cc900aad2d7737073cc01fe574bf55 Reviewed-on: https://go-review.googlesource.com/8623Reviewed-by:
-
Shenghou Ma authored
Fixes #10107. Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f6 Reviewed-on: https://go-review.googlesource.com/8453Reviewed-by:
-
Shenghou Ma authored
Update #10373. Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f5 Reviewed-on: https://go-review.googlesource.com/8452Reviewed-by:
David Crawshaw <crawshaw@golang.org>
-
Shenghou Ma authored
Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f4 Reviewed-on: https://go-review.googlesource.com/8451Reviewed-by:
-
Shenghou Ma authored
Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f3 Reviewed-on: https://go-review.googlesource.com/8450Reviewed-by:
-
Shenghou Ma authored
Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f2 Reviewed-on: https://go-review.googlesource.com/8439Reviewed-by:
-
Shenghou Ma authored
This CL introduces R_ADDRARM64, which is similar to R_ADDRPOWER. Fixes #10112. Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f1 Reviewed-on: https://go-review.googlesource.com/8438Reviewed-by:
-
Shenghou Ma authored
Based on Michael Hudson-Doyle's patch: https://github.com/4ad/go/commit/b735215ee41b4237ec393d4669f6f55d4b27babc Change-Id: I309e3df7608b9eef9339196fdc50dedf5f9439f0 Reviewed-on: https://go-review.googlesource.com/8437Reviewed-by:
-
Shenghou Ma authored
Previously, the TestCompareAPI test would fail if runtime.Version() is "dev", or, more importantly, "go1.5"; because compareAPI depends on runtime.Version and -allow_new flag. Move that logic out make its test more robust. Change-Id: I8f40daa1838b8acd26adac8848762d95315053b0 Reviewed-on: https://go-review.googlesource.com/8622Reviewed-by:
-
Shenghou Ma authored
Based on cmd/9g/peep.go. Go 1 benchmark comparison: benchmark old ns/op new ns/op delta BenchmarkBinaryTree17 24328574000 18351639000 -24.57% BenchmarkFannkuch11 17029365000 10817758000 -36.48% BenchmarkFmtFprintfEmpty 291 223 -23.37% BenchmarkFmtFprintfString 1073 799 -25.54% BenchmarkFmtFprintfInt 1024 778 -24.02% BenchmarkFmtFprintfIntInt 1654 1277 -22.79% BenchmarkFmtFprintfPrefixedInt 1360 1083 -20.37% BenchmarkFmtFprintfFloat 2272 1415 -37.72% BenchmarkFmtManyArgs 5933 4742 -20.07% BenchmarkGobDecode 53166003 38584736 -27.43% BenchmarkGobEncode 37930156 30074874 -20.71% BenchmarkGzip 1880638900 1286832100 -31.57% BenchmarkGunzip 386343633 292194480 -24.37% BenchmarkHTTPClientServer 237077 179776 -24.17% BenchmarkJSONEncode 101731690 73116925 -28.13% BenchmarkJSONDecode 344655360 241277600 -29.99% BenchmarkMandelbrot200 28329778 12950809 -54.29% BenchmarkGoParse 21670755 16554244 -23.61% BenchmarkRegexpMatchEasy0_32 557 484 -13.11% BenchmarkRegexpMatchEasy0_1K 4687 4832 +3.09% BenchmarkRegexpMatchEasy1_32 539 483 -10.39% BenchmarkRegexpMatchEasy1_1K 5100 5080 -0.39% BenchmarkRegexpMatchMedium_32 796 651 -18.22% BenchmarkRegexpMatchMedium_1K 233099 182047 -21.90% BenchmarkRegexpMatchHard_32 13202 9897 -25.03% BenchmarkRegexpMatchHard_1K 401027 303602 -24.29% BenchmarkRevcomp 3837679666 2816546600 -26.61% BenchmarkTemplate 440608300 324831040 -26.28% BenchmarkTimeParse 1460 1019 -30.21% BenchmarkTimeFormat 1609 1174 -27.04% benchmark old MB/s new MB/s speedup BenchmarkGobDecode 14.44 19.89 1.38x BenchmarkGobEncode 20.24 25.52 1.26x BenchmarkGzip 10.32 15.08 1.46x BenchmarkGunzip 50.23 66.41 1.32x BenchmarkJSONEncode 19.07 26.54 1.39x BenchmarkJSONDecode 5.63 8.04 1.43x BenchmarkGoParse 2.67 3.50 1.31x BenchmarkRegexpMatchEasy0_32 57.38 66.05 1.15x BenchmarkRegexpMatchEasy0_1K 218.47 211.91 0.97x BenchmarkRegexpMatchEasy1_32 59.29 66.21 1.12x BenchmarkRegexpMatchEasy1_1K 200.76 201.54 1.00x BenchmarkRegexpMatchMedium_32 1.26 1.53 1.21x BenchmarkRegexpMatchMedium_1K 4.39 5.62 1.28x BenchmarkRegexpMatchHard_32 2.42 3.23 1.33x BenchmarkRegexpMatchHard_1K 2.55 3.37 1.32x BenchmarkRevcomp 66.23 90.24 1.36x BenchmarkTemplate 4.40 5.97 1.36x Fixes #10105. Change-Id: I353cc9fdf922e431821508c9dbbe4d9a85d64bd4 Signed-off-by:
Dave Cheney <dave@cheney.net>
-
Shenghou Ma authored
Fixes #10369. Change-Id: If0a6d2b33c6862c9f7f862bdc997f2204072c6dc Reviewed-on: https://go-review.googlesource.com/8620 Run-TryBot: Minux Ma <minux@golang.org> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by:
