delta/go-git.git/src/archive/zip, branch dev.inline

archive/zip: skip large concurrent tests in race mode

2016-11-21T20:51:42+00:00

We recently added these large zip64 tests. They're slow-ish already,
but fast enough in non-race mode with t.Parallel. But in race mode,
the concurrency makes them much slower than the normal
non-race-to-race multiplier.

They're taking so long now that it's causing test failures when it
sometimes is over the test timeout threshold.

Change-Id: I02f4ceaa9d6cab826708eb3860f47a57b05bdfee
Reviewed-on: https://go-review.googlesource.com/33423
Run-TryBot: Brad Fitzpatrick 
TryBot-Result: Gobot Gobot 
Reviewed-by: Ian Lance Taylor

archive/zip: avoid overflow in record count and byte offset fields

2016-11-17T18:54:33+00:00

This is Quentin's https://golang.org/cl/33012 with updated tests.

Fixes #14186

Change-Id: Ib51deaab0368c6bad32ce9d6345119ff44f3c2d6
Reviewed-on: https://go-review.googlesource.com/33291
Reviewed-by: Quentin Smith 
Run-TryBot: Brad Fitzpatrick 
TryBot-Result: Gobot Gobot

archive/zip: only use Extended Timestamp on non-zero MS-DOS timestamps

2016-10-11T21:31:25+00:00

We should preserve the fact that a roundtrip read on fields with the zero
value should remain the zero for those that are reasonable to stay that way.
If the zero value for a MS-DOS timestamp was used, then it is sensible for
that zero value to also be read back later.

Fixes #17403

Change-Id: I32c3915eab180e91ddd2499007374f7b85f0bd76
Reviewed-on: https://go-review.googlesource.com/30811
Run-TryBot: Joe Tsai 
TryBot-Result: Gobot Gobot 
Reviewed-by: Brad Fitzpatrick

archive/zip: handle mtime in NTFS/UNIX/ExtendedTS extra fields

2016-10-06T19:05:52+00:00

Handle NTFS timestamp, UNIX timestamp, Extended extra timestamp.
Writer supports only Extended extra timestamp field, matching most
zip creators.

Fixes #10242.

Change-Id: Id665db274e63def98659231391fb77392267ac1e
Reviewed-on: https://go-review.googlesource.com/18274
Run-TryBot: Russ Cox 
TryBot-Result: Gobot Gobot 
Reviewed-by: Russ Cox

archive/zip: use HTTPS for documentation link

2016-05-14T17:27:44+00:00

The resource is available over (and redirects to) HTTPS, it seems like a good
idea to save a redirect and ensure an encrypted connection.

Change-Id: I262c7616ae289cdd756b6f67573ba6bd7e3e0ca6
Reviewed-on: https://go-review.googlesource.com/23104
Reviewed-by: Brad Fitzpatrick

archive/zip: pool flate readers

2016-05-04T14:28:27+00:00

Similar to the flate Writer pools already used,
this adds pooling for flate Readers.

compress/flate allows re-using of Readers, see
https://codereview.appspot.com/97140043/

In a real-world scenario when reading ~ 500 small files from a ZIP
archive this gives a speedup of 1.5x-2x.

Fixes #14289

Change-Id: I2d98ad983e95ab7d97e06fd0145f619b4f47caa4
Reviewed-on: https://go-review.googlesource.com/19416
Run-TryBot: Brad Fitzpatrick 
TryBot-Result: Gobot Gobot 
Reviewed-by: Brad Fitzpatrick

archive/zip: improve BenchmarkCompressedZipGarbage

2016-05-01T04:42:00+00:00

Before this CL:

$ go test -bench=CompressedZipGarbage -count=5 -run=NONE archive/zip
BenchmarkCompressedZipGarbage-8        50  20677087 ns/op   42973 B/op      47 allocs/op
BenchmarkCompressedZipGarbage-8       100  20584764 ns/op   24294 B/op      47 allocs/op
BenchmarkCompressedZipGarbage-8        50  20859221 ns/op   42973 B/op      47 allocs/op
BenchmarkCompressedZipGarbage-8       100  20901176 ns/op   24294 B/op      47 allocs/op
BenchmarkCompressedZipGarbage-8        50  21282409 ns/op   42973 B/op      47 allocs/op

The B/op number is effectively meaningless. There
is a surprisingly large one-time cost that gets
divided by the number of iterations that your
machine can get through in a second.

This CL discards the first run, which helps.
It is not a panacea. Running with -benchtime=10s
will allow the sync.Pool to be emptied,
which brings the problem back.
However, since there are more iterations to divide
the cost through, it’s not quite as bad,
and running with a high benchtime is rare.

This CL changes the meaning of the B/op number,
which is unfortunate, since it won’t have the
same order of magnitude as previous Go versions.
But it wasn’t really comparable before anyway,
since it didn’t have any reliable meaning at all.

After this CL:

$ go test -bench=CompressedZipGarbage -count=5 -run=NONE archive/zip
BenchmarkCompressedZipGarbage-8   	     100	  20881890 ns/op	    5616 B/op	      47 allocs/op
BenchmarkCompressedZipGarbage-8   	      50	  20622757 ns/op	    5616 B/op	      47 allocs/op
BenchmarkCompressedZipGarbage-8   	      50	  20628193 ns/op	    5616 B/op	      47 allocs/op
BenchmarkCompressedZipGarbage-8   	     100	  20756612 ns/op	    5616 B/op	      47 allocs/op
BenchmarkCompressedZipGarbage-8   	     100	  20639774 ns/op	    5616 B/op	      47 allocs/op

Change-Id: Iedee04f39328974c7fa272a6113d423e7ffce50f
Reviewed-on: https://go-review.googlesource.com/22585
Reviewed-by: Brad Fitzpatrick

all: use new io.SeekFoo constants instead of os.SEEK_FOO

2016-04-13T05:25:23+00:00

Automated change.

Fixes #15269

Change-Id: I8deb2ac0101d3f7c390467ceb0a1561b72edbb2f
Reviewed-on: https://go-review.googlesource.com/21962
Run-TryBot: Brad Fitzpatrick 
Reviewed-by: Andrew Gerrand 
TryBot-Result: Gobot Gobot

all: delete dead test code

2016-03-21T07:10:08+00:00

This deletes unused code and helpers from tests.

Change-Id: Ie31d46115f558ceb8da6efbf90c3c204e03b0d7e
Reviewed-on: https://go-review.googlesource.com/20927
Reviewed-by: Brad Fitzpatrick 
Run-TryBot: Brad Fitzpatrick 
TryBot-Result: Gobot Gobot

archive/zip: add missing argument to error message

2016-03-15T19:13:21+00:00

Silence vet.

Change-Id: I987438847389500cf3b5bc545ef918c66917b51a
Reviewed-on: https://go-review.googlesource.com/20683
Reviewed-by: Brad Fitzpatrick