# The problem
There were many "path tricks":
* By default, Gitea uses its program directory as its work path
* Gitea tries to use the "work path" to guess its "custom path" and
"custom conf (app.ini)"
* Users might want to use other directories as work path
* The non-default work path should be passed to Gitea by GITEA_WORK_DIR
or "--work-path"
* But some Gitea processes are started without these values
* The "serv" process started by OpenSSH server
* The CLI sub-commands started by site admin
* The paths are guessed by SetCustomPathAndConf again and again
* The default values of "work path / custom path / custom conf" can be
changed when compiling
# The solution
* Use `InitWorkPathAndCommonConfig` to handle these path tricks, and use
test code to cover its behaviors.
* When Gitea's web server runs, write the WORK_PATH to "app.ini", this
value must be the most correct one, because if this value is not right,
users would find that the web UI doesn't work and then they should be
able to fix it.
* Then all other sub-commands can use the WORK_PATH in app.ini to
initialize their paths.
* By the way, when Gitea starts for git protocol, it shouldn't output
any log, otherwise the git protocol gets broken and client blocks
forever.
The "work path" priority is: WORK_PATH in app.ini > cmd arg --work-path
> env var GITEA_WORK_DIR > builtin default
The "app.ini" searching order is: cmd arg --config > cmd arg "work path
/ custom path" > env var "work path / custom path" > builtin default
## ⚠️ BREAKING
If your instance's "work path / custom path / custom conf" doesn't meet
the requirements (eg: work path must be absolute), Gitea will report a
fatal error and exit. You need to set these values according to the
error log.
----
Close#24818Close#24222Close#21606Close#21498Close#25107Close#24981
Maybe close#24503
Replace #23301
Replace #22754
And maybe more
That's a longstanding INI package problem: the "MustXxx" calls change
the option values, and the following "Save" will save a lot of garbage
options into the user's config file.
Ideally we should refactor the INI package to a clear solution, but it's
a huge work.
A clear workaround is what this PR does: when "Save", load a clear INI
instance and save it.
Partially fix#25377, the "install" page needs more fine tunes.
1. The "web" package shouldn't depends on "modules/context" package,
instead, let each "web context" register themselves to the "web"
package.
2. The old Init/Free doesn't make sense, so simplify it
* The ctx in "Init(ctx)" is never used, and shouldn't be used that way
* The "Free" is never called and shouldn't be called because the SSPI
instance is shared
---------
Co-authored-by: Giteabot <teabot@gitea.io>
Fix#24856
Rename "context.contextKey" to "context.WebContextKey", this context is
for web context only. But the Context itself is not renamed, otherwise
it would cause a lot of changes (if we really want to rename it, there
could be a separate PR).
The old test code doesn't really test, the "install page" gets broken
not only one time, so use new test code to make sure the "install page"
could work.
## ⚠️ Breaking
The `log.<mode>.<logger>` style config has been dropped. If you used it,
please check the new config manual & app.example.ini to make your
instance output logs as expected.
Although many legacy options still work, it's encouraged to upgrade to
the new options.
The SMTP logger is deleted because SMTP is not suitable to collect logs.
If you have manually configured Gitea log options, please confirm the
logger system works as expected after upgrading.
## Description
Close#12082 and maybe more log-related issues, resolve some related
FIXMEs in old code (which seems unfixable before)
Just like rewriting queue #24505 : make code maintainable, clear legacy
bugs, and add the ability to support more writers (eg: JSON, structured
log)
There is a new document (with examples): `logging-config.en-us.md`
This PR is safer than the queue rewriting, because it's just for
logging, it won't break other logic.
## The old problems
The logging system is quite old and difficult to maintain:
* Unclear concepts: Logger, NamedLogger, MultiChannelledLogger,
SubLogger, EventLogger, WriterLogger etc
* Some code is diffuclt to konw whether it is right:
`log.DelNamedLogger("console")` vs `log.DelNamedLogger(log.DEFAULT)` vs
`log.DelLogger("console")`
* The old system heavily depends on ini config system, it's difficult to
create new logger for different purpose, and it's very fragile.
* The "color" trick is difficult to use and read, many colors are
unnecessary, and in the future structured log could help
* It's difficult to add other log formats, eg: JSON format
* The log outputer doesn't have full control of its goroutine, it's
difficult to make outputer have advanced behaviors
* The logs could be lost in some cases: eg: no Fatal error when using
CLI.
* Config options are passed by JSON, which is quite fragile.
