13111: Serve "site filesystem" at keep-web root URL.
[arvados.git] / services / keep-web / server_test.go
index 1c36f983b0555334226e2a5816a4dfbd021c36a1..ee585ad5b212af1f12f2bad3f162f8c1c11f3a2f 100644 (file)
@@ -1,30 +1,32 @@
+// Copyright (C) The Arvados Authors. All rights reserved.
+//
+// SPDX-License-Identifier: AGPL-3.0
+
 package main
 
 import (
        "crypto/md5"
        "fmt"
+       "io"
+       "io/ioutil"
+       "net"
+       "os"
        "os/exec"
        "strings"
        "testing"
 
+       "git.curoverse.com/arvados.git/sdk/go/arvados"
        "git.curoverse.com/arvados.git/sdk/go/arvadosclient"
        "git.curoverse.com/arvados.git/sdk/go/arvadostest"
        "git.curoverse.com/arvados.git/sdk/go/keepclient"
        check "gopkg.in/check.v1"
 )
 
-var _ = check.Suite(&IntegrationSuite{})
+var testAPIHost = os.Getenv("ARVADOS_API_HOST")
 
-const (
-       spectatorToken  = "zw2f4gwx8hw8cjre7yp6v1zylhrhn3m5gvjq73rtpwhmknrybu"
-       activeToken     = "3kg6k6lzmp9kj5cpkcoxie963cmvjahbt2fod9zru30k1jqdmi"
-       anonymousToken  = "4kg6k6lzmp9kj4cpkcoxie964cmvjahbt4fod9zru44k4jqdmi"
-       fooCollection   = "zzzzz-4zz18-fy296fx3hot09f7"
-       bogusCollection = "zzzzz-4zz18-totallynotexist"
-       hwCollection    = "zzzzz-4zz18-4en62shvi99lxd4"
-)
+var _ = check.Suite(&IntegrationSuite{})
 
-// IntegrationSuite tests need an API server and an arv-git-httpd server
+// IntegrationSuite tests need an API server and a keep-web server
 type IntegrationSuite struct {
        testServer *server
 }
@@ -34,17 +36,17 @@ func (s *IntegrationSuite) TestNoToken(c *check.C) {
                "",
                "bogustoken",
        } {
-               hdr, body := s.runCurl(c, token, "/collections/"+fooCollection+"/foo")
-               c.Check(hdr, check.Matches, `(?s)HTTP/1.1 401 Unauthorized\r\n.*`)
+               hdr, body, _ := s.runCurl(c, token, "collections.example.com", "/collections/"+arvadostest.FooCollection+"/foo")
+               c.Check(hdr, check.Matches, `(?s)HTTP/1.1 404 Not Found\r\n.*`)
                c.Check(body, check.Equals, "")
 
                if token != "" {
-                       hdr, body = s.runCurl(c, token, "/collections/download/"+fooCollection+"/"+token+"/foo")
+                       hdr, body, _ = s.runCurl(c, token, "collections.example.com", "/collections/download/"+arvadostest.FooCollection+"/"+token+"/foo")
                        c.Check(hdr, check.Matches, `(?s)HTTP/1.1 404 Not Found\r\n.*`)
                        c.Check(body, check.Equals, "")
                }
 
