felixge/godrone 193
GoDrone is a free software alternative firmware for the Parrot AR Drone 2.0.
Go port of zopfli, a zlib-compatible compression library
A command-line reader for 1Password.opvault files
Summer student project to add group management to Personhood
An open source mobile agent for achieving SSI
Utility to find AES keys in running processes
Explorations in AES Finding and Hiding
jeffallen/aries-framework-dotnet 0
Aries Framework .NET for building multiplatfrom SSI services
A fork of Golang.org/x/crypto/acme/autocert to take CSRs
An embedded key/value database for Go.
created repositoryjeffallen/aes-hider
Explorations in AES Finding and Hiding
created time in 10 days
push eventjeffallen/aes-finder
commit sha 0e38945b26eb4d191b633de44234553dd6697eeb
Add gitignore.
push time in 10 days
issue openeddedis/cothority
Eventlog contract calls time.Now()
Contracts must be time invariant, but eventlog is not.
See also #1331.
created time in 11 days
issue commentdedis/cothority
Subscription for eventlog (SecureKG)
Take a look, maybe, @pierluca .
comment created time in 11 days
issue openeddedis/student20_pop
Choose a license, get Prof Ford's approval, and commit it to the repo.
created time in 15 days
push eventdedis/kyber
commit sha 9ac80102d756a21f318685e230e33791c44b5e2e
Add deprecation warning While investigating #436, I found that the CoSi deprecation warning was only in Cothority, not in Kyber.
commit sha 2fc887469d46c88ba2fa83d7b3e4c951715a32d9
Merge pull request #438 from dedis/issue436 Add deprecation warning
push time in 21 days
PR merged dedis/kyber
While investigating #436, I found that the CoSi deprecation warning was only in Cothority, not in Kyber.
pr closed time in 21 days
push eventdedis/onet
commit sha c8453d478e61b621ae780bf5035694da5fc4c913
Replace satori/go.uuid with google/uuid (#661) * Replace satori/go.uuid with google/uuid Closes #660
commit sha 3a28acc588c3856475f3e0d0b9d41dfc93d6adf9
Merge branch 'master' into issue601
push time in 25 days
issue closeddedis/kyber
Investigate impact of ROS solver on CoSi and DKG
This paper claims attacks on CoSi and DKG: https://eprint.iacr.org/2020/945.pdf
This issue is to track our findings as we analyse what if anything needs to be done in Kyber in response.
cc @LefKok
closed time in 25 days
jeffallen
push eventdedis/kyber
commit sha 814a25b9fabf0fe70b94e2a917e947ab13b260c2
Merge pull request #437 from dedis/go15 Upgrade to Go 1.15
commit sha 9ac80102d756a21f318685e230e33791c44b5e2e
Add deprecation warning While investigating #436, I found that the CoSi deprecation warning was only in Cothority, not in Kyber.
push time in 25 days
PR opened dedis/kyber
While investigating #436, I found that the CoSi deprecation warning was only in Cothority, not in Kyber.
pr created time in 25 days
push eventdedis/onet
commit sha f61455dd79da98100bacd767c203f706ed148b6f
Cancel the context, as suggested by the docs and required by go vet.
push time in 25 days
Pull request review commentdedis/onet
Remove dependency on tylerb/graceful
func (w *WebSocket) stop() { if !w.started { return }- log.Lvl3("Stopping", w.server.Server.Addr)- w.server.Stop(100 * time.Millisecond)+ log.Lvl3("Stopping", w.server.Addr)++ d := time.Now().Add(100 * time.Millisecond)+ ctx, _ := context.WithDeadline(context.Background(), d)+ w.server.Shutdown(ctx)+ <-w.startstop
I figured out my confusion, I was looking at the signalling via w.startstop, and got confused thinking that WebSocket.Stop() was trying to be asynchronous, when in fact it blocking and sa @Gilthoniel says the call to net/http.Shutdown is blocking (until the context expires) as well.
