I contributed a few days of work to an open source project. It was merged into a sub-branch, that was then squashed and merged by the library's original author to keep a clean git history, erasing my status as a contributor.

Now when I send people the library and tell them I'm a contributor, they don't believe me, because I've been reduced to a footnote in a commit message.

:(

Maybe when a squash and merge is done through the UI at least, the information about who contributed those changes could be tracked.

I think there's a bug in your new feature from August 15th. The diff showed 5 file changes but it introduced a lot more. When you revert the merging commit, there are 26 changed files. Causes confusion and headache when investigating. I'd rather not have the feature appear at all, vs. having to solve the mystery of why previously reverted commits are showing up in master.

All of the API v3 endpoints on api.github.com support CORS (https://developer.github.com/v3/#cross-origin-resource-sharing): however, the github.com endpoints involved in creating an OAuth2 access token via the Web Flow, specifically https://github.com/login/oauth/access_token, don't support the OPTIONS method or CORS headers necessary to POST to it and get the access_token back.

This is a bug. Denying cross-origin access here doesn't help security, as it is most easily worked around by doing the POST requests through a potentially-insecure third-party open reverse proxy (as implemented using https://cors-anywhere.herokuapp.com/https://github.com/login/oauth/access_token on http://stuartpb.github.io/gistachio/demo.html).

samtools/hts-specs#339 was a pull request consisting of a single commit:

commit 28c558e0bc7e112b8b7674baf9da5c5cdbd860f8 (range-desc)
Author:     John Marshall <John.W.Marshall@glasgow.ac.uk>
AuthorDate: Fri Sep 14 22:26:33 2018 +0100
Commit:     John Marshall <John.W.Marshall@glasgow.ac.uk>
CommitDate: Fri Sep 14 22:26:33 2018 +0100

It was merged using the GitHub UI "Squash and merge" button (so that it would be rebased onto the current tip of origin/master) leading to the following commit on the main repository's master:

commit 98de888fb25b856fac74ae27023f1f5300752240 (origin/master, origin/HEAD)
Author:     John Marshall <jmarshall@users.sourceforge.net>
AuthorDate: Wed Sep 19 09:41:21 2018 +0100
Commit:     Andrew Yates <andrewyatz@users.noreply.github.com>
CommitDate: Wed Sep 19 09:41:21 2018 +0100

As expected, the Committer and CommitDate have been altered to reflect the person who pressed the merge button.

The AuthorDate has been altered by GitHub to be the same as CommitDate, but this is incorrect — the commit was still originally authored on Friday 14th.

The AuthorEmail has been altered by GitHub to be the default email address associated with the @jmarshall GitHub account at the time. This is very incorrect, as I do not use this email address when doing work on samtools/hts-specs.

When the "Squash and merge" button is being used to squash several commits from different authors, there is an argument to be had that Author/AuthorDate could be updated in some way (or GitHub could just follow what git does in this case).

When the "Squash and merge" button is being used to squash several commits from the same author (so AuthorName and AuthorEmail are the same, which is trivially true when there is only one commit), it is a bug to change the author's listed AuthorName or AuthorEmail.

If a pull request branch is updated via a push (force or otherwise), it should invalidate any approvals submitted via the new review system, because the update could introduce new changes which the previous approvers might not approve. Currently this invalidation does not happen, which means that there is a high risk of these new changes getting merged even though they were never reviewed. This significantly weakens the reliability of the review system. Gerrit is an example of a code review system which handles this correctly: only the reviews on the latest patch set are taken into account.

As a result of this approach, Gerrit additionally invalidates disapprovals as well as approvals. There are arguments both for and against that behaviour (and hence whether it should be replicated on GitHub):

If someone submits a review of an early version of a PR which requires changes (i.e. a -1 vote), then if that disapproval is not reset when a new version of the PR is pushed, it means that a follow-up re-review from that specific person is required before the PR can be merged. That has one advantage of ensuring that the concerns they expressed are fixed by the new version; however there is a risk that that person may go AWOL and leave the PR blocked forever. In that case, a workaround might be to close the PR and resubmit it. This might seem like an overly noisy approach, but OTOH it might be good to make the process of effectively manually overriding someone's disapproval a noisy operation.

