Closed farzadbekran closed 2 years ago
Oh, we had a similar bug some years ago, but I thought it was fixed. Do you receive a specific exception? This is probably due to the recent upgrade to the streamly library- it should be easy to fix.
This is all I get: project-m36-server: Network.Socket.recvBuf: end of file (end of file)
I may need some additional details in order to reproduce this locally. Which operating system are you running?
The following steps work for me on Linux locally:
cabal run project-m36-server -- -n spam -d /tmp/db
#in terminal B
cabal run tutd -- -h 127.0.0.1 -n spam
TutorialD (master/main): x:=true
TutorialD (master/main): :commit
#in terminal C
killall -9 tutd
#in terminal B, witness that tutd was killed
cabal run tutd -- -h 127.0.0.1 -n spam
TutorialD (master/main): :showexpr x
┌┐
││
├┤
└┘
There must be some additional detail in your configuration. Are you using notifications perhaps?
I'm using ArchLinux. This is all the steps I take (the directory does not exist, so a fresh db):
first terminal:
stack exec -- project-m36-server -n shopper-db -d shopper-db --fsync
second terminal:
stack exec -- tutd -h 127.0.0.1 -n shopper-db
TutorialD (master/main): :quit
Goodbye.
in the first terminal:
project-m36-server: Network.Socket.recvBuf: end of file (end of file)
my stack.yaml:
resolver:
url: https://raw.githubusercontent.com/commercialhaskell/stackage-snapshots/master/lts/16/31.yaml
packages:
- .
extra-deps:
extra-deps:
- git: https://github.com/agentm/project-m36.git
commit: 81ff8a6d38973c26ce4a42151dc387fbd89ff7b5
- git: https://github.com/agentm/curryer.git
commit: 74ecaac619ad1165a4517d5c0f4e530493bda130
- git: https://github.com/composewell/streamly.git
commit: ac3af8749194f1788704dda8667d0b3807075cc2
- fast-builder-0.1.2.1
- winery-1.3.2
- barbies-2.0.2.0
- barbies-th-0.1.7
- extended-reals-0.2.3.0
- data-interval-2.0.1
- stm-containers-1.1.0.4
- stm-hamt-1.2.0.4
- rset-1.0.0
- hashable-time-0.2.0.2
- network-transport-tcp-0.7.0
- megaparsec-7.0.5
- network-2.8.0.1
- primitive-extras-0.8
- primitive-unlifted-0.1.3.0
- prettyprinter-1.7.0
- prettyprinter-ansi-terminal-1.1.2
- base16-bytestring-1.0.1.0
- modern-uri-0.3.4.0
An old version of the network module in the stack configurations turned out to be the cause of the crash. Fixed now in all stack configurations.
What the title says. This combined with dev mode on yesod and #312 really breaks the workflow (having to delete the db files, starting the server again and initializing db on every code change)