comment created time in 25 days
Pull request review commentdedis/onet
Remove dependency on tylerb/graceful
func (w *WebSocket) stop() { if !w.started { return }- log.Lvl3("Stopping", w.server.Server.Addr)- w.server.Stop(100 * time.Millisecond)+ log.Lvl3("Stopping", w.server.Addr)++ d := time.Now().Add(100 * time.Millisecond)+ ctx, _ := context.WithDeadline(context.Background(), d)+ w.server.Shutdown(ctx)+ <-w.startstop
Defer will trigger when the function returns, but the shutdown is not expected to be done by then, it would shortcut the timeout. Still reading and thinking to see if I can convince myself either that the timeout is not useful or that the cancel can be called so fast.
comment created time in a month
pull request commentdedis/onet
Remove dependency on tylerb/graceful
So go vet is mad because ./websocket.go:224:7: the cancel function returned by context.WithDeadline should be called, not discarded, to avoid a context leak. It seemed like due to async concerns I wouldn't be able to call that. Back to investigate more I guess.
comment created time in a month
PR opened dedis/onet
Fixes #601.
pr created time in a month
push eventdedis/onet
commit sha 635958689ca9a926f358721c597432b613cd665f
Warn when a failed connection might be from a known bug Fixes #657
commit sha 655fed7d2ae90b01887a8c963374b8ffd3aeeb5d
Update network/tcp.go Co-authored-by: Linus Gasser <[email protected]>
commit sha 3ac7df92a83745bd23dafe55e5dd045626ae14f2
Merge pull request #662 from dedis/issue657 Warn when a failed connection might be from a known bug
push time in a month
PR merged dedis/onet
Fixes #657
pr closed time in a month
issue closeddedis/onet
Make better error message for TLS 1.5 fix
The fix for go 1.15 with regard to TLS, #655, made the test-network crash. Some more information can be found in #656, which is, however, not capable of reproducing the error in a way that we can try to fix it.
So as the minimal service to users of our code, I propose we do the following:
Improve the error message given by an incompatible onet-connection and be more clear what might have happened. E.g., something like:
TLS handshake failed - this might be due to mismatching go-version numbers, e.g., 1.14 and 1.15, between nodes. Another reason might be nodes running with onet (3.2.4) that cannot communicate with onet (3.2.5)
We should also consider removing the 3.2.5 tag and replacing it with a 3.3.0 tag, and write something in the CHANGELOG.md
closed time in a month
ineiti
pull request commentdedis/onet
Warn when a failed connection might be from a known bug
The build failed due to a data race that seems to be associated with graceful:
WARNING: DATA RACE
314Read at 0x00c0001189f8 by goroutine 106:
315 net/http.(*conn).setState()
316 /home/travis/.gimme/versions/go1.15.linux.amd64/src/net/http/server.go:1753 +0x110
317 net/http.(*conn).serve.func1()
318 /home/travis/.gimme/versions/go1.15.linux.amd64/src/net/http/server.go:1806 +0xfd
319 net/http.(*conn).serve()
320 /home/travis/.gimme/versions/go1.15.linux.amd64/src/net/http/server.go:1945 +0xaff
321
322Previous write at 0x00c0001189f8 by goroutine 97:
323 gopkg.in/tylerb/graceful%2ev1.(*Server).manageConnections()
324 /home/travis/gopath/pkg/mod/gopkg.in/tylerb/[email protected]/graceful.go:399 +0x644
325
326Goroutine 106 (running) created at:
327 net/http.(*Server).Serve()
328 /home/travis/.gimme/versions/go1.15.linux.amd64/src/net/http/server.go:2969 +0x5d3
329 gopkg.in/tylerb/graceful%2ev1.(*Server).Serve()
330 /home/travis/gopath/pkg/mod/gopkg.in/tylerb/[email protected]/graceful.go:309 +0x472
331 gopkg.in/tylerb/graceful%2ev1.(*Server).ListenAndServe()
332 /home/travis/gopath/pkg/mod/gopkg.in/tylerb/[email protected]/graceful.go:147 +0xf2
333 go.dedis.ch/onet/v3.(*WebSocket).start.func1()
334 /home/travis/gopath/src/go.dedis.ch/onet/v3/websocket.go:194 +0x15c
335
336Goroutine 97 (finished) created at:
337 gopkg.in/tylerb/graceful%2ev1.(*Server).Serve()
338 /home/travis/gopath/pkg/mod/gopkg.in/tylerb/[email protected]/graceful.go:297 +0x377
339 gopkg.in/tylerb/graceful%2ev1.(*Server).ListenAndServe()
340 /home/travis/gopath/pkg/mod/gopkg.in/tylerb/[email protected]/graceful.go:147 +0xf2
341 go.dedis.ch/onet/v3.(*WebSocket).start.func1()
342 /home/travis/gopath/src/go.dedis.ch/onet/v3/websocket.go:194 +0x15c
343==================
344 testing.go:1023: race detected during execution of test
345--- FAIL: TestLocalTCPGenConnectableRoster (0.48s)
346=== CONT
347 testing.go:1023: race detected during execution of test
348
We already have #601 reminding us to get rid of graceful, so I'm going to merge this anyway.
