[aur-dev] Proposal: AUR bug fix release
Hey guys. It seems a few little bugs are popping up in AUR. Would it be a good idea to do just a few little patches on the master branch and release 1.5.1? We could do this so that the users have a little better interface while we get into the heavy testing development. Things that could use fixing: 1. JSON 2. Search results out of order. FS#9615 3. Problems searching orphans. [1] 4. Wrong search link for versioned dependencies. FS#8622 5. Updated logo. 6. ? I know we already have patches for 1 and 2. Items 3 and 4 are pretty trivial as well. So what do you think? [1] http://archlinux.org/pipermail/aur-general/2008-February/007109.html
On Thu, Feb 28, 2008 at 12:53 PM, Loui
Hey guys. It seems a few little bugs are popping up in AUR. Would it be a good idea to do just a few little patches on the master branch and release 1.5.1? We could do this so that the users have a little better interface while we get into the heavy testing development.
Things that could use fixing: 1. JSON 2. Search results out of order. FS#9615 3. Problems searching orphans. [1] 4. Wrong search link for versioned dependencies. FS#8622 5. Updated logo. 6. ?
I know we already have patches for 1 and 2. Items 3 and 4 are pretty trivial as well. So what do you think?
[1] http://archlinux.org/pipermail/aur-general/2008-February/007109.html
I could be wrong but it sounds like number 3 is just another bug because of number 2. Unless we really have a problem with the new logo I think a new release could be done with just the first 3 points fixed, these a causing the most problems, have been patched and would put the AUR back into a somewhat stable condition again. -- Callan 'wizzomafizzo' Barrett
On 2/27/08, Callan Barrett
I could be wrong but it sounds like number 3 is just another bug because of number 2. Unless we really have a problem with the new logo I think a new release could be done with just the first 3 points fixed, these a causing the most problems, have been patched and would put the AUR back into a somewhat stable condition again.
This looks like the cause of bug 3. In lib/pkgfuncs.inc in the master branch: do_Orphans is set to "Orphans" (line 757) but the check (line 634) doesn't recognise that value.
On Wed, Feb 27, 2008 at 10:53:11PM -0500, Loui wrote:
Hey guys. It seems a few little bugs are popping up in AUR. Would it be a good idea to do just a few little patches on the master branch and release 1.5.1? We could do this so that the users have a little better interface while we get into the heavy testing development.
Things that could use fixing: 1. JSON 2. Search results out of order. FS#9615 3. Problems searching orphans. [1] 4. Wrong search link for versioned dependencies. FS#8622 5. Updated logo. 6. ?
I know we already have patches for 1 and 2. Items 3 and 4 are pretty trivial as well. So what do you think?
New logo was in testing branch last weekish, so we've got 5 done too. And yes, I think this is a good idea. Let's get these together and *glances at calendar* i could push it out tonight or tomorrow night. Sorry i've been a bit behind applying patches, school's keepin me busy this week, they're all flagged in my inbox for applying. -S
On Wed, 27 Feb 2008 22:23:05 -0600
Simo Leone
Things that could use fixing: 1. JSON 2. Search results out of order. FS#9615 3. Problems searching orphans. [1] 4. Wrong search link for versioned dependencies. FS#8622 5. Updated logo. 6. ?
I know we already have patches for 1 and 2. Items 3 and 4 are pretty trivial as well. So what do you think?
New logo was in testing branch last weekish, so we've got 5 done too.
And yes, I think this is a good idea. Let's get these together and *glances at calendar* i could push it out tonight or tomorrow night. Sorry i've been a bit behind applying patches, school's keepin me busy this week, they're all flagged in my inbox for applying.
-S
Good stuff. I organised a bunch of patches into one place for convenience: http://louipc.dontexist.org/aur-patches These should take care of items 1, 2, 3, and 5
On Thu, 28 Feb 2008 12:36:42 -0500
Loui
Good stuff. I organised a bunch of patches into one place for convenience: http://louipc.dontexist.org/aur-patches
These should take care of items 1, 2, 3, and 5
There's also two patches (0008 0009) that make setting up a DB from scratch a lot easier. There were problems with gendummydata.py and I needed to generate some data to test. I chose to use 'AUR' as the db name because it makes it distinct from the 'aur' user.
And yes, I think this is a good idea. Let's get these together and *glances at calendar* i could push it out tonight or tomorrow night. Sorry i've been a bit behind applying patches, school's keepin me busy this week, they're all flagged in my inbox for applying. Hey hey. How's the school treating you? Will they give you a break in
On Wed, 27 Feb 2008 22:23:05 -0600
Simo Leone
On Sat, Mar 15, 2008 at 05:16:18PM -0400, Loui wrote:
Hey hey. How's the school treating you? Will they give you a break in the near future?
You've got impeccable timing. Happy spring break ;) Patches applied and pushed to testing branch tonight. aur.neotuli.net is running that branch. If nobody breaks it by tomorrow, I'll push it live. -S *sigh* if we nicknamed AUR releases, I might name this one "guilt-trip".