-
Shenghou Ma authored
Add test, and while we're at here, also add a test for ARM. Fixes #10343. Change-Id: Ic914df8233d4f1f495e2cc0743fbd37b7671bc91 Signed-off-by:
-
Robert Griesemer authored
Change-Id: I58de76c49de6e43befb30b2bf677934e7952b5ab Reviewed-on: https://go-review.googlesource.com/8610Reviewed-by:
-
Robert Griesemer authored
The old code checked out a specific version of go/types from the x/tools repo. With go/types being part of the std repo, this is not necessary anymore. Also, for the same reason, the api tool is now built like any other regular command. There's no need to build it for each run. Removed the respective +build tags. Change-Id: I5088e4867223d676957084c24651ec05452ac495 Reviewed-on: https://go-review.googlesource.com/8564Reviewed-by:
-
Robert Griesemer authored
Temporary work-around so we can start using go/types in the std repo. Change-Id: I661465af791171b01cd23abf34dcb7eea6e26173 Reviewed-on: https://go-review.googlesource.com/8594Reviewed-by:
-
Robert Griesemer authored
This is a first step towards moving go/types from the tools repo into the std repo. The files were brought over via the added src/go/types.bash script for reproducability. The script can be removed once all dependencies on go/types have moved to the std repo go/types. The script moved packages as follows: - x/tools/go/types => go/types (type-checker) - x/tools/go/exact => go/exact (constants) - x/tools/go/gcimporter => go/types/internal/gcimporter The gcimporter is needed to be able to run tests. go/types should probably have some factory function to provide an appropriate importer. Some of the go/types tests fail for a handful of platforms (windows and nacl). In order to keep this change "clean" from manual changes, the next change will disable those tests for now so we can move forward. Change-Id: I448d8f7faa39ad2e04811911b699f7682627c224 Reviewed-on: https://go-review.googlesource.com/8530Reviewed-by:
-
Nigel Tao authored
Change-Id: I3621527c924a43724032f80a072505c60d929ab3 Reviewed-on: https://go-review.googlesource.com/8180Reviewed-by:
-
Alex Brainman authored
Fixes #9753 Change-Id: I6c641ed7ef4f687a108e7d937ab4b9c24d5baf5d Reviewed-on: https://go-review.googlesource.com/4940Reviewed-by:
-
- 07 Apr, 2015 7 commits
-
-
Robert Griesemer authored
Change-Id: Iff943d15e83e7db1f3c77f509a60e00ee2041d69 Reviewed-on: https://go-review.googlesource.com/8592Reviewed-by:
-
Robert Griesemer authored
For very out-of-range floating-point constants (1e100000000), precise formatting of the offending value for error messages is not needed and potentially extremely slow. This change resurrects an adjusted variant of the original code which uses float64 formatting in the common case (in-range values), and a fast manual approximation for out-of-range values. Change-Id: I2f6e53040929b8bf924dac4bb27c4d811ede48e2 Reviewed-on: https://go-review.googlesource.com/8470Reviewed-by:
-
Michael Hudson-Doyle authored
Adds $DOLLAR to doc.go Change-Id: Ia0ccdd9b5292010297ae95debab4acf50d3c4d9c Reviewed-on: https://go-review.googlesource.com/8600Reviewed-by:
-
ALTree authored
Fixed bug that caused Exp(x, y, m) ( i.e. x**y (mod m) ) to return x instead of x (mod m) when y == 1. See issue page on github for more details. Added test case Fixes #9826 Change-Id: Ibabb58275a20c4231c9474199b7f1c10e54241ce Reviewed-on: https://go-review.googlesource.com/8409Reviewed-by:
-
Rob Pike authored
Fix the other places the slice length was being believed, and refactor the code to use a single function to unify the check. Fixes #10273. Change-Id: Ia62b25203fbe87c95d71a70ebc1db8d202eaa4a4 Reviewed-on: https://go-review.googlesource.com/8511Reviewed-by:
-
Dave Cheney authored
Add OGETG to the list of ignored operations. We don't instrument the runtime package, but calls to runtime.getg can appear in other packages, for example, after inlining runtime.LockOSThread. Change-Id: I8d6e91f1f3c8fd1302b596bdead42d588c059911 Reviewed-on: https://go-review.googlesource.com/8553Reviewed-by:
Dmitry Vyukov <dvyukov@google.com>
-
Dave Cheney authored
Add special case for OGETG which should never be instrumented. Change-Id: I7d082abb8608537f82b03362b687baf2a1d809dc Reviewed-on: https://go-review.googlesource.com/8551Reviewed-by:
-