* INI package makes the KEY in `[log]` section visible in `[log.sub1]`
and `[log.sub1.subA]`, this behavior is quite fragile and would cause
more unclear problems, and there is no strong requirement to support
`log.<mode>.<logger>` syntax.
## The new design
See `logger.go` for documents.
## Screenshot
<details>
</details>
## TODO
* [x] add some new tests
* [x] fix some tests
* [x] test some sub-commands (manually ....)
---------
Co-authored-by: Jason Song <i@wolfogre.com>
Co-authored-by: delvh <dev.lh@web.de>
Co-authored-by: Giteabot <teabot@gitea.io>
Creating pid file should not belong to setting package and only web
command needs that. So this PR moves pidfile creation from setting
package to web command package to keep setting package more readable.
I marked this as `break` because the PIDFile path moved. For those who
have used the pid build argument, it has to be changed.
---------
Co-authored-by: KN4CK3R <admin@oldschoolhack.me>
Some bugs caused by less unit tests in fundamental packages. This PR
refactor `setting` package so that create a unit test will be easier
than before.
- All `LoadFromXXX` files has been splited as two functions, one is
`InitProviderFromXXX` and `LoadCommonSettings`. The first functions will
only include the code to create or new a ini file. The second function
will load common settings.
- It also renames all functions in setting from `newXXXService` to
`loadXXXSetting` or `loadXXXFrom` to make the function name less
confusing.
- Move `XORMLog` to `SQLLog` because it's a better name for that.
Maybe we should finally move these `loadXXXSetting` into the `XXXInit`
function? Any idea?
---------
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: delvh <dev.lh@web.de>
Change all license headers to comply with REUSE specification.
Fix#16132
Co-authored-by: flynnnnnnnnnn <flynnnnnnnnnn@github>
Co-authored-by: John Olheiser <john.olheiser@gmail.com>
Only load SECRET_KEY and INTERNAL_TOKEN if they exist.
Never write the config file if the keys do not exist, which was only a fallback for Gitea upgraded from < 1.5
Co-authored-by: wxiaoguang <wxiaoguang@gmail.com>
The recovery, API, Web and package frameworks all create their own HTML
Renderers. This increases the memory requirements of Gitea
unnecessarily with duplicate templates being kept in memory.
Further the reloading framework in dev mode for these involves locking
and recompiling all of the templates on each load. This will potentially
hide concurrency issues and it is inefficient.
This PR stores the templates renderer in the context and stores this
context in the NormalRoutes, it then creates a fsnotify.Watcher
framework to watch files.
The watching framework is then extended to the mailer templates which
were previously not being reloaded in dev.
Then the locales are simplified to a similar structure.
Fix#20210Fix#20211Fix#20217
Signed-off-by: Andrew Thornton <art27@cantab.net>
This PR adds functionality to allow Gitea to sit behind an
HAProxy and HAProxy protocolled connections directly.
Fix#7508
Signed-off-by: Andrew Thornton <art27@cantab.net>
fgprof is a sampling Go profiler that allows you to analyze On-CPU as
well as Off-CPU (e.g. I/O) time together.
Go's builtin sampling CPU profiler can only show On-CPU time, but it's
better than fgprof at that. Go also includes tracing profilers that can
analyze I/O, but they can't be combined with the CPU profiler.
fgprof is designed for analyzing applications with mixed I/O and CPU
workloads. This kind of profiling is also known as wall-clock profiling.
Whilst fgprof can cause significant STW latencies in applications with a
lot of goroutines (> 1-10k), these latencies only occur if the profile
is requested - it doesn't cause a delay by simply being available.
The fgprof profile is mounted on
`http://localhost:6060/debug/fgprof?seconds=3`
Signed-off-by: Andrew Thornton <art27@cantab.net>
Continues on from #19202.
Following the addition of pprof labels we can now more easily understand the relationship between a goroutine and the requests that spawn them.
This PR takes advantage of the labels and adds a few others, then provides a mechanism for the monitoring page to query the pprof goroutine profile.
The binary profile that results from this profile is immediately piped in to the google library for parsing this and then stack traces are formed for the goroutines.
If the goroutine is within a context or has been created from a goroutine within a process context it will acquire the process description labels for that process.
The goroutines are mapped with there associate pids and any that do not have an associated pid are placed in a group at the bottom as unbound.
In this way we should be able to more easily examine goroutines that have been stuck.
A manager command `gitea manager processes` is also provided that can export the processes (with or without stacktraces) to the command line.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Make router logger more friendly, show the related function name/file/line.