comment created time in a month
push eventdedis/onet
commit sha 655fed7d2ae90b01887a8c963374b8ffd3aeeb5d
Update network/tcp.go Co-authored-by: Linus Gasser <[email protected]>
push time in a month
PR opened dedis/onet
Fixes #657
pr created time in a month
issue commentdedis/onet
Make better error message for TLS 1.5 fix
I tested this with a 3 node Cothority where two were old Onet and one was a new Onet. I used this patch to Cothortiy's run_nodes.sh to do so:
diff --git a/conode/run_nodes.sh b/conode/run_nodes.sh
index fdaca7951..2a84da372 100755
--- a/conode/run_nodes.sh
+++ b/conode/run_nodes.sh
@@ -92,10 +92,14 @@ for n in $( seq $nbr_nodes ); do
export CONODE_SERVICE_PATH=$(pwd)
while [[ -f running ]]; do
echo "Starting conode $LOG"
+ cb=$CONODE_BIN
+ if [ -x $CONODE_BIN-$co ]; then
+ cb=$CONODE_BIN-$co
+ fi
if [[ "$SHOW" ]]; then
- $CONODE_BIN -d $verbose -c $co/private.toml server 2>&1 | tee $LOG-$(date +%y%m%d-%H%M).log
+ $cb -d $verbose -c $co/private.toml server 2>&1 | tee $LOG-$(date +%y%m%d-%H%M).log
else
- $CONODE_BIN -d $verbose -c $co/private.toml server > $LOG-$(date +%y%m%d-%H%M).log 2>&1
+ $cb -d $verbose -c $co/private.toml server > $LOG-$(date +%y%m%d-%H%M).log 2>&1
fi
if [[ "$single" ]]; then
echo "Will not restart conode in single mode."
comment created time in a month
issue openedjschaf/bibtex
Backslash does not work in in double quote strings
If a URL has a literal tilde in it, it must be possible to escape that, instead of turning it into an NBSP.
Here's a unit test to show what I mean:
diff --git a/parser/parser_test.go b/parser/parser_test.go
index 0029289..474970c 100644
--- a/parser/parser_test.go
+++ b/parser/parser_test.go
@@ -209,6 +209,13 @@ func TestParseFile_BibDecl_ModeParseStrings(t *testing.T) {
asts.WithBibTags("howPublished",
asts.QuotedTextExpr(0, asts.Text(`\url`),
asts.BraceTextExpr(1, asts.Text("http://example.com/foo-bar/"), asts.NBSP(), asts.Text("baz/"))))},
+ {`@article{cite_key,
+ howPublished = "\url{http://example.com/foo-bar/\~baz/}"
+ }`,
+ asts.WithBibKeys("cite_key"),
+ asts.WithBibTags("howPublished",
+ asts.QuotedTextExpr(0, asts.Text(`\url`),
+ asts.BraceTextExpr(1, asts.Text("http://example.com/foo-bar/~baz/"))))},
{`@article { cite_key, title = {\href{https://nyt.com/}{Dollar \$140}} }`,
asts.WithBibKeys("cite_key"),
asts.WithBibTags("title",
created time in a month
pull request commentjschaf/bibtex
Sorry, missed your comments while rebasing, removing backslash, etc. Fixed now.