-               hdr, body = s.runCurl(c, token, "/bad-route")
+               hdr, body, _ = s.runCurl(c, token, "collections.example.com", "/bad-route")
                c.Check(hdr, check.Matches, `(?s)HTTP/1.1 404 Not Found\r\n.*`)
                c.Check(body, check.Equals, "")
        }
@@ -56,73 +58,227 @@ func (s *IntegrationSuite) TestNoToken(c *check.C) {
 // really works against the server.
 func (s *IntegrationSuite) Test404(c *check.C) {
        for _, uri := range []string{
-               // Routing errors
-               "/",
+               // Routing errors (always 404 regardless of what's stored in Keep)
                "/foo",
                "/download",
                "/collections",
                "/collections/",
-               "/collections/" + fooCollection,
-               "/collections/" + fooCollection + "/",
+               // Implicit/generated index is not implemented yet;
+               // until then, return 404.
+               "/collections/" + arvadostest.FooCollection,
+               "/collections/" + arvadostest.FooCollection + "/",
+               "/collections/" + arvadostest.FooBarDirCollection + "/dir1",
+               "/collections/" + arvadostest.FooBarDirCollection + "/dir1/",
                // Non-existent file in collection
-               "/collections/" + fooCollection + "/theperthcountyconspiracy",
-               "/collections/download/" + fooCollection + "/" + activeToken + "/theperthcountyconspiracy",
+               "/collections/" + arvadostest.FooCollection + "/theperthcountyconspiracy",
+               "/collections/download/" + arvadostest.FooCollection + "/" + arvadostest.ActiveToken + "/theperthcountyconspiracy",
                // Non-existent collection
-               "/collections/" + bogusCollection,
-               "/collections/" + bogusCollection + "/",
-               "/collections/" + bogusCollection + "/theperthcountyconspiracy",
-               "/collections/download/" + bogusCollection + "/" + activeToken + "/theperthcountyconspiracy",
+               "/collections/" + arvadostest.NonexistentCollection,
+               "/collections/" + arvadostest.NonexistentCollection + "/",
+               "/collections/" + arvadostest.NonexistentCollection + "/theperthcountyconspiracy",
+               "/collections/download/" + arvadostest.NonexistentCollection + "/" + arvadostest.ActiveToken + "/theperthcountyconspiracy",
        } {
-               hdr, body := s.runCurl(c, activeToken, uri)
+               hdr, body, _ := s.runCurl(c, arvadostest.ActiveToken, "collections.example.com", uri)
                c.Check(hdr, check.Matches, "(?s)HTTP/1.1 404 Not Found\r\n.*")
-               c.Check(body, check.Equals, "")
+               if len(body) > 0 {
+                       c.Check(body, check.Equals, "404 page not found\n")
+               }
        }
 }
 