[BREAKING]
This PR substantially changes the logging format of the router logger. If you use this logging for monitoring e.g. fail2ban you will need to update this to match the new format.
* Improve install code to avoid low-level mistakes.
If a user tries to do a re-install in a Gitea database, they gets a warning and double check.
When Gitea runs, it never create empty app.ini automatically.
Also some small (related) refactoring:
* Refactor db.InitEngine related logic make it more clean (especially for the install code)
* Move some i18n strings out from setting.go to make the setting.go can be easily maintained.
* Show errors in CLI code if an incorrect app.ini is used.
* APP_DATA_PATH is created when installing, and checked when starting (no empty directory is created any more).
* Simplify Gothic to use our session store instead of creating a different store
We have been using xormstore to provide a separate session store for our OAuth2 logins
however, this relies on using gorilla context and some doubling of our session storing.
We can however, simplify and simply use our own chi-based session store. Thus removing
a cookie and some of the weirdness with missing contexts.
Signed-off-by: Andrew Thornton <art27@cantab.net>
* as per review
Signed-off-by: Andrew Thornton <art27@cantab.net>
* as per review
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Handle MaxTokenLength
Signed-off-by: Andrew Thornton <art27@cantab.net>
* oops
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Co-authored-by: Lauris BH <lauris@nix.lv>
Although panics within the rendering pipeline are caught and dealt with,
panics that occur before that starts are unprotected and will kill Gitea
without being sent to the logs.
This PR adds a basic recovery handler to catch panics that occur after
the logger is initialised and ensure that they're sent to the logger.
Signed-off-by: Andrew Thornton <art27@cantab.net>
One of the repeatedly reported issues has been that gitea produces too much console
logging during set up even if the console logger is turned off.
Fundamentally this is due to some otherwise very helpful logging that has to occur
before logging is set up. This has come to a head with the merging of #16243 where
otherwise potentially helpful Trace logging in the git module now appears on the
console.
This PR proposes three things:
1. Change the initial default logger to Info not Trace.
2. Change the logging for the AppPath things to Info in recompense.
3. Add two new command line options to gitea web: --quiet and --verbose
`gitea web -q` or `gitea web --quiet` will only log Fatal level initially.
`gitea web -verbose` will log at Trace.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
* refactor routers directory
* move func used for web and api to common
* make corsHandler a function to prohibit side efects
* rm unused func
Co-authored-by: 6543 <6543@obermui.de>
* Make SSH server host key path configurable
* make it possible to have multiple keys
* Make gitea.rsa the default key
* Add some more logging
Signed-off-by: Andrew Thornton <art27@cantab.net>
Use [chi](https://github.com/go-chi/chi) instead of the forked [macaron](https://gitea.com/macaron/macaron). Since macaron and chi have conflicts with session share, this big PR becomes a have-to thing. According my previous idea, we can replace macaron step by step but I'm wrong. :( Below is a list of big changes on this PR.
- [x] Define `context.ResponseWriter` interface with an implementation `context.Response`.
- [x] Use chi instead of macaron, and also a customize `Route` to wrap chi so that the router usage is similar as before.
- [x] Create different routers for `web`, `api`, `internal` and `install` so that the codes will be more clear and no magic .
- [x] Use https://github.com/unrolled/render instead of macaron's internal render
- [x] Use https://github.com/NYTimes/gziphandler instead of https://gitea.com/macaron/gzip
- [x] Use https://gitea.com/go-chi/session which is a modified version of https://gitea.com/macaron/session and removed `nodb` support since it will not be maintained. **BREAK**
- [x] Use https://gitea.com/go-chi/captcha which is a modified version of https://gitea.com/macaron/captcha
- [x] Use https://gitea.com/go-chi/cache which is a modified version of https://gitea.com/macaron/cache
- [x] Use https://gitea.com/go-chi/binding which is a modified version of https://gitea.com/macaron/binding
- [x] Use https://github.com/go-chi/cors instead of https://gitea.com/macaron/cors
- [x] Dropped https://gitea.com/macaron/i18n and make a new one in `code.gitea.io/gitea/modules/translation`
- [x] Move validation form structs from `code.gitea.io/gitea/modules/auth` to `code.gitea.io/gitea/modules/forms` to avoid dependency cycle.
- [x] Removed macaron log service because it's not need any more. **BREAK**
- [x] All form structs have to be get by `web.GetForm(ctx)` in the route function but not as a function parameter on routes definition.