comment created time in a month
push eventjeffallen/bibtex
commit sha 06dd2ef28291976e3c8240f14e1a2bdedc9755e5
Correctly populate the author suffix. Fix a crash with empty author name. Fixes #1.
push time in a month
push eventjeffallen/bibtex
commit sha a17d7107bf00be5774b47a8a7820a48eabbc3a6d
Allow escaped chars in strings - Allow strings like `"\$"` - Allow trailing commas after bibtex declarations
commit sha c0bf8bf75bce222408aca120fff9d7a5794cab6b
Fix endless loop on illegal tokens Fail in main loop if next token is illegal
commit sha 180d9226ab0c3b0d943574448b7bfd2e22d593f2
Correctly populate the author suffix. Fix a crash with empty author name. Fixes #1.
push time in a month
push eventjeffallen/bibtex
commit sha 8f3f81abb95bcefa9e542bd6f20b2daf8e2edea1
Better error message, remove debugging.
push time in a month
push eventjeffallen/bibtex
commit sha f1064b1c896cad76d5091ae7e4b8ec8d70363497
Fix a crash with empty author name.
push time in a month
issue commentjschaf/bibtex
crashing problems from parsing and resolving
The escaping works now, but the parser can still be hung with unterminated things.
comment created time in a month
push eventjeffallen/bibtex
commit sha b4be74119fe199b51d0d84691cd6cbd5f8da9499
Adds backslash escapaing.
commit sha 9901b123c62d208f40ee21ad538944e1cf210e6b
The problem is not only unterminated math, but also unterminated quotes.
push time in a month
issue commentjschaf/bibtex
crashing problems from parsing and resolving
The $ problem is actually 2 problems. An unterminated math expression will hang the parser. And backslash to escape characters is not implemented yet in the scanner. The characters that need to be escaped are listed in the end of this page: http://www.bibtex.org/SpecialSymbols/
comment created time in a month
push eventjeffallen/bibtex
commit sha 6bc3a6daffdc089209197be3e7984bd95289e336
Correctly populate the suffix.
push time in a month
Pull request review commentjschaf/bibtex
func ResolveFile(fset *gotok.FileSet, filename string, src interface{}) ([]Entry // ResolveAuthors extracts the authors from the parsed text of a bibtex field, // usually the author or editor field.-func ResolveAuthors(txt *ast.ParsedText) ([]Author, error) {+func ResolveAuthors(txt *ast.ParsedText) (a []Author, err error) {+ defer func() {+ if r := recover(); r != nil {
The problem is that the panic comes from inside of resolveAuthor, because it has no way to return errors. I suppose what you mean is, "let's change the signature of resolveAuthor to return errors", so I'll do that.
comment created time in a month
push eventdedis/bib
commit sha 16427220365a806374aa724cc660a83c5d8cf80c
Correct the bib files to be more parsable The bib files were full of stray tags and commas. Some lines also had problems making github.com/jschaf/bibtex unable to parse them, so they have been subtly modified to avoid tickling the bugs in that parser until we manage to fix jschaf/bibtex#1.
push time in a month
issue commentjschaf/bibtex
Moving on through my list of bibtex entries, I found one which causes the parser to go into an infinite loop and I write a unite test for it. It is unclear how dollar should be handled, when it is not starting a TeX formula. But it should not be possible to put the parser into an infinite look anyway. :(
comment created time in a month
push eventjeffallen/bibtex
commit sha bb262f68980efd2019ec361d20eb769ee42ae0a3
A unit test to show an infinite loop in the parser.
push time in a month
issue commentjschaf/bibtex
I found a related problem with commas in author names, and included a second commit with a test and some ideas on how to fix these. Hope this is helpful.
comment created time in a month
push eventjeffallen/bibtex
commit sha 91f00c0b4b7a18041a744c83c46018f6f5c97cf6
Another example of a related problem.
push time in a month
issue openedjschaf/bibtex
Hello, thanks for this, it's right what I need, right when I need it.
