[SOLVED] AUR push of git reading non existant lines from .SRCINFO

SampsonCrowley · 2017-07-21 18:58:03

remote: error: The following errors occurred when parsing .SRCINFO in commit
remote: error: ab5a7ca68bbafbda50cd8836f3ed046e1d3950c1:
remote: error: line 21: failed to parse line: "'6f1996c211e6f512289565e162752b04'"
remote: error: line 22: failed to parse line: "'74e1450ec40a5915341d7901736aca45'"
remote: error: line 23: failed to parse line: "'b11938850095c777c80238ed878e158d')"

.SRCINFO

pkgbase = pgpool-ii
        pkgdesc = Middleware that works between PostgreSQL servers and a PostgreSQL database client.
        pkgver = 3.6.4
        pkgrel = 2
        url = [url]http://www.pgpool.net[/url]
        arch = i686
        arch = x86_64
        license = custom
        depends = openssl
        depends = postgresql-libs
        replaces = pgpool
        options = !libtool
        backup = etc/conf.d/pgpool-ii
        source = [url]http://www.pgpool.net/download.php?f=pgpool-II-3.6.4.tar.gz[/url]
        source = pgpool-ii.service
        source = pgpool-ii.conf.d
        source = pgpool-ii.diff
        md5sums = e31eedc9cf00b0b27596449a20f4c83d
        md5sums = 6f1996c211e6f512289565e162752b04
        md5sums = 74e1450ec40a5915341d7901736aca45
        md5sums = b11938850095c777c80238ed878e158d

pkgname = pgpool-ii

git show ab5a7ca68bbafbda50cd8836f3ed046e1d3950c1
commit ab5a7ca68bbafbda50cd8836f3ed046e1d3950c1
Author: Sampson Crowley <REDACTED>
Date:   Fri Jul 21 12:43:18 2017 -0600

    remove prebuilt pkg

diff --git a/.SRCINFO b/.SRCINFO
index 1bf85bb..bdb3699 100644
--- a/.SRCINFO
+++ b/.SRCINFO
@@ -1,7 +1,7 @@
 pkgbase = pgpool-ii
        pkgdesc = pgpool-II is a connection pool server for PostgreSQL.
-       pkgver = 3.2.5
-       pkgrel = 1
+       pkgver = 3.6.4
+       pkgrel = 2
        url = http://www.pgpool.net
        arch = i686
        arch = x86_64
@@ -9,13 +9,19 @@ pkgbase = pgpool-ii
        depends = postgresql-libs
        replaces = pgpool
        options = !libtool
-       backup = etc/conf.d/pgpool-ii
-       source = http://www.pgpool.net/download.php?f=pgpool-II-3.2.5.tar.gz
-       source = pgpool-ii.service
+       backup = etc/pgpool
+       source = http://www.pgpool.net/download.php?f=pgpool-II-3.6.4.tar.gz
        source = pgpool-ii.conf.d
-       sha256sums = aecac952fd3d292584c9aa359d72f89b144c29f45c9f848cb030e46215a814f7
-       sha256sums = 6c797a94c1648890a613e201c2ce2e3e2b602103daecc009ebf189578e8e15a0
+       source = pgpool-ii.diff
+       source = pgpool-ii.service
+       sha256sums = 6ad78b9878d126e8da1e03e8edf3c8d56bb0754fe6432f633e81f6faaf8e6dbf
        sha256sums = 9d0c12d735595424fad799499d50bc4da8b7f1e7ffc2c21317f8824559bb7aea
+       sha256sums = c62fc1e31383b875701ffa40c07a5b9d7c8ff36d4a9c9356e831fab612ca8b20
+       sha256sums = 6c797a94c1648890a613e201c2ce2e3e2b602103daecc009ebf189578e8e15a0
+         '6f1996c211e6f512289565e162752b04'
+         '74e1450ec40a5915341d7901736aca45'
+         'b11938850095c777c80238ed878e158d')
+

 pkgname = pgpool-ii

diff --git a/pgpool-ii-3.6.4-2-x86_64.pkg.tar.xz b/pgpool-ii-3.6.4-2-x86_64.pkg.tar.xz
deleted file mode 100644
index 3758613..0000000
Binary files a/pgpool-ii-3.6.4-2-x86_64.pkg.tar.xz and /dev/null differ
(END)

Last edited by SampsonCrowley (2017-07-21 20:32:07)

WorMzy · 2017-07-21 19:09:55

Please use code tags when pasting output and text file content. Please edit your post to use them.

Also, post the output of

git show ab5a7ca68bbafbda50cd8836f3ed046e1d3950c1

from the directory you are trying to push.

SampsonCrowley · 2017-07-21 19:24:39

It seriously is because it tries to parse commits before the one being pushed??? so I have to go back, and delete the commits that aren't even part of the final push..

that's retarded

also why does the forum use BBCode instead of standard markdown?

Slithery · 2017-07-21 19:30:10

SampsonCrowley wrote:

also why does the forum use BBCode instead of standard markdown?

