Philip Martin wrote:
>Branko Čibej <email@example.com> writes:
>>Jay, thanks for your analysis, it's extremely valuable!
>>Which leads to a short sermon: In the past few weeks a lot of new
>>stuff has gone into the code, mostly in the shape of fixes for
>>bite-sized tasks. In hindsight, most of those changes were poorly
>>thought out, at least as far as portability is concerned. Both
>>copy-with-permissions and readonly admin files have managed to hose
>>the Win32 build. While the file copy part is arguably an APR concern,
>>the read-only file part is not. I think we have to be more careful
>>about what goes in the code, otherwise we'll have a 1.0 that works on
>>(some flavours of) Unix and nowhere else.
>>Although I do try to keep up with the proposed patches, the sad fact
>>is that one pair of eyes and a limited amount of time is simply not
>>enough. That's why I find Jay's analysis of this particular problem so
>>valuable. It's also a bit discouraging that people seem to overlook
>>the broad picture so easily and think in terms of Unixy semantics
>>instead. QED, that won't work for long.
>>I do understand that thinking beyond "worksforme" makes it harder for
>>many people to contribute, and I'm not criticising these contributions
>>themselves. But I'm starting to see a trend here, and the closer we
>>are to 1.0 in terms of functionality, the worse it'll get, unless we
>>do something about it.
>Expecting Unix developers to provide Win32 support just by thinking
>about it is not going to work. We need input from Win32 developers.
>Consider the read-only problem (which prompted the sermon) this was
>first raised last year. Since then there has been no Win32 code
>proposed to provide it. Not a single line. We had another discussion
>last month, and the Win32 issue was raised again. I asked for a
>possible Win32 interface (not an implementation, just an interface)
>but nothing appeared. I asked for a description of the Win32
>filesystem behaviour, still nothing. So in the absence of any other
>input I proposed an interface, nobody said it was not suitable. In
>fact, the comment I got was "this seems the way to go".
>Now it is quite possible that the interface won't work on Win32. I
>don't know, I'm not a Win32 developer. Perhaps the interface is OK and
>it's the application code that needs to change, again it needs a Win32
>person to point out the problems. Even better would be to provide a
>patch, but describing the problem will do. If a problem is identified
>I will try to fix it, but references to "sheer sloppy programming" do
>not encourage me. The changes in question did not come out of the
>blue, they were posted to the list precisely because I suspected that
>there could be platform issues.
Pleas note that I did say I'm *not* criticising contributions (or
contributors). In fact, since I reviewed and o.k.'d your patch, part of
the blame for the current state of affairs is definitely mine.
As to sloppy programming ... sure, I should've put that differently.
Write it off to the bad mood of someone who has spent far too many
sleepless nights trying to get the dam' thing working on Win32, and
hasn't the good sense to calm down before posting. (Damn, it's past
The problem, as usual, is not people but processes. In "doing something
about it" we shouldn't ask people who contribute patches to test on more
than one platform; we /should/, however try to make sure a change does
get tested on as broad a set of platforms as we can achieve. Note I said
"change", not "contribution". This will get more and more important the
closer we are to 1.0, because the cost of errors gets higher.
>Within days of my starting to use apr_file_attrs_set an APR Win32
>patch appears, so there are Win32 developers interested in Subversion.
>I guess that Subversion currently has more developers using Unix than
>Win32. I don't know how to encourage people with Win32 know-how to
>contribute, but I wish they would. Until they do Win32 support is
>inevitably going to be a second class citizen. I don't like that, I
>want Win32 support (and OS X support, 64-bit support, etc.) However
>it won't just happen, it requires appropriate developer interest.
I'm afraid developer interest is not enough. Our goal for 1.0 is to have
a replacement for CVS. From my point of view that doesn't involve just
functionality; it means 1.0 has to run on at least the 10 most popular
platforms CVS runs on. It's not just Win32, I'm pretty sure we have
problems right now on popular Unixes, too.
In short: we have to define a set of target platforms for 1.0, and try
to find machines to test on. I can help with Win32, HP-UX and Solaris,
maybe AIX with a bit of luck.
Brane Čibej <brane_at_xbc.nu> http://www.xbc.nu/brane/
To unsubscribe, e-mail: firstname.lastname@example.org
For additional commands, e-mail: email@example.com
Received on Sat Oct 21 14:37:06 2006