Release 0.10.0

[glensc]

0.9.0 was released Nov 15, 2017, there are 120 commits to master since this tag.

We should make a new release, so that the project does not look stale and the new changes can be used.

Merged branches

• Merge pull request Backend decoupling #244 from 1ma/backend-decoupling
• Merge pull request Fix incorrectly named 'profiler.enable' in config documentation #279 from FreeLSD/patch-1
• Merge pull request Support setting mongodb driver options by config #280 from researchgate/driveroptions
• Merge pull request Flamegraph: Various clean-up in prep for d3.flameGraph update  #215 from Krinkle/flamegraph
• Merge pull request Use non-aggregated data when generating the flamegraph. #219 from perftools/flamegraph-data
• Merge pull request Support for tideways_xhprof, but keep tideways support BC. #223 from beberlei/TidewaysXhprof
• Merge pull request Use collector script provided by xhgui-collector #224 from perftools/use-collector
• Merge pull request add autoload file #225 from Ruesin/master
• Merge pull request add parameter for mongodb cursor #228 from skors/fix/#221_mongodb_cursor
• Merge pull request Update README.md #233 from tm1000/patch-1
• Merge pull request Workdir-independent tests #235 from 1ma/absolute-paths
• Merge pull request Resolve copypaste conflicts #236 from 1ma/resolve-cp-conflicts
• Merge pull request Allow a Run to be imported over the network #230 from comandi/add-profile-import-over-network
• Merge pull request add link to delete run from the main paginated list page #238 from dumityty/delete-runs
• Merge pull request Add link to delete all run from the main paginated list page #239 from dumityty/delete-all-runs
• Merge pull request added mandatory 'cursor' parameter to mongo aggregate method call #242 from 1ma/fix-aggregate-call
• Merge pull request Working on #151 #243 from Mte90/patch-1
• Merge pull request flamegraph: Remove the flamegraph feature #246 from Krinkle/rm-flamegraph
• Merge pull request config: remove unused vars, restore defaults, add documentation #249 from Krinkle/config
• Merge pull request Remove unused delete.twig #250 from Krinkle/rm-unused-twig
• Merge pull request Restore Twig template path #251 from Krinkle/restore-twig
• Merge pull request Fix false positive PHP 7 job in Travis CI #254 from Krinkle/fail-pass
• Merge pull request Use POST for /runs/delete_all and add a confirmation form #252 from Krinkle/post-deleteall
• Merge pull request Require PHP 5.6+, Upgrade PHPUnit #255 from Krinkle/tests
• Merge pull request Fix code coverage reporting #256 from Krinkle/coverage
• Merge pull request Use POST for /run/delete and add confirmation form #257 from Krinkle/post-delete
• Merge pull request Add docker-compose setup #263 from glensc/docker-stab2
• Merge pull request Added new method in config.default file #264 from Mte90/patch-2
• Merge pull request Filter by methods #265 from Mte90/filter-methods
• Merge pull request Set Cache-Control headers for runs list and runs view #270 from Krinkle/runslist-cc
• Merge pull request docs: small typos in readme #272 from guygaver/guygaver/fix-typos-in-readme
• Merge pull request Skipping built-in functions #277 from panychek/feature/skip-built-in
• Merge pull request Update README.md #282 from flip111/patch-1
• Merge pull request env vars for save.handler and pdo #284 from 1ma/pdo-env-vars

[glensc]

@markstory do you have any checklist of goals what should be solved before 0.10.0 is tagged?

The release highlight is probably #244

[Krinkle]

Perhaps it's time for 1.0?

[markstory]

do you have any checklist of goals what should be solved before 0.10.0 is tagged?

No. I've not had much time to devote to this project lately. I'm happy to do a 1.0 release as I think the current state is pretty stable. The next big changes are likely to do large dependency upgrades and potentially stop using mongodb.

Mongodb has long been challenging as many users are not familiar with operating it, and I think for what this project does mongo isn't a great fit. There have been several pull requests to add additional storage adapters, and entirely work around folks not wanting to operate mongodb. This all leads me to believe that other simpler storage systems may be a better fit.

[glensc]

@Krinkle I'd rather polish the multiple backend support a bit. i.e so that the xhgui stays closed box solution that you operate over HTTP via the new client library:

• https://github.com/perftools/php-profiler

using pdo, mongo, etc seems un-neccessary difficulty if you just want to measure your application.

having client layer thin, I don't think people care what xhgui uses internally. just launch the docker container and don't worry about it.

[Krinkle]

That'd be cool yeah, would reduce coupling I suppose. Although it does mean that the XHGui server would no longer be read-only and cleanly isolated network-wise, which is something I currently enjoy, but I'd probably prefer the HTTP solution indeed.

As for Docker, would that be for the web server or the database as well? If both, then it'd likely not be something we'd use at Wikimedia. I'd rather keep the data out that box and treat it as stateless.

Would adding that API be a breaking change though? Or would it be something we add-on in a minor release?

[markstory]

having client layer thin, I don't think people care what xhgui uses internally. just launch the docker container and don't worry about it.

I don't think 'run and forget' is viable for production and most pre-production environments outside of local development.

Supporting both a document store and SQL based storage solutions makes future changes harder as contributors need to be aware of the nuances of both. It also impedes feature work as new features would need to be compatible with both storage models, and often sql/mongo have conceptual differences that are hard to reconcile.

[markstory]

I published 0.10.0, we can always do a 1.0.0 in a bit if things are feeling more settled for other folks.

[glensc]

All right. Thanks for finishing this!