pgRouting/pgrouting

Public chat group of the pgRouting repository.

Vicky Vergara

@cvvergara

999998, 3

SELECT ARRAY[1,2,3,4,5 ..... 1000000] AS sources, ARRAY[1000000, ...... 5,4,3,2,1] as targets

Ashish Kumar

@krashish8

When are we having the next meeting? I'm sorry I didn't listen.
Is it Tomorrow Tuesday (19 May), 2pm UTC = 9am in Mexico = 7:30pm in India? @cvvergara

Vicky Vergara

@cvvergara

I didn't set up a date, but just be attentive for the last minute annoucement

Ashish Kumar

@krashish8

Okay. Thanks!

Vicky Vergara

@cvvergara

@krashish8 @rajhim2 @prakashupes please review pgRouting/pgrouting#1349

Vicky Vergara

@cvvergara

Topic: updating your working branch

The master branch has been updated, so you need to update the working branch

work on your GSoC-pgRouting clone
suppose remotes are:
- origin your GSoC-pgRouting fork
- upstream pgRouting repo
- gsoc the GSoC-pgRouting repo

Steps:

If you haven't started coding:

git fetch gsoc
git checkout gsoc/<your branch> 
git checkout -b <branch name>
git fetch upstream
git rebase upstream/master
git push --set-upstream origin <branch name>

And make a PR from <branch name> to gsoc/<your branch>

If you already started coding on branch: <branch name>
```
git checkout <branch name>
git fetch upstream
git rebase upstream/master
git push
```
And make a PR from <branch name> to gsoc/<your branch>

I will check those PR (in my) tonight

end of topic

Vicky Vergara

@cvvergara

@rajhim2 why are you reverting the merge of the PR on the main repository

Ashish Kumar

@krashish8

@cvvergara I think we should force push in case "If you already started coding on branch: <branch name>"
git push -f
This is because rebasing would reorder the commits history, therefore the remote branch can't be fast-forwarded to local, without a force push.

Vicky Vergara

@cvvergara

yeah I forgot the -f, but that doesnt try to revert the master branch of the main repository

@krashish8 can you make the PR

Ashish Kumar

@krashish8

@cvvergara Done!

Vicky Vergara

@cvvergara

Thanks, The merge will be done before mailing a report next weekend
The opened PR will allow to follow your work

Ashish Kumar

@krashish8

Okay, understood!

Himanshu Raj

@rajhim2

@rajhim2 why are you reverting the merge of the PR on the main repository

It was clicked by mistake mam.

Ashish Kumar

@krashish8

@cvvergara There are some extra commits made by transfix-integration [bot]. So, we should rebase the master branch of pgrouting repository again, right? (repeating the above steps)

Himanshu Raj

@rajhim2

Can anyone please explain me why is this build failing. https://travis-ci.org/github/rajhim2/GSoC-pgRouting/jobs/689126458 This repository is on my profile rajhim2.

Himanshu Raj

@rajhim2

The command "if [ "$DOCUMENTATION" == "OFF" ] && [ "$BUILD_TYPE" == "Debug" ]; then bash ./ci/travis/doc-quieries_test.sh $POSTGRESQL_VERSION postgres; fi" exited with 1.

Himanshu Raj

@rajhim2

Travis Build: https://travis-ci.org/github/rajhim2/GSoC-pgRouting/jobs/689202922

Vicky Vergara

@cvvergara

https://www.osgeo.org/foundation-news/release-of-pgrouting-3-0-0/

Daniel Kastl

@dkastl

Thank you @cvvergara !

Next I think it's on my to encourage a few more pgRouting businesses to financially contribute:
https://funding.communitybridge.org/projects/pgrouting

Rajat Shinde

@omshinde

Thank you @cvvergara . Congratulations to the team. :)

Vicky Vergara

@cvvergara

Next meeting: tuesday May 26 at 2pm UTC = 9am in Mexico = 7:30pm in India

Himanshu Raj

@rajhim2

Why the same build passes on appveyor and fails on travis-ci.

https://travis-ci.org/github/rajhim2/GSoC-pgRouting/jobs/691040785

https://ci.appveyor.com/project/rajhim2/gsoc-pgrouting-twjjp/builds/33105985

Ashish Kumar

@krashish8

Why the same build passes on appveyor and fails on travis-ci.

@rajhim2 I don't have any idea how it passes on appveyor, however, there's an error in your added files.

In the file src/kargersContraction/kargersContraction.c, the function's name shall be in all lowercase, as PostgreSQL converts all unquoted names to lowercase (_pgr_kargerscontraction instead of _pgr_kargersContraction).