On Mon, 17 Mar 2008 02:53:09 -0500
Simo Leone
You've got impeccable timing. Happy spring break ;) Patches applied and pushed to testing branch tonight. aur.neotuli.net is running that branch. If nobody breaks it by tomorrow, I'll push it live.
*sigh* if we nicknamed AUR releases, I might name this one "guilt-trip".
Orphans searching seems to be broken on your test site, but is working fine on mine when I'm running the same testing branch. It might have something to do with the database. AURMaintainer vs. Maintainer or something. I'm looking into it. Hey don't be so hard on yourself. Just let us know what's up. Happy St. Patty's and Cheers ;)
On Mon, 17 Mar 2008 14:35:47 -0400
Loui
Orphans searching seems to be broken on your test site, but is working fine on mine when I'm running the same testing branch. It might have something to do with the database. AURMaintainer vs. Maintainer or something. I'm looking into it. Hmm nevermind I spoke too soon. Mine is still buggy.
On Mon, 17 Mar 2008 15:06:34 -0400
Loui
Hmm nevermind I spoke too soon. Mine is still buggy. The attached patch fixes one apparent bug. Ooops.
On Mon, Mar 17, 2008 at 02:35:47PM -0400, Loui wrote:
Orphans searching seems to be broken on your test site, but is working fine on mine when I'm running the same testing branch. It might have something to do with the database. AURMaintainer vs. Maintainer or something. I'm looking into it.
Hey don't be so hard on yourself. Just let us know what's up. Happy St. Patty's and Cheers ;)
Whoahhh nelly. pkg_search_page() is one very large steaming pile of [moose, ox, horse, cow] shit. Did you know, for instance, that every search page load since 2005 has been pulling a list of _every_ AUR user out of the DB for _every_ query? I didn't, and I'll be damned if it stays that way. I'm going to wrecking-ball pkg_search_page(), should be getting back to you on that very soon. Poke me with a large stick if I dont 8) -S
On Wed, 19 Mar 2008 02:28:51 -0500
Simo Leone
Did you know, for instance, that every search page load since 2005 has been pulling a list of _every_ AUR user out of the DB for _every_ query? I didn't, and I'll be damned if it stays that way.
I'm going to wrecking-ball pkg_search_page(), should be getting back to you on that very soon. Poke me with a large stick if I dont 8)
Daaaamn. I knew it was pretty bad but I don't remember noticing that one. I'm not a huge fan of the wrecking ball approach anymore though... I've come to realise that it can be a real pain to figure out what's going on and it's a lot easier to test and be confident of small changes. But lots of small changes add up!
On Wed, Mar 19, 2008 at 10:53:41AM -0400, Loui wrote:
Daaaamn. I knew it was pretty bad but I don't remember noticing that one. I'm not a huge fan of the wrecking ball approach anymore though... I've come to realise that it can be a real pain to figure out what's going on and it's a lot easier to test and be confident of small changes. But lots of small changes add up!
Well, a little progress update. The first half hour was spent swearing (alot) and documenting every single request variable the function deals with. This is probably useful to you guys as well... /* pkg_search_page(SID) * outputs the body of search/search results page * * parameters: * SID - current Session ID * preconditions: * package search page has been accessed * request variables have not been sanitized * * request vars: * O - starting result number * PP - number of search hits per page * L - package location ID number * C - package category ID number * K - package search string * SO - search hit sort order: * values: a - ascending * d - descending * SB - sort search hits by: * values: l - package location * c - package category * n - package name * v - number of votes * m - maintainer username * SeB- property that search string (K) represents * values: nd - package name&description * m - package maintainer's username * s - package submitter's username * do_MyPackages - boolean. whether to only include * logged in's packages * do_Orphans - boolean. whether to search packages * without a maintainer * do_Search - boolean. whether to execute a search * * * These two are actually handled in packages.php. * * IDs- integer array of ticked packages' IDs * action - action to be taken on ticked packages * values: do_Flag - Flag out-of-date * do_UnFlag - Remove out-of-date flag * do_Adopt - Adopt * do_Disown - Disown * do_Delete - Delete * do_Notify - Toggle notification */
On Wed, 19 Mar 2008 11:18:17 -0500
Simo Leone
Well, a little progress update. The first half hour was spent swearing (alot) and documenting every single request variable the function deals with. This is probably useful to you guys as well...
Good stuff. That'll be handy.
On Wed, Mar 19, 2008 at 11:18:17AM -0500, Simo Leone wrote:
On Wed, Mar 19, 2008 at 10:53:41AM -0400, Loui wrote:
Well, a little progress update. The first half hour was spent swearing (alot) and documenting every single request variable the function deals with. This is probably useful to you guys as well...
What i've got so far will appear in my working branch within a few hours. It appears to work relatively well, and will be running on my test site. As usual, please break it. -S
participants (3)
-
Callan Barrett
-
Loui
-
Simo Leone