-func (s *IntegrationSuite) Test200(c *check.C) {
-       anonymousTokens = []string{anonymousToken}
+func (s *IntegrationSuite) Test1GBFile(c *check.C) {
+       if testing.Short() {
+               c.Skip("skipping 1GB integration test in short mode")
+       }
+       s.test100BlockFile(c, 10000000)
+}
+
+func (s *IntegrationSuite) Test100BlockFile(c *check.C) {
+       if testing.Short() {
+               // 3 MB
+               s.test100BlockFile(c, 30000)
+       } else {
+               // 300 MB
+               s.test100BlockFile(c, 3000000)
+       }
+}
+
+func (s *IntegrationSuite) test100BlockFile(c *check.C, blocksize int) {
+       testdata := make([]byte, blocksize)
+       for i := 0; i < blocksize; i++ {
+               testdata[i] = byte(' ')
+       }
        arv, err := arvadosclient.MakeArvadosClient()
        c.Assert(err, check.Equals, nil)
-       arv.ApiToken = activeToken
-       kc, err := keepclient.MakeKeepClient(&arv)
+       arv.ApiToken = arvadostest.ActiveToken
+       kc, err := keepclient.MakeKeepClient(arv)
        c.Assert(err, check.Equals, nil)
-       kc.PutB([]byte("Hello world\n"))
-       kc.PutB([]byte("foo"))
-       for _, spec := range [][]string{
+       loc, _, err := kc.PutB(testdata[:])
+       c.Assert(err, check.Equals, nil)
+       mtext := "."
+       for i := 0; i < 100; i++ {
+               mtext = mtext + " " + loc
+       }
+       mtext = mtext + fmt.Sprintf(" 0:%d00:testdata.bin\n", blocksize)
+       coll := map[string]interface{}{}
+       err = arv.Create("collections",
+               map[string]interface{}{
+                       "collection": map[string]interface{}{
+                               "name":          fmt.Sprintf("testdata blocksize=%d", blocksize),
+                               "manifest_text": mtext,
+                       },
+               }, &coll)
+       c.Assert(err, check.Equals, nil)
+       uuid := coll["uuid"].(string)
+
+       hdr, body, size := s.runCurl(c, arv.ApiToken, uuid+".collections.example.com", "/testdata.bin")
+       c.Check(hdr, check.Matches, `(?s)HTTP/1.1 200 OK\r\n.*`)
+       c.Check(hdr, check.Matches, `(?si).*Content-length: `+fmt.Sprintf("%d00", blocksize)+`\r\n.*`)
+       c.Check([]byte(body)[:1234], check.DeepEquals, testdata[:1234])
+       c.Check(size, check.Equals, int64(blocksize)*100)
+}
+
+type curlCase struct {
+       auth    string
+       host    string
+       path    string
+       dataMD5 string
+}
+
+func (s *IntegrationSuite) Test200(c *check.C) {
+       s.testServer.Config.AnonymousTokens = []string{arvadostest.AnonymousToken}
+       for _, spec := range []curlCase{
                // My collection
-               {activeToken, "/collections/" + fooCollection + "/foo", "acbd18db4cc2f85cedef654fccc4a4d8"},
-               {"", "/collections/download/" + fooCollection + "/" + activeToken + "/foo", "acbd18db4cc2f85cedef654fccc4a4d8"},
-               {"tokensobogus", "/collections/download/" + fooCollection + "/" + activeToken + "/foo", "acbd18db4cc2f85cedef654fccc4a4d8"},
-               {activeToken, "/collections/download/" + fooCollection + "/" + activeToken + "/foo", "acbd18db4cc2f85cedef654fccc4a4d8"},
-               {anonymousToken, "/collections/download/" + fooCollection + "/" + activeToken + "/foo", "acbd18db4cc2f85cedef654fccc4a4d8"},
-               // Anonymously accessible user agreement. These should
-               // start working when CollectionFileReader provides
-               // real data instead of fake/stub data.
-               {"", "/collections/"+hwCollection+"/Hello%20world.txt", "f0ef7081e1539ac00ef5b761b4fb01b3"},
-               {activeToken, "/collections/"+hwCollection+"/Hello%20world.txt", "f0ef7081e1539ac00ef5b761b4fb01b3"},
-               {spectatorToken, "/collections/"+hwCollection+"/Hello%20world.txt", "f0ef7081e1539ac00ef5b761b4fb01b3"},
-               {spectatorToken, "/collections/download/"+hwCollection+"/"+spectatorToken+"/Hello%20world.txt", "f0ef7081e1539ac00ef5b761b4fb01b3"},
+               {
+                       auth:    arvadostest.ActiveToken,
+                       host:    arvadostest.FooCollection + "--collections.example.com",
+                       path:    "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       auth:    arvadostest.ActiveToken,
+                       host:    arvadostest.FooCollection + ".collections.example.com",
+                       path:    "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       host:    strings.Replace(arvadostest.FooPdh, "+", "-", 1) + ".collections.example.com",
+                       path:    "/t=" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       path:    "/c=" + arvadostest.FooPdh + "/t=" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       path:    "/c=" + strings.Replace(arvadostest.FooPdh, "+", "-", 1) + "/t=" + arvadostest.ActiveToken + "/_/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       path:    "/collections/download/" + arvadostest.FooCollection + "/" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       auth:    "tokensobogus",
+                       path:    "/collections/download/" + arvadostest.FooCollection + "/" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       auth:    arvadostest.ActiveToken,
+                       path:    "/collections/download/" + arvadostest.FooCollection + "/" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+               {
+                       auth:    arvadostest.AnonymousToken,
+                       path:    "/collections/download/" + arvadostest.FooCollection + "/" + arvadostest.ActiveToken + "/foo",
+                       dataMD5: "acbd18db4cc2f85cedef654fccc4a4d8",
+               },
+
+               // Anonymously accessible data
+               {
+                       path:    "/c=" + arvadostest.HelloWorldCollection + "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       host:    arvadostest.HelloWorldCollection + ".collections.example.com",
+                       path:    "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       host:    arvadostest.HelloWorldCollection + ".collections.example.com",
+                       path:    "/_/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       path:    "/collections/" + arvadostest.HelloWorldCollection + "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       auth:    arvadostest.ActiveToken,
+                       path:    "/collections/" + arvadostest.HelloWorldCollection + "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       auth:    arvadostest.SpectatorToken,
+                       path:    "/collections/" + arvadostest.HelloWorldCollection + "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       auth:    arvadostest.SpectatorToken,
+                       host:    arvadostest.HelloWorldCollection + "--collections.example.com",
+                       path:    "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
+               {
+                       auth:    arvadostest.SpectatorToken,
+                       path:    "/collections/download/" + arvadostest.HelloWorldCollection + "/" + arvadostest.SpectatorToken + "/Hello%20world.txt",
+                       dataMD5: "f0ef7081e1539ac00ef5b761b4fb01b3",
+               },
        } {
-               hdr, body := s.runCurl(c, spec[0], spec[1])
-               if strings.HasPrefix(hdr, "HTTP/1.1 501 Not Implemented\r\n") && body == "" {
-                       c.Log("Not implemented!")
-                       continue
+               host := spec.host
+               if host == "" {
+                       host = "collections.example.com"
                }
+               hdr, body, _ := s.runCurl(c, spec.auth, host, spec.path)
                c.Check(hdr, check.Matches, `(?s)HTTP/1.1 200 OK\r\n.*`)
-               c.Check(fmt.Sprintf("%x", md5.Sum([]byte(body))), check.Equals, spec[2])
+               if strings.HasSuffix(spec.path, ".txt") {
+                       c.Check(hdr, check.Matches, `(?s).*\r\nContent-Type: text/plain.*`)
+                       // TODO: Check some types that aren't
+                       // automatically detected by Go's http server
+                       // by sniffing the content.
+               }
+               c.Check(fmt.Sprintf("%x", md5.Sum([]byte(body))), check.Equals, spec.dataMD5)
        }
 }
 