Normally when I push a branch I'll see a message like this in my Slack:

Now if I push a branch called ":shower:" I'll see this instead:

And when I go to the home page of my repo it mentions that I pushed branches, but doesn't show the name:

Clicking the "Compare & Pull Request" button doesn't work; I have to manually open a PR for the branch the hard way by looking through the list of all my branches and finding the appropriate one.

Anyways... I don't know the full extent of this issue (whether it extends to unicode characters in general or not). But I would like if GitHub better-supported emoji names for branches since a picture is worth a thousand words, as they say.

What the GUI exposes

When browsing the source code in a repository, GitHub offers a button that says "Create New File":

Clicking on it brings the user to a page that looks something like this:

The specific link is as follows:

https://github.com/NetTopologySuite/NetTopologySuite/new/develop/NetTopologySuite

What we found

The dotnet/docfx developers discovered that you can use query string parameters to pre-fill some of the values for you.

• This allows that project to give end users a link that will let them overwrite the auto-generated documentation and/or add more information that isn't in the code, by creating files in an "overwrite" folder with the machine-readable name and initial contents so that the tool will be able to see that on the next run and incorporate it into the documentation site.

Specifically, the value query string parameter fills the main text box, and the filename query string parameter can be used to fill in the filename.

Issues

The value parameter seems to work just fine, but there are some issues with filename.

Goes up a level

When filename is specified at all, the last folder part of the URL seems to be ignored. Compare the following URL to the URL from above:

https://github.com/NetTopologySuite/NetTopologySuite/new/develop/NetTopologySuite?filename=Callees.cs&value=public+static+class+Callees%0A%7B%0A++++public+const+string+PrimaryCallee+=+%22Ghostbusters%22%3B%0A%7D

It takes you to a page that looks like this (compare the folder path to what we had above):

Workaround 1 breaks "cancel" link

The DocFX folks also saw this, but they found a workaround: append some dummy path after it like "new", so that it puts you into the same folder.

Compare this URL:

https://github.com/NetTopologySuite/NetTopologySuite/new/develop/NetTopologySuite/new?filename=Callees.cs&value=public+static+class+Callees%0A%7B%0A++++public+const+string+PrimaryCallee+=+%22Ghostbusters%22%3B%0A%7D

That brings us here:

However, clicking the encircled "cancel" link takes us to a nonexistent page. See discussion and other details in dotnet/docfx#3156:

https://github.com/NetTopologySuite/NetTopologySuite/tree/develop/NetTopologySuite/new

Workaround 2 doesn't stick through "cancel" and "back button"

So I had the idea that, instead of appending some dummy that doesn't exist, we could try duplicating the last folder in the filename query string parameter:

https://github.com/NetTopologySuite/NetTopologySuite/new/develop/NetTopologySuite?filename=NetTopologySuite/Callees.cs&value=public+static+class+Callees%0A%7B%0A++++public+const+string+PrimaryCallee+=+%22Ghostbusters%22%3B%0A%7D

That brought me here, and it seemed perfect!

"Cancel" worked just fine, bringing me back to the folder that would have contained the file. Then I clicked on my browser's "Back" button (Firefox 65), and:

...darn. It seems to have ignored the path information from the filename parameter for some reason.

What we really need

What we need is a supported way to pre-fill some of the information on the "Create New File" page in order to be able to help users contribute incremental, targeted improvements to documentation that can be consumed by a tool.

Currently, the person who opened the PR cannot self-approve it. However someone else can push code to the same branch, approve, and merge (bypassing branch protection meant to require review by another person before merging).

We'd love a way to prevent this through Github. We found a probot app which does basically what we want, but it would be nice to not have to host this infrastructure ourselves: https://github.com/Raul6469/self-approve

Please fix this issue "Without adding a message to the release, it only counts as a "tag" and not as a full release". So I'm not getting "releases" notifications for projects that not include a release messages.

