Package Details: mattermost 5.10.0-1

Git Clone URL: (read-only)
Package Base: mattermost
Description: Open source Slack-alternative in Golang and React
Upstream URL:
Keywords: chat mattermost messenger networking slack
Licenses: Apache, AGPL
Submitter: mtorromeo
Maintainer: wget (caleb)
Last Packager: caleb
Votes: 27
Popularity: 0.236908
First Submitted: 2015-08-24 15:14
Last Updated: 2019-04-26 08:25

Latest Comments

1 2 3 4 5 6 ... Next › Last »

caleb commented on 2019-04-26 07:38

@marcool04 I'm sorry but that change doesn't make sense to me. The expression already finds the tabs that exist ((\t*)) and uses them as a guide to indent the added lines (\1). Adding another tab there should not be necessary.

Furthermore I'm not seeing the same make error. Of course I'm working on mattermost 5.10.0 — were you working with the last 5.8.0 or something of your own inbetween?

marcool04 commented on 2019-03-27 18:44

Hi there,

There seems to be a tab missing in a sed line in the PKGBUILD, the one dealing with gifsicle. This causes a make error:

Makefile:21: *** missing separator

The required change to fix it is:

diff --git a/PKGBUILD b/PKGBUILD
index 59d59eb..8028e77 100644
@@ -115,7 +115,7 @@ prepare() {
         gifsicleSystem="$(which gifsicle)"
         sed -r -i Makefile \
-            -e "s/(\t*)npm install(.*)/\0\n\1rm \"$gifsicleNpm\"\n\tln -s \"$gifsicleSystem\" \"$gifsicleNpm\"/"
+            -e "s/(\t*)npm install(.*)/\0\n\1\trm \"$gifsicleNpm\"\n\tln -s \"$gifsicleSystem\" \"$gifsicleNpm\"/"

Regards, Mark.

asere commented on 2019-03-24 14:47

Hi, I'm having these error:

Build Linux arm64
env GOOS=linux GOARCH=arm64 go install -i  -ldflags "-linkmode external -extldflags '-Wl,-O1,
--sort-common,--as-needed,-z,relro,-z,now' -s -w -X
odel.BuildNumber=5.8.0-1 -X ' Mar 
24 03:29:33 UTC 2019' -X ' Arc
h Linux (aarch64)' -X '
0-1 Arch Linux (aarch64)' -X
dy=false" ./...
loadinternal: cannot find runtime/cgo
loadinternal: cannot find runtime/cgo
/usr/lib/go/pkg/tool/linux_arm64/link: signal: killed
make: *** [build/ build-linux] Error 1
==> ERROR: A failure occurred in build().

I'm using Archlinux on a Raspberry Pi (arm64).

ColdSphinX commented on 2019-02-22 11:03

@baltitenger: you need to change ClientDirectory to /usr/share/webapps/mattermost/client/plugins

baltitenger commented on 2019-02-07 18:38

It seems like the /usr/share/webapps/mattermost/bin/client/ directory doesn't get created on install, and mattermost shows this warning: "msg":"Failed to start up plugins","error":"mkdir ./client/plugins: no such file or directory".

EDIT: I've tried creating it, reset the database, and now it only shows a 404 page, but no warnings. If I delete it and reset the database again, it works but shows the warnings.

Dr0ID commented on 2019-02-01 10:07

Hi, please provide wget in makedepends packages.

/bin/sh: line 1: wget: command not found
/bin/sh: line 1: wget: command not found
make: *** [build/ package] Error 127

ArchangeGabriel commented on 2018-11-24 11:45

Building without -i indeed works.

ArchangeGabriel commented on 2018-11-21 16:06

We don’t have reproducible builds yet, building tools still need to be updated. However there is a test server that try to assess whether our builds are reproducible using the new devtools code, you can see it in action here:

Regarding the way to build Go packages, I don’t know, things are moving quite fast there (I think the today standard is go modules, replacing go dep and whatever there was before).

wget commented on 2018-11-12 18:57

@ArchangeGabriel Also, wrt. the aforementioned Go issues you linked to me. It is written in the issues comments we shouldn't be using go install -i any more. Should I report/fix this upstream i.e. directly at Mattermost?

wget commented on 2018-11-12 18:56

@ArchangeGabriel Thanks. Actually I know what these go related compilation trimming functions did, but what I didn't was their purpose: going a step forward to reproducible builds. And for the record, I usually got warnings when building Mattermost. The warnings said local paths were leaking in the final executable. It seems this is even fixing these. Do we have a reproducible build now, or yet further work is required?