- [x] Move Git HTTP protocol implementation to use routers directly.
- [x] Fix the problem that chi routes don't support trailing slash but macaron did.
- [x] `/api/v1/swagger` now will be redirect to `/api/swagger` but not render directly so that `APIContext` will not create a html render.
Notices:
- Chi router don't support request with trailing slash
- Integration test `TestUserHeatmap` maybe mysql version related. It's failed on my macOS(mysql 5.7.29 installed via brew) but succeed on CI.
Co-authored-by: 6543 <6543@obermui.de>
* use certmagic for more extensible/robust ACME cert handling
* accept TOS based on config option
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lauris BH <lauris@nix.lv>
* Use mount but not register for chi routes
* Fix test
* Fix test
* Fix test
* Fix comment
* turn back unnecessary change
* Remove the timout middleware since some operations may spend much time.
* Fix --port setting
Unfortunately there was an error in #13195 which set the --port
option before the settings were read. This PR fixes this by
moving applying this option to after the the settings are read
However, on looking further into this code I believe that the setPort
code was slightly odd.
Firstly, it may make sense to run the install page on a different
temporary port to the full system and this should be possible with
a --install-port option.
Secondy, if the --port option is provided we should apply it to both
otherwise there will be unusual behaviour on graceful restart
Thirdly, the documentation for --port says that the setting is
temporary - it should therefore not save its result to the configuration
(This however, does mean that authorized_keys and internal links may
not be correct. - I think we need to discuss this option further.)
Fix#13277
Signed-off-by: Andrew Thornton <art27@cantab.net>
* Update cmd/web.go
* Apply suggestions from code review
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
#12391 offered to change the default PID file from /var/run/gitea.pid however in discussion it was decided that this could break users of older systems. An alternative was offered that we could make the PID file compile/link time settable.
This PR does this, and changes the name of the setting from CustomPID to simply PIDFile. It also updates the from-source docs to show how to change the compiler settings to do this.
Closes#12391
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: Florian Klink <flokli@flokli.de>
* Change graceful to use a singleton obtained through GetManager instead of a global.
* Graceful: Make TestPullRequests shutdownable
* Graceful: Make the cron tasks graceful
* Graceful: AddTestPullRequest run in graceful ctx
* Graceful: SyncMirrors shutdown
* Graceful: SetDefaultContext for Xorm to be HammerContext
* Avoid starting graceful for migrate commands and checkout
* Graceful: DeliverHooks now can be shutdown
* Fix multiple syncing errors in modules/sync/UniqueQueue & Make UniqueQueue closable
* Begin the process of making the repo indexer shutdown gracefully
* Graceful: Create callbacks to with contexts
* Graceful: Say when Gitea is completely finished
* Graceful: Git and Process within HammerTime
Force all git commands to terminate at HammerTime
Force all process commands to terminate at HammerTime
Move almost all git processes to run as git Commands
* Graceful: Always Hammer after Shutdown
* ProcessManager: Add cancel functionality
* Fix tests
* Make sure that process.Manager.Kill() cancels
* Make threadsafe access to Processes and remove own unused Kill
* Remove cmd from the process manager as it is no longer used
* the default context is the correct context
* get rid of double till
* Graceful: Allow graceful restart for fcgi
My previous interpretation was incorrect - we do not handle sockets
being passed in over stdin
* Update web.go
* Graceful Shutdown for windows and others
Restructures modules/graceful, adding shutdown for windows, removing and
replacing the old minwinsvc code.
Creates a new waitGroup - terminate which allows for goroutines to
finish up after the shutdown of the servers.
Shutdown and terminate hooks are added for goroutines.
* Remove unused functions - these can be added in a different PR
* Add startup timeout functionality
* Document STARTUP_TIMEOUT
* Only attempt to kill parent once
* Apply suggestions from code review
Co-Authored-By: guillep2k <18600385+guillep2k@users.noreply.github.com>
* Add waitgroup for running servers
* Prevent deadlock in indexer initialisation during graceful restart
* Move from gracehttp to our own service to add graceful ssh
* Add timeout for start of indexers and make hammer time configurable
* Fix issue with re-initialization in indexer during tests
* move the code to detect use of closed to graceful
* Handle logs gracefully - add a pid suffix just before restart
* Move to using a cond and a holder for indexers
* use time.Since
* Add some comments and attribution
* update modules.txt
* Use zero to disable timeout
* Move RestartProcess to its own file
* Add cleanup routine