You need to change this in all the three lines:

https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L43
https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L44
https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L116

Himanshu Raj

@rajhim2

Why the same build passes on appveyor and fails on travis-ci.

@rajhim2 I don't have any idea how it passes on appveyor, however, there's an error in your added files.
In the file src/kargersContraction/kargersContraction.c, the function's name shall be in all lowercase, as PostgreSQL converts all unquoted names to lowercase (_pgr_kargerscontraction instead of _pgr_kargersContraction).
You need to change this in all the three lines:
https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L43
https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L44
https://github.com/rajhim2/GSoC-pgRouting/blob/him/src/kargersContraction/kargersContraction.c#L116

Thank you @krashish8 ! It worked like a charm.

Himanshu Raj

@rajhim2

Mahmoud will open 1 issue:

For the functionality request of the pgr_dijkstra with the combinations sql

the issue has an example

step 2) Other issue (maybe there is already one that talks about using ctlr-C and in the comment put the possible solution.

Then once we located both issues, Mahmoud will open a PR with tilte "Closes #xx and closes #yyy" with a small comment on the message

After that, @cvvergara will open an issue about using the method for stoping execution on each function.

Himanshu Raj

@rajhim2

Open issues about the code that Mahmoud made

For example: "Normal" flag is not needed

other example: To create pgr_dikstraCost with combinations

Himanshu Raj

@rajhim2

doc/src/pgRouting-introduction.rst

Himanshu Raj

@rajhim2

https://gitter.im/pgRouting/pgrouting?at=5ec3fbbb7da13f3a0ac31d60

Vicky Vergara

@cvvergara

Topic: @mahmsakr preparation for PR

Once that

Contributor name has been added to modified files
Contributor name is on the license of new files
Contributor name is been added to the Contributors List

From the console terminal:

# make sure that you are in the correct branch
git checkout combinationsSQL
# update to latest main repo contents
git fetch upstream
# Put your work on top of what develop
git rebase upstream/develop

Once you have updated the repo, the following that I forgot to mention on our last meeting must be done:

in doc/src/release_notes.rst in the corresponding 3.1.0 release notes add
```
.. rubric:: New functions

* pgr_dijkstra(combinations)
```
Commit and push

From the github website:

make a PR to develop branch
end of topic

Vicky Vergara

@cvvergara

@rajhim2

Why the same build passes on appveyor and fails on travis-ci.

appveyor does not perform the tests it only compiles the tests travis besides compiling does the testing.

Vicky Vergara

@cvvergara

https://workshop.pgrouting.org

Vicky Vergara

@cvvergara

G=(E,V)

Ashish Kumar

@krashish8

I was adding the boost functionality to the function pgr_depthFirstSearch (pgRouting/GSoC-pgRouting#38) I was implementing.
For this I was supposed to use boost::depth_first_search - for directed graphs, and boost::undirected_dfs - for undirected graphs.

In the current implementation, I have used pgrouting::UndirectedGraph(#L117) and called boost::depth_first_search, yet the function gives the correct result, as expected, for the undirected graph. The documentation for boost::depth_first_search says the graph Graph& g should be directed, however in the current implementation, I use undirected graph with it, and it is working alright.

Moreover, in the pgrouting, we have the functions pgr_primDFS and pgr_kruskalDFS, both of which work only for Undirected Graphs, but in the implementation, I see that the function calls the boost::depth_first_search which is supposed to take only Directed Graphs as input.

As the current implementation works good on calling boost::depth_first_search for both directed and undirected graphs, I don't consider the need to use boost::undirected_dfs as everything is working fine.

Anyone has any idea why does the boost::depth_first_search still works for undirected graphs?

Vicky Vergara

@cvvergara

Yes, I have an idea, because you are using this general graph class that I created so you get the graph that you need:
https://docs.pgrouting.org/doxy/3.0/classpgrouting_1_1graph_1_1Pgr__base__graph.html

So, with the if-then-else you need to make sure you are calling to the correct boost function

Vicky Vergara

@cvvergara

TOPIC: Interruptions

(Not for the GSoC students)
The branch https://github.com/mahmsakr/pgrouting/tree/combinationsSQL is almost ready and to merge the interruptions to that branch the following has to be done:

Use the website

Where communities thrive

People

Repo info

Activity

Topic: updating your working branch

end of topic

Topic: @mahmsakr preparation for PR

end of topic

TOPIC: Interruptions