I tried to parse a huge list of entries, and I got a panic from ResolveAuthors in the case where there are {} inside of an author value. I will post a unit test showing the problem in a PR. I don't understand enough about the AST to fix it myself yet.
created time in a month
push eventdedis/cothority
commit sha 34c34c7f83b1ee0a2f8a52786a9301f7dc56cb2b
Making Jenkins pass Update the jenkins-docker to 1.15 Improve two tests that were often failing on Jenkins due to protocol-timeouts
commit sha 3f644ddb8fbb1ad4df26f99e411176db214018b0
Update Dockerfile-int
commit sha 2dd82e7edc702aec70de27571e1831733de4c525
Update byzcoin/viewchange_test.go
commit sha 2d50ed5d53eef319bbe361cf6b962678679b8e3b
Merge pull request #2366 from dedis/jenkins_fix Making Jenkins pass
push time in a month
PR merged dedis/cothority
Update the jenkins-docker to 1.15 Improve two tests that were often failing on Jenkins due to protocol-timeouts
OR-review - whoever accepts first shall rebase...
pr closed time in a month
push eventdedis/cothority
commit sha 2dd82e7edc702aec70de27571e1831733de4c525
Update byzcoin/viewchange_test.go
push time in a month
Pull request review commentdedis/cothority
func TestViewChange_Basic3(t *testing.T) { testViewChange(t, 10, 3, 2*testInterval) } +// This test is brittle with regard to timeouts:+// - the given interval has been used for the 'blockInterval',+// which doesn't exist anymore, but is still used for+// - the timeout when adding a transaction+// - calculating the timeouts when asking for a signature+// - the time to wait to propagate to children+// So it's using the `SetPropagationTimeout` to tweak it a bit.+// But it's a bi...
comment created time in a month
push eventdedis/cothority
commit sha 3f644ddb8fbb1ad4df26f99e411176db214018b0
Update Dockerfile-int
push time in a month
Pull request review commentdedis/cothority
# # As you can see, you can run the long tests using the tag build argument. -FROM golang:1.13+FROM golang:1.15
Dockerfile-int too, right?
comment created time in a month
Pull request review commentdedis/cothority
func TestViewChange_Basic3(t *testing.T) { testViewChange(t, 10, 3, 2*testInterval) } +// This test is brittle with regard to timeouts:+// - the given interval has been used for the 'blockInterval',+// which doesn't exist anymore, but is still used for+// - the timeout when adding a transaction+// - calculating the timeouts when asking for a signature+// - the time to wait to propagate to children+// So it's using the `SetPropagationTimeout` to tweak it a bit.+// But it's a bi...
A bit what?
comment created time in a month
pull request commentdedis/onet
So I'm a little unclear on where the changes to the error logging came from, don't know if they need to be there, in which case we need to fix the tests.
Also, not clear on what this tells us: it seems like this proves that this change is backwards compatible, but clearly the experience with the production network says it was not.
comment created time in a month
push eventdedis/onet
commit sha fc444347a5305ca932fda433672163002223fe2f
Had to roll-back this change as well.
push time in a month
push eventdedis/onet
commit sha b29523d03524f241b5ca2851cf08cbc5f97bcf76
Fix last failing test?
push time in a month
Pull request review commentdedis/onet
func NewTCPConn(addr Address, suite Suite) (conn *TCPConn, err error) { func (c *TCPConn) Receive() (env *Envelope, e error) { buff, err := c.receiveRaw() if err != nil {- return nil, xerrors.Errorf("receiving: %w", err)
I don't know why, but this change causes TestRouterErrorHandling to fail.
comment created time in a month
push eventdedis/onet
commit sha 9502752c5925ace91567c1e63781d5dca47d9917
Try with Go 1.14.
push time in a month
Pull request review commentdedis/onet
func pubToCN(pub kyber.Point) string { // tlsConfig returns a generic config that has things set as both the server // and client need them. The returned config is customized after tlsConfig returns.-func tlsConfig(suite Suite, us *ServerIdentity) (*tls.Config, error) {+func tlsConfig(suite Suite, us *ServerIdentity) (*tls.Config,+ error) {
pre-commit, I hate you.
comment created time in a month
Pull request review commentdedis/onet
type verifier func(rawCerts [][]byte, vrf [][]*x509.Certificate) (err error) // gives us a certificate back, crypto/tls calls the verifier with arguments we // can't control. But the verifier still has access to the nonce because it's in the // closure.-func makeVerifier(suite Suite, them *ServerIdentity) (verifier, []byte) {+func makeVerifier(suite Suite, them *ServerIdentity) (verifier,
This is bad formatting. If this is the fault of the pre-commit stuff, it needs to be fixed somehow.
