1
true
2008-08-19T23:54:09+01:00
11378
9903
865
fix-count_records
321
8994
resolved
activerecord bug patch
fix count_records
2008-08-25T09:31:01+01:00
17393
Repository
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
This is a patch related to "this thread in rubyonrails-core":http://groups.google.com/group/rubyonrails-core/browse_thread/thread/15d2917c8fe02bec.
In a has many association `count_records` basically counts by SQL. If there are no children in the database it assumes that's enough to say there are no children at all. That's not necessarily the case because the association could have new records in @target.
There are several ways to observe this behaviour, for example rendering a collection that has only new records (via #build) would render nothing at all.
This minimal test depicts the issue as well:
@@@ ruby
post = Post.create
post.comments.build
post.comments.size # => 1
post.comments.size # => 0
@@@
The fix turns out to be simple. I've added a regression test as well.
The documentation I mention in a comment will eventually be merged from "this patch in docrails":http://github.com/lifo/docrails/commit/2ee60b0f70febc8c28a90e8361d9e9c2b86d4833.
This is a patch related to "this thread in rubyonrails-core":http://groups.google.com/group/rubyonrails-core/browse_thread/thread/15d2917c8fe02bec.
In a has many association `count_records` basically counts by SQL. If there are no children in the database it assumes that's enough to say there are no children at all. That's not necessarily the case because the association could have new records in @target.
There are several ways to observe this behaviour, for example rendering a collection that has only new records (via #build) would render nothing at all.
This minimal test depicts the issue as well:
@@@ ruby
post = Post.create
post.comments.build
post.comments.size # => 1
post.comments.size # => 0
@@@
The fix turns out to be simple. I've added a regression test as well.
The documentation I mention in a comment will eventually be merged from "this patch in docrails":http://github.com/lifo/docrails/commit/2ee60b0f70febc8c28a90e8361d9e9c2b86d4833.
<div><p>This is a patch related to <a href="http://groups.google.com/group/rubyonrails-core/browse_thread/thread/15d2917c8fe02bec">this thread in rubyonrails-core</a>.</p>
<p>In a has many association <code>count_records</code> basically counts by SQL. If there are no children in the database it assumes that's enough to say there are no children at all. That's not necessarily the case because the association could have new records in @target.</p>
<p>There are several ways to observe this behaviour, for example rendering a collection that has only new records (via #build) would render nothing at all.</p>
<p>This minimal test depicts the issue as well:</p>
<pre><code class="ruby">
post = Post.create
post.comments.build
post.comments.size # => 1
post.comments.size # => 0
</code></pre>
<p>The fix turns out to be simple. I've added a regression test as well.</p>
<p>The documentation I mention in a comment will eventually be merged from <a href="http://github.com/lifo/docrails/commit/2ee60b0f70febc8c28a90e8361d9e9c2b86d4833">this patch in docrails</a>.</p></div>
1
This is a patch related to "this thread in rubyonrails-core":http://groups.google.com/group/rubyonrails-core/browse_thread/thread/15d2917c8fe02bec.
In a has many association `count_records` basically counts by SQL. If there are no children in the database it assumes that's enough to say there are no children at all. That's not necessarily the case because the association could have new records in @target.
There are several ways to observe this behaviour, for example rendering a collection that has only new records (via #build) would render nothing at all.
This minimal test depicts the issue as well:
@@@ ruby
post = Post.create
post.comments.build
post.comments.size # => 1
post.comments.size # => 0
@@@
The fix turns out to be simple. I've added a regression test as well.
The documentation I mention in a comment will eventually be merged from "this patch in docrails":http://github.com/lifo/docrails/commit/2ee60b0f70febc8c28a90e8361d9e9c2b86d4833.
<div><p>This is a patch related to <a href="http://groups.google.com/group/rubyonrails-core/browse_thread/thread/15d2917c8fe02bec">this thread in rubyonrails-core</a>.</p>
<p>In a has many association <code>count_records</code> basically counts by SQL. If there are no children in the database it assumes that's enough to say there are no children at all. That's not necessarily the case because the association could have new records in @target.</p>
<p>There are several ways to observe this behaviour, for example rendering a collection that has only new records (via #build) would render nothing at all.</p>
<p>This minimal test depicts the issue as well:</p>
<pre><code class="ruby">
post = Post.create
post.comments.build
post.comments.size # => 1
post.comments.size # => 0
</code></pre>
<p>The fix turns out to be simple. I've added a regression test as well.</p>
<p>The documentation I mention in a comment will eventually be merged from <a href="http://github.com/lifo/docrails/commit/2ee60b0f70febc8c28a90e8361d9e9c2b86d4833">this patch in docrails</a>.</p></div>
false
2008-08-19T23:54:09+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
new
activerecord bug patch
fix count_records
2008-08-19T23:54:09+01:00
11378
Xavier Noria
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
2
Patch updated with Fred's proposal. That one is better because we certainly can assert `loaded` anyway.