Because it uses vanilla FluxBB. If you want these forums to support markdown then take it up with the FluxBB devs.

SampsonCrowley · 2017-07-21 19:31:43

I guess we can close this one, I just rebuilt the repo from scratch and it worked

SampsonCrowley · 2017-07-21 19:33:48

And I don't have a problem with it using BBCode, this is just the first forum i've joined that doesn't support markdown so i was curious

eschwartz · 2017-07-21 20:13:12

SampsonCrowley wrote:

It seriously is because it tries to parse commits before the one being pushed??? so I have to go back, and delete the commits that aren't even part of the final push..
that's retarded

What, because people who try checking out an old version of the AUR package should be able to use it without actual syntax errors? The idea of running tests on every commit is not a new one, many software projects do the same thing.

It is hardly retarded, even if it isn't something you personally would choose.

SampsonCrowley wrote:

I guess we can close this one, I just rebuilt the repo from scratch and it worked

Edit your first post, and prepend the word "[SOLVED]" to the thread title.

Last edited by eschwartz (2017-07-21 20:14:26)

SampsonCrowley · 2017-07-21 20:31:32

What, because people who try checking out an old version of the AUR package should be able to use it without actual syntax errors?

a commit is not the same thing as a release. a commit just means I wanted to save my changes where I can access them on any of my boxes before shutting down

eschwartz · 2017-07-21 20:40:12

Please continue to cherry-pick the parts of my sentence that you think you can successfully argue with.

Also consider adding `git commit --amend` to your workflow. (Either that or --fixup|--squash and rebase --interactive --autosquash.)
Abusing commits in order to continue working on your working treee on another box, is, well, sort of abusing git in ways it was not meant to be abused.

SampsonCrowley · 2017-07-21 21:07:28

Please continue to cherry-pick the parts of my sentence that you think you can successfully argue with.

in what way is quoting the part that summarizes the paragraph 'cherry picking'? you're the one who came into argue. i expressed my opinion on the matter and said we can close it, then you came in to try to tell me that per commit testing instead of release testing isn't retarded

not really. even if you work on a dev branch, when you merge it into a master branch, it will still have the commits from your dev branch unless you specifically rebase or squash it. feel free to keep thinking that commits should be releases, but everywhere else that is not correct.

the right way to use git is to commit early and commit often, then release when you are ready. and commiting and pushing so you can work somewhere else is not an `abuse of git` and it absolutely was meant to be used that way. ever heard of pair programming?

the more commits you have to more fine grained control you have over changes you can revert.

Bye now. I won't be coming back here

Last edited by SampsonCrowley (2017-07-21 21:08:18)

eschwartz · 2017-07-21 22:24:34

SampsonCrowley wrote:

in what way is quoting the part that summarizes the paragraph 'cherry picking'? you're the one who came into argue. i expressed my opinion on the matter and said we can close it, then you came in to try to tell me that per commit testing instead of release testing isn't retarded

Mostly because it didn't even obliquely reference the main part of the paragraph.

And yes, I came in to argue over your accusation of retardation on the part of the AUR codebase.

...

Okay. I'm happy for you that that workflow is useful for you. Might also be worth mentioning that the actual developers of git disagree with you (they expect and demand one commit per logically separate change and will tell you to rebase/squash), but... tools were made to be used by their users. Whatever.

Anyway, the "commit early and commit often" paradigm refers to committing new, properly developed features, not borked code that will eventually pass through a compiler without syntax errors once it is finished. Do what you want, even... but there are others who have managed to achieve a workflow where each commit still builds, and new features are hooked up once ready.
So again, your assertion that per-commit testing is *retarded* can be proven wrong.

And the AUR policy is to validate every commit, just like the policy of some projects out there is that every commit must be self-contained.

Git certainly was not "meant" to be used for pair programming, especially as there is no need to commit in order to make sure the person spying over your shoulder can see the code. (If you're doing some sort of virtual pair programming, you would be better off with remote desktop sharing or tmux anyway.)

Last edited by eschwartz (2017-07-21 22:36:00)

Scimmia · 2017-07-22 02:12:39

SampsonCrowley wrote:

It seriously is because it tries to parse commits before the one being pushed??? so I have to go back, and delete the commits that aren't even part of the final push..

I think this is where you're getting tripped up. Of course it doesn't validate commits that aren't part of the final push, but you're not just pushing one commit. You're pushing all of the commits that preceded the current HEAD.

Arch Linux

#1 2017-07-21 18:58:03

[SOLVED] AUR push of git reading non existant lines from .SRCINFO

#2 2017-07-21 19:09:55

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#3 2017-07-21 19:24:39

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#4 2017-07-21 19:30:10

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#5 2017-07-21 19:31:43

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#6 2017-07-21 19:33:48

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#7 2017-07-21 20:13:12

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#8 2017-07-21 20:31:32

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#9 2017-07-21 20:40:12

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#10 2017-07-21 21:07:28

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#11 2017-07-21 22:24:34

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

#12 2017-07-22 02:12:39

Re: [SOLVED] AUR push of git reading non existant lines from .SRCINFO

Board footer