Minimal synthetic examples:

• https://github.com/cirosantilli/test-adoc-readme-large | http://web.archive.org/web/20190701071234/https://github.com/cirosantilli/test-adoc-readme-large
• https://github.com/cirosantilli/test-adoc-readme-large/blob/master/README.adoc | http://web.archive.org/web/20190701071238/https://github.com/cirosantilli/test-adoc-readme-large/blob/master/README.adoc
• https://github.com/cirosantilli/test-md-readme-large | http://web.archive.org/web/20190701071248/https://github.com/cirosantilli/test-md-readme-large
• https://github.com/cirosantilli/test-md-readme-large/blob/master/README.md | http://web.archive.org/web/20190701071256/https://github.com/cirosantilli/test-md-readme-large/blob/master/README.md

On the above examples, it can be seen that the README is cut off on the repo main pages at: 00000000031999, which according to wc is exactly 512000 bytes.

The blob view however shows up to the correct value of 00000000063999. Edit: as of 2020-06-04 there was an update, and now the blob view just says "Sorry about that, but we can’t show files that are this big right now.", which I'm fine with as it is a clear result. I'd rather have the cropped readme with a clear warning though.

Please either render the full thing, or give a clear warning somewhere that the README is being cut down.

I wonder if it's truly not technically feasible to render it somehow, since the corresponding github pages of those websites render fine: https://cirosantilli.com/linux-kernel-module-cheat/ I know this is a dynamic website, but maybe there's a way.

Real world examples:

• https://github.com/2019ncovmemory/nCovMemory "由于Github的错误，在初始页面README底部不会完整显示。请移步我们的网站或完整README阅读。"
• cirosantilli/linux-kernel-module-cheat#75

Oh, and the blame unicorns "This page is taking way too long to load": https://github.com/cirosantilli/linux-kernel-module-cheat/blame/d6d7f15c912ed6371c8cf15a93f43488d6fc5efb/README.adoc (archive).

I was just looking at the most starred repos of all time and used the following command:

https://github.com/search?o=desc&q=stars%3A%3E0&s=stars&type=Repositories
This search produces inconsistent results, where a library may not be shown.

Notice that the first image has both vue and react but the second one doesn't have vue.

I refreshed the pages several times and the issue persisted. Checked the console but there was no error.

Is it just me or did somebody else also noticed this as well?

This is a bug, and I haven't reported it to Github yet, because, given the caliber of their response to the last eight bugs I've reported to them, I'm certain they're just going to brush me off, misunderstand the issue without reading it, blame me, and deny anything's wrong, and I just do not have the energy to deal with that crap right now.

What I have below is the draft of what I would send to GitHub, if I weren't absolutely certain it'd end up in nothing but headaches for me.

The issue

Requesting a user's organization membership (see the docs) with a personal access token gives you the complete list of all organizations the user is a member of, public or private. This is good. This is what I want.

Requesting the same organization membership with a token obtained through an OAuth application returns only the organizations that the user allowed the application access to - not the ones they have private membership to (which are included with a personal access token), and not even the ones their membership is already public for.

Considering that part of the dialog when granting access says that the application will be able to read "public and private organization membership" or something like that, I'd say this is pretty messed up. I need API access to all the organizations I'm a member to as they emerge, and I don't want to have to go through the agony of re-approving every organization, to every application. every time I join a new org, just to keep this current.

A .tsv file that contains quotes triggers this error message:
"We can make this file beautiful and searchable if this error is
corrected: Illegal quoting in line 2."
Example: https://github.com/julianladisch/tsv/blob/master/example.tsv

However, quotes in .tsv files do not have any special meaning so there cannot be any illegal quoting.
See these references for .tsv files:
https://en.wikipedia.org/wiki/Tab-separated_values
https://www.iana.org/assignments/media-types/text/tab-separated-values

I would be very pleased if you can change the .tsv rendering accordingly.

• https://github.com/cirosantilli/test-octopus-100k | http://archive.is/gsR8I 500
• master: cirosantilli/test-octopus-100k@8582d01 | http://archive.is/5JwXq hangs for a while, and then cloudfare. OMG, am I / archive.is going to be blacklisted?
• one of the 100k: cirosantilli/test-octopus-100k@20c0610 | http://archive.is/xsboa works
• root: cirosantilli/test-octopus-100k@95bb02e | http://archive.is/yuQoR works
• repo generated with: https://github.com/cirosantilli/test-git-web-interface/blob/9374e5e18217982a13e354d2812d34a831ee3558/other-test-repos/many-commits-octopus.py

If the hanging page is killing your server, feel free to delete that repo while you fix the issue. I didn't think it would be that bad.

