I develop using Rails on win32 in a corporate environment. While most
stuff works, there are a few annoyances that I deal with on a regular
basis. It would be great if stuff like this could get cleaned up.
I have filed a few bugs against Rails that have gotten no attention, and
asked a question on ruby-talk that never got answered (though I posted
via the newsgroup, maybe the gateway is one-way?). I don’t mean to
sound bitter, I’m aware that the problem is most likely that no one
can really test these problems.
Do y’all have any advice regarding what I can do to help get bugs like
these fixed? Are my expectations regarding turnaround time just
unrealistic?
Rails bugs:
Ruby bug (is there a Ruby bug tracker?):
http://groups.google.com/group/comp.lang.ruby/browse_thread/thread/102cad122462d930
Thanks,
Will R. wrote:
Do y’all have any advice regarding what I can do to help get bugs like
these fixed? Are my expectations regarding turnaround time just
unrealistic?
There are a very small number of people who have control over an open
source project like. The rest of us work on whatever we want and submit
patches. If you want something fixed in rails right away, the best way
is to fix it yourself and then submit a patch. Given the 12 core
members and thousands of tickets you cant expect them to get back to
everyone. Open source projects usually dont happen quickly.
So delve through some rails code and post some patches in the trac.
Alex W. wrote:
So delve through some rails code and post some patches in the trac.
I have already thoroughly investigated all three problems I linked to
and figured out how to fix them. Both the Rails tickets I linked to
have code for my fixes included in them. I guess my question was more
about how exactly to polish up the patches to reduce the effort required
to apply them as much as possible. Are additional unit tests required?
I’m guessing I should actually attach a diff against current rails svn
in the one bug where I just pasted a patch inline. If there is a
document out there titled “How to create correct Rails patches” then
please point me to it.
Wilson B. wrote:
- Clean patches against trunk that pass all the AR tests on all
platforms.
- Patches that are minimally intrusive
Once you’ve got that, you can throw an email at the Rails-Core list
and get some attention paid to your patches.
Thanks for the info, Wilson, that’s basically what I wanted to know.
Whenever I find some spare time (hah) I will try to clean this stuff up.
On 6/28/06, Will R. [email protected] wrote:
Do y’all have any advice regarding what I can do to help get bugs like
Thanks,
- Clean patches against trunk that pass all the AR tests on all
platforms.
- Patches that are minimally intrusive (For example, I think the
:order change in the first patch is a little scary, and may need an
extra test or two to make sure it doesn’t break anything for, say,
Oracle users.)
Once you’ve got that, you can throw an email at the Rails-Core list
and get some attention paid to your patches.
Another possibility here, at least for your first patch is that there
is a movement underway to move SQL generation into the connection
adapters:
http://wrath.rubyonrails.org/pipermail/rails-core/2006-May/001522.html
If that takes hold, it will be much easier to deal with funny
DBMS-specific stuff without breaking things.
Windows definitely plays second (or 4th) fiddle in Rails. To my
knowledge, the full ActiveRecord test suite doesn’t pass on Win32
using any database adapter.
I just tried it with MySQL, and got:
904 tests, 1911 assertions, 6 failures, 402 errors
Are they passing for you with SQL Server? If not, it’s hard to submit
a well-tested patch.
Bill W. wrote:
Alex W. wrote:
So delve through some rails code and post some patches in the trac.
+1
Bill, it’s clear you didn’t actually read the tickets I linked to, or
even read my second post in this thread. If you had, you would see that
I have delved through the code, and I have posted my patches. This is
really what this thread was about–if I have done those things, how do I
get someone to read the tickets and act on them?
Which is the question that Wilson answered, and I’m going to fix up my
tickets as he suggested. Thanks again, Wilson, for actually reading
what I posted instead of just making a snappy response.
On 6/28/06, Will R. [email protected] wrote:
I have filed a few bugs against Rails that have gotten no attention, and
asked a question on ruby-talk that never got answered (though I posted
via the newsgroup, maybe the gateway is one-way?). I don’t mean to
sound bitter, I’m aware that the problem is most likely that no one
can really test these problems.
Rails bugs:
I don’t believe any of the core team use SQL Server, so patches to the
adapter can take longer to apply than other areas of the system. In
recent months this process seems to be taking longer and longer, and I
know there is a small backlog building up of issues and patches
against the adapter.
I’m planning to sort through all these by the end of the weekend and
then prod someone with commit rights to get around to applying them.
I’ll try and make sure both your patches are included in this.
Tom
Tom W. wrote:
On 6/28/06, Will R. [email protected] wrote:
Rails bugs:
I’m planning to sort through all these by the end of the weekend and
then prod someone with commit rights to get around to applying them.
I’ll try and make sure both your patches are included in this.
Tom,
Thanks for your interest. I just posted an updated patch to ticket 4684
that applies cleanly against latest trunk. Applying this patch reduced
the number of errors in “rake test_sqlserver” from 19 to 4 for me.
Ticket 4650 is a bit more of a problem. The code there has changed
since I posted the ticket three months ago, and the changes I proposed
have essentially already been made. BUT, it still causes the same error
on SQL Server. See my new comment in that ticket for an explanation
why. It should be pretty clear why it’s not so obvious how to fix this
cleanly. If you can offer some help, I’d be grateful. I removed the
[PATCH] marker from this ticket for now.
I should also mention that there is what appears to be a related ticket
4850 here: http://dev.rubyonrails.org/ticket/4850.
Now that I am set up to run the ActiveRecord unit tests, I’ll try to
take a look at the other failures and see if I can be of any help on
them.