 // Return header block and body.
-func (s *IntegrationSuite) runCurl(c *check.C, token, uri string, args ...string) (hdr, body string) {
+func (s *IntegrationSuite) runCurl(c *check.C, token, host, uri string, args ...string) (hdr, bodyPart string, bodySize int64) {
        curlArgs := []string{"--silent", "--show-error", "--include"}
+       testHost, testPort, _ := net.SplitHostPort(s.testServer.Addr)
+       curlArgs = append(curlArgs, "--resolve", host+":"+testPort+":"+testHost)
        if token != "" {
                curlArgs = append(curlArgs, "-H", "Authorization: OAuth2 "+token)
        }
        curlArgs = append(curlArgs, args...)
-       curlArgs = append(curlArgs, "http://"+s.testServer.Addr+uri)
+       curlArgs = append(curlArgs, "http://"+host+":"+testPort+uri)
        c.Log(fmt.Sprintf("curlArgs == %#v", curlArgs))
-       output, err := exec.Command("curl", curlArgs...).CombinedOutput()
+       cmd := exec.Command("curl", curlArgs...)
+       stdout, err := cmd.StdoutPipe()
+       c.Assert(err, check.Equals, nil)
+       cmd.Stderr = cmd.Stdout
+       go cmd.Start()
+       buf := make([]byte, 2<<27)
+       n, err := io.ReadFull(stdout, buf)
+       // Discard (but measure size of) anything past 128 MiB.
+       var discarded int64
+       if err == io.ErrUnexpectedEOF {
+               buf = buf[:n]
+       } else {
+               c.Assert(err, check.Equals, nil)
+               discarded, err = io.Copy(ioutil.Discard, stdout)
+               c.Assert(err, check.Equals, nil)
+       }
+       err = cmd.Wait()
        // Without "-f", curl exits 0 as long as it gets a valid HTTP
        // response from the server, even if the response status
        // indicates that the request failed. In our test suite, we
@@ -130,29 +286,43 @@ func (s *IntegrationSuite) runCurl(c *check.C, token, uri string, args ...string
        // headers ourselves. If curl exits non-zero, our testing
        // environment is broken.
        c.Assert(err, check.Equals, nil)
-       hdrsAndBody := strings.SplitN(string(output), "\r\n\r\n", 2)
+       hdrsAndBody := strings.SplitN(string(buf), "\r\n\r\n", 2)
        c.Assert(len(hdrsAndBody), check.Equals, 2)
        hdr = hdrsAndBody[0]
-       body = hdrsAndBody[1]
+       bodyPart = hdrsAndBody[1]
+       bodySize = int64(len(bodyPart)) + discarded
        return
 }
 
 func (s *IntegrationSuite) SetUpSuite(c *check.C) {
        arvadostest.StartAPI()
-       arvadostest.StartKeep()
+       arvadostest.StartKeep(2, true)
+
+       arv, err := arvadosclient.MakeArvadosClient()
+       c.Assert(err, check.Equals, nil)
+       arv.ApiToken = arvadostest.ActiveToken
+       kc, err := keepclient.MakeKeepClient(arv)
+       c.Assert(err, check.Equals, nil)
+       kc.PutB([]byte("Hello world\n"))
+       kc.PutB([]byte("foo"))
+       kc.PutB([]byte("foobar"))
 }
 
 func (s *IntegrationSuite) TearDownSuite(c *check.C) {
-       arvadostest.StopKeep()
+       arvadostest.StopKeep(2)
        arvadostest.StopAPI()
 }
 
 func (s *IntegrationSuite) SetUpTest(c *check.C) {
        arvadostest.ResetEnv()
-       s.testServer = &server{}
-       var err error
-       address = "127.0.0.1:0"
-       err = s.testServer.Start()
+       cfg := DefaultConfig()
+       cfg.Client = arvados.Client{
+               APIHost:  testAPIHost,
+               Insecure: true,
+       }
+       cfg.Listen = "127.0.0.1:0"
+       s.testServer = &server{Config: cfg}
+       err := s.testServer.Start()
        c.Assert(err, check.Equals, nil)
 }