And Linus called a 66-way merge a Cthulhu merge... too young too naive: https://softwareengineering.stackexchange.com/questions/314215/can-a-git-commit-have-more-than-2-parents/377903#377903

Update 2018-11-15: https://github.com/cirosantilli/test-octopus-100k dit not 500 once for the first time. The head commit with many children took forever but did load after a few minutes! Children did not show on UI though: cirosantilli/test-octopus-100k@8582d01

Update 2019-01-13: OK, now cirosantilli/test-octopus-100k@8582d01 opens, but keeps loading the list of children forever.

Update 2019-05: nevermind, unicorn is back after a while of loading.

Update 2020-12-04: Pushing 1M failed with "Connection to github.com closed by remote host. fatal: the remote end hung up unexpectedly".

I have noticed that at the bottom of all GitHub comment boxes there is a GitHub Markdown link that takes the user here: https://guides.github.com/features/mastering-markdown/

If the user scrolls down to the very bottom of the GitHub Mastering Markdown page to "Emoji" there is a link that takes the user to an "Emoji Cheat Sheet" page here: https://www.webpagefx.com/tools/emoji-cheat-sheet/

Issue #1:

• The issue is that there is no :simple_smile: available in GitHub comments.

Issue #2:

• A lot of the emoji's on this list do not match what the user can do in a GitHub comment.

Please resolve this and make this consistent. It will go a LONG way with community building as Emoji's are a standard of communication in the age we live in.

Thank you.

For example, the following .rst will not render correctly:

a directory of `reStructuredText <restructuredtext_>`_ files.

.. _restructuredtext: http://docutils.sourceforge.net/rst.html

The link should point to docutils.sourceforge.net; instead, it points to ./restructuredtext_.

It’s imposible to use astral (= supplementary) Unicode symbols in comments on GitHub. This is very annoying, since lots of software has bugs with these symbols, and this GitHub bug makes it impossible to write a decent bug report about it.

Astral code points range from U+010000 to U+10FFFF. Try U+1D306 TETRAGRAM FOR CENTRE or U+1F4A9 PILE OF POO, for example.

Reproduce:

1. Open the pull requests page: https://github.com/joyent/node/pulls
2. Click on a pull request. Shows the "Discussion" tab.
3. Click the "Commits" tab. The url changes, and the page updates.
4. Click the "Back" button.

Expect: Return to the "Discussion" tab.
Actual: Returns to the https://github.com/joyent/node/pulls page

There is a bug in parsing two the same links in different ways.

[my path](path/to/doc.md)
- [my path](path/to/doc.md)

Result:
my path

• my path

In the list expected the same link as in regular text block.

It will show a bug on using links from repository root page in README.md file.
So, in the list an added link will not contain blob/master/ in the beginning.

1. change profile url to http://piér.re
2. link doesn't show up on save
3. change profile url to http://pi%C3%A9r.re
4. link shows up

When a notification comes from discussions section, the API doesn't populate the link correctly and it's null when we receive it.

Confirmed by @codebytere at gitify-app/gitify#501

This wiki was created 4 years ago so that might be the problem or possibly there is some malformed content within the pages.

See screenshots of wiki pages sidebar and cloned result

The repo: https://github.com/cirosantilli/test-commit-many-parents-1m

The 502: cirosantilli/test-commit-many-parents-1m@e57a199

Parent list should be paged / capped.

ShareX/ShareX@v13.4.0...master#files_bucket is not
https://github.com/ShareX/ShareX/compare/v13.4.0...master#files_bucket

Hi there,

I'd like to raise this issue on Github when I upload files through the platform.

When I go to "Add file" drop-down menu > Upload files" and upload a file, I see the progress bar showing the file is being uploaded, however the tool gives the option to "Commit change" while the file is still being uploaded, which takes me to a possible unhandled page error showing:
"error page (HTTP ERROR 400): This page isn't working at the momentIf the problem continues, contact the site owner."