<div><p>Patch updated with Fred's proposal. That one is better because we certainly can assert <code>loaded</code> anyway.</p></div>
false
2008-08-20T00:22:45+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
new
activerecord bug patch
fix count_records
2008-08-20T00:22:45+01:00
11378
Xavier Noria
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
1
Looks good to me, tests pass. Perhaps the assert should assert that the size is 1 (or is that already covered by existing tests ?
<div><p>Looks good to me, tests pass. Perhaps the assert should assert that the size is 1 (or is that already covered by existing tests ?</p></div>
false
2008-08-20T11:33:52+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
new
activerecord bug patch
fix count_records
2008-08-20T11:33:52+01:00
17477
Frederick Cheung
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
1
Yeah, the test assumes size1 is correct because there are unit tests for #size in test_build and test_collection_size_after_building.
<div><p>Yeah, the test assumes size1 is correct because there are unit tests for #size in test_build and test_collection_size_after_building.</p></div>
false
2008-08-20T11:55:31+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
new
activerecord bug patch
fix count_records
2008-08-20T11:55:31+01:00
11378
Xavier Noria
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
1
I don't think it's necessary to refer to google groups threads in a comment. Either explain the issue in a comment (if it really needs explaining) or don't add a comment. When the patch is commited the added code will already be tied to the commit (git annotate) which will refer to this ticket which already refers to the google groups thread.
Other than that, +1
<div><p>I don't think it's necessary to refer to google groups threads in a comment. Either explain the issue in a comment (if it really needs explaining) or don't add a comment. When the patch is commited the added code will already be tied to the commit (git annotate) which will refer to this ticket which already refers to the google groups thread.</p>
<p>Other than that, +1</p></div>
false
2008-08-20T18:49:33+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
new
activerecord bug patch
fix count_records
2008-08-20T18:49:33+01:00
14890
Tarmo Tänav
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
1
(from [a970f916fb1e05376733e2d42d9bcc2b873af355]) Fix has_many#count_records. [#865 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
http://github.com/rails/rails/commit/a970f916fb1e05376733e2d42d9bcc2b873af355
<div><p>(from [a970f916fb1e05376733e2d42d9bcc2b873af355]) Fix has_many#count_records. [<a href="/projects/8994/tickets/865" title="Ticket #865">#865</a> state:resolved]</p>
<p>Signed-off-by: Pratik Naik <a>pratiknaik@gmail.com</a>
<a href="http://github.com/rails/rails/commit/a970f916fb1e05376733e2d42d9bcc2b873af355">http://github.com/rails/rails/co...</a></p></div>
true
2008-08-21T15:51:04+01:00
11378
---
:state: new
9903
865
fix-count_records
0
8994
resolved
activerecord bug patch
fix count_records
2008-08-21T15:51:04+01:00
17393
Repository
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
1
(from [0048f558e69cfc2675552dd31ea773b2b8271e9b]) Fix has_many#count_records. [#865 state:resolved]
Signed-off-by: Pratik Naik <pratiknaik@gmail.com>
http://github.com/rails/rails/commit/0048f558e69cfc2675552dd31ea773b2b8271e9b
<div><p>(from [0048f558e69cfc2675552dd31ea773b2b8271e9b]) Fix has_many#count_records. [<a href="/projects/8994/tickets/865" title="Ticket #865">#865</a> state:resolved]</p>
<p>Signed-off-by: Pratik Naik <a>pratiknaik@gmail.com</a>
<a href="http://github.com/rails/rails/commit/0048f558e69cfc2675552dd31ea773b2b8271e9b">http://github.com/rails/rails/co...</a></p></div>
true
2008-08-25T09:31:01+01:00
11378
--- {}
9903
865
fix-count_records
0
8994
resolved
activerecord bug patch
fix count_records
2008-08-25T09:31:01+01:00
17393
Repository
Xavier Noria
http://rails.lighthouseapp.com/projects/8994/tickets/865
2.x
97e7c7e94849d411ebc6f07879877e0812c20d52
text/plain
2008-08-20T00:22:45+01:00
fix_count_records_2.diff
40285
2623
11378
http://rails.lighthouseapp.com/attachments/40285/fix_count_records_2.diff