comment created time in a month
Pull request review commentdedis/onet
func (cm *certMaker) get(nonce []byte) (*tls.Certificate, error) { // test code to get ahold of and modify. if testNoURIs { tmpl.URIs = nil+ tmpl.MaxPathLen = 1
This is necessary as a true repro of v3.2.4 behaviour, but it is not important. It will cause Go 1.15's server-side TLS to fail with an "internal error".
comment created time in a month
push eventdedis/onet
commit sha bf6194ed84e71660d6321f93102f445c5dad66f7
A simpler version; this passes with Go 1.14, not with Go 1.15.
push time in a month
Pull request review commentdedis/onet
func makeVerifier(suite Suite, them *ServerIdentity) (verifier, []byte) { // nil) check that the public key advertsied by the far side is the same // as the public key we expect. if them != nil {- if len(cert.URIs) > 0 {- // see explanation of the URL scheme above for why we prepend :.- cn := fmt.Sprintf(":%v", pubToCN(them.Public))- found := false- for _, u := range cert.URIs {- if u.Scheme == "onet-pubkey" && u.Opaque == cn {- found = true- break- }- }- if !found {- return xerrors.Errorf("No onet-pubkey URIs match the expected public key %v", pubToCN(them.Public))+ if noURIs {+ err = cert.VerifyHostname(pubToCN(them.Public))
This call will not pass, no matter what, if you are testing with Go 1.15. The idea of this PR is to provide two implementations, one like v3.2.4 and one like v3.2.5 and mix and match them. That's cool, but then you can only get it to pass with Go 1.14.
I installed Go 1.14 and then ran the unit test like this:
brew install [email protected]
/usr/local/Cellar/[email protected]/1.14.7/bin/go test -run TestTLS -v
comment created time in a month
push eventdedis/cothority
commit sha 81539db35dbb84cbbe88abcc70e283b4b053cc79
Go 1.15 update.
commit sha c7ec23bc49349d06fd4888c13d4987d5258b686b
Merge pull request #2359 from dedis/go15 Upgrade to Go 1.15
push time in a month
PR merged dedis/cothority
pr closed time in a month
push eventdedis/cothority
commit sha d14a4542bf6f359875390c489f9b51fde00f79cd
Documentation Added description of pre-commit Removed while copying its content to
commit sha 48d81a8935f1a53995bd5305d1b83242911c7a66
Merge pull request #2362 from dedis/contributing Documentation
push time in a month
PR merged dedis/cothority
Added description of pre-commit Removed while copying its content to
pr closed time in a month
push eventdedis/cothority
commit sha 6e091325f840c3683f90c4c94cf116449112f9e5
Improve result/event unpack (#2348)
commit sha f4705cabd6b2eb23c956dee35b5ab3d5f3466cd8
Updates kyber dependencies Fixes CVE-2020-13822 (elliptic) and CVE-2020-7660 (serialize-javascript)
commit sha 716e371ef7afd88940a9368a70aa1eaa00e769f4
Add default pre-commit-config
commit sha 951d0617c05cd0b664b608346a079956d3bb6432
Run pre-commit on all files
commit sha 710f7c71161d343c006b7f29522660e65d726432
Travis: use pre-commit during lint stage
commit sha 8ccf34bcf9a49e6b40841030a8f0e367105d91d6
Add YAML check in pre-commit
commit sha 778da8219651eb9856da9c85de3be0cbae048f4d
Add JSON and some GoLang checks
commit sha 867b7148e075f448e3c8161db39fae4008b1f6cd
Merge branch 'master' into kyber-js-update
commit sha bad37ffbbd3902715b7b06f923a99815b3729211
Merge pull request #2360 from dedis/kyber-js-update Updates kyber dependencies
commit sha 81539db35dbb84cbbe88abcc70e283b4b053cc79
Go 1.15 update.
push time in a month
PR opened dedis/cothority
pr created time in a month