When this issue happens, I also need to return to the previous page and upload the file again.

I imagine that is not very common for people do this in the platform but I'd like to suggest that the "Commit changes" button get disabled while the upload is still occurring, and enable only after the upload is complete to avoid this issue.

Thanks

I want to search only issues (normal issues, not PRs) for a given repo or organization.

Unfortunately is:issue finds both PRs and issues.

So I don't know which query I should type and whether it's possible.

See image

See title

This is most easily demonstrated by viewing the "rich diff" of a commit that works around this bug by switching to 1-indexing: OctoLinker/OctoLinker@9b2119b?short_path=6a33714#diff-6a3371457528722a734f3c51d9238c13

The source diff shows the unrendered changes: OctoLinker/OctoLinker@9b2119b#diff-6a3371457528722a734f3c51d9238c13

https://developer.github.com/v3/git/refs/#create-a-reference

Look in the parameters type. The type of some parameters is type when it should be string.

I am also currently testing if there is a technical issue with this API call.

Working in a repository "git push origin master" didn't work once, some seconds later it worked.

$ git push origin master
Counting objects: 8, done.
Delta compression using up to 8 threads.
Compressing objects: 100% (8/8), done.
Writing objects: 100% (8/8), 756 bytes | 0 bytes/s, done.
Total 8 (delta 5), reused 0 (delta 0)
remote: Resolving deltas: 100% (5/5), completed with 5 local objects.
remote: ruby-jemalloc: symbol lookup error: /data/github/current/vendor/gems/2.1.7/ruby/2.1.0/gems/json-1.8.3/lib/json/ext/generator.so: undefined symbol: rb_data
_typed_object_alloc
To git@github.com:pasosdeJesus/sal7711_gen
! [remote rejected] master -> master (pre-receive hook declined)
error: failed to push some refs to 'git@github.com:pasosdeJesus/sal7711_gen'

$ git remote -v
origin git@github.com:pasosdeJesus/sal7711_gen (fetch)
origin git@github.com:pasosdeJesus/sal7711_gen (push)

$ date
jue ago 18 05:28:34 COT 2016

(this error happened like 5 minutes ago).

I came across a potential bug of the web-based diff tool in github where for one specific diff the web-based diff tool somehow shows something completely different (in fact a NOP) than using the cmd-line "git diff" which actually should be the gold standard.

See here:
Thinstation/thinstation@b6b8b13

Using cmd-line "git diff" for this commit show the following diff, however:

# git diff b6b8b13c4e22955d1cc022b53a07ed4d590d732e ts/build/build
diff --git a/ts/build/build b/ts/build/build
index d3f1c4d..e7efd65 100755
--- a/ts/build/build
+++ b/ts/build/build
@@ -995,7 +995,7 @@ do
                done
        fi
        if [ `make_caps X$args` == "XFORCE" ]; then
-               echo $name >> ./tmp-tree/etc/modules
+               echo $name `echo $args | cut -f1 -d#` >> ./tmp-tree/etc/modules
        fi
        ;;
     module_pkg)
@@ -1523,7 +1523,7 @@ for filename in `ls -1 $PKGDIR/` ; do
                rm -Rf build
        fi
        echo "Building $filename.pkg"
-       tar -cJf ../$filename.pkg *
+       tar -cz * > ../$filename.pkg
        cd ../../..
        rm -Rf $PKGDIR/$filename
        if [ -n "$PKGFILES" ] ; then

• stars:1: works, good https://github.com/search?utf8=✓&q=stars:1&type=Repositories&ref=searchresults
• stars:asdf: "We could not perform this search", good https://github.com/search?utf8=✓&q=stars:asdf&type=Repositories&ref=searchresults
• stars:0x1: "We couldn’t find any repositories matching 'stars:0x1'", bad https://github.com/search?utf8=✓&q=stars:0x1&type=Repositories&ref=searchresults

E.g.:

<http://example.com/ab'cd>

renders to nothing:

http://example.com/ab'cd

but I'd expect a link instead.

Wayback machine here

Further tests: https://github.com/cirosantilli/test-git-web-interface/blob/504d212433cea225871d3e6251fbdf323747cf58/md.md#link || wayback machine

I think those are all valid URLs: http://stackoverflow.com/a/25570235/895245

All work on CommonMark 0.21: http://spec.commonmark.org/dingus/

This came from a real-life application: http://wiki.osdev.org/James_Molloy's_Tutorial_Known_Bugs

Description

Currently, I can't search any username (including mine) in organization that have a lot of members.

For instance, in the EpicGames organization, it will show you no result if you search based on the username.

And if you force accessing via url with query param (https://github.com/orgs/EpicGames/people?query=ezralazuardy), it will give you an 🦄 error.

FYI, This issue actually have been reported here but there is no further bug fix.

Steps to Reproduce

• You must be logged in.
• Just access the EpicGames organization's poeple page.
• Search any member username, and get no response 😔
• Force access via url with query param (https://github.com/orgs/EpicGames/people?query=ezralazuardy)
• Get your 🦄

I am using following url for retrieving issues : https://api.github.com/repos/owner/repo_name/issues?state=open&direction=asc&page=0
For repositories with large number of issues e.g. https://github.com/microsoft/vscode ; when request is sent for fetching issues with direction ascending , it is often failing with exception server-error. If direction is default which is desc , it is working.
For less number of issues , it is working

If you go into the Projects section of a repository and try to drag and drop the columns, the native drag and drop in Google Chrome may cause the drag image to contain extra unnecessary elements. This has been solved in SortableJS as of version 1.8.1. After looking at the source, I believe Github is using an older version of Sortable for this feature.

Here is how this issue may be reproduced:

1. Go to any Projects section that you have permission to sort in
2. Find a column that has so many items that there is a scrollbar, and scroll down a bit without clicking anything
3. Click on a card in that column
4. Try dragging the column itself, the issue will occur

On my GitHub homepage, I see this:

In fact, that first comment was 14 hours ago, much earlier than the creation of the issue listed second (shown as older activity). It was marked as spam "7 minutes ago."

"Commented" is not the proper summary to report for the event "comment was marked as spam."

I'm developing a GitHub App that uses the GitHub Checks API, and I think I might have found a bug:

I originally create the check run as "completed/neutral", and then I mark it as "in_progress", but it seems like both statuses are shown in the UI. Looking at the HTML, it seems like there should be some logic that actually checks if status == "completed" before showing the conclusion.

Hello,
the name file is not changed when it's just type case is changed.
That can cause a trouble when we come back to old version or when other people download the file.
Or maybe I miss something to update the filename when the changement is only a upper case or lowr case.

https://github.com/opennota shows 888 stars, but at https://github.com/stars I see 889 of them.

Text sample:

> Quote 1

> Quote 2

Inside "Preview" tab everything works as expected, each quote is rendered separately:

However, in the actual page we get one quote with merged contents:

Quote 1

Quote 2

In the index page 'github.com' , when I click Tab button , there is a 'a tag' on the left & the content is 'skip to content' .

A PR I sent - prism-break/prism-break#1627 - has some merge conflicts. In the GitHub web interface I see a "Resolve conflicts" button at the bottom (in the place where the green "Merge" button would appear). However, clicking this button takes me to https://github.com/nylira/prism-break/pull/1627/conflicts, which 404s.

Note that I have no write access whatsoever to the repository in question.

User info is getting bumped into the next line and into user's icons. Makes it unreadable and visually unappealing. Example of this behavior shown below.

It used to be the case that comments made by collaborators on issues and pull requests received a label or badge that indicated they were collaborators similar to owners/members. However, this appears to have been removed. This occurred at the same time owners were renamed to members so may have been a regression that slipped into that change.

I have raised this with GitHub Support via email and recommend anyone interested in this does the same. Currently, they are looking into it.

Steps to reproduce:

1. Have a working $organization.github.io repository
2. Have a working apex custom domain, with both @ and www records set in the DNS with A records, with the CNAME file containing the non-www version of the domain
3. Rename the organization
4. $organization.github.io will redirect to the custom domain, but at that point the Pages server returns HTTP 404.

Note that forcing a Pages rebuild fixed this issue.

Just let me smile! 💃