Alan W. Irwin wrote:
> Hi Hazen:
> Your comment below got me thinking about this new subject.
> On 2009-09-03 10:07-0400 Hazen Babcock wrote:
>> Since this [command-line parsing] issue has been around for years and I'm the
>> only one who seems to struggle with it I'm not inclined to delay the
>> release until it gets sorted out.
> On the whole, I prefer to stick to release dates once we set them, but in
> this case of four (!) issues that affect our premier devices it may be
> better for us to back off and release a week later instead depending on
> whether there is a good chance an extra week would allow us to sort out
> some of these issues. Let me know what you think about this.
> Here are those issues again:
> (1) The above command-line parsing issue. The valgrind investigations I
> suggested for this case may allow you to find the reason for this error and
> fix it in a relatively small amount of time. Do you have time to work on
> this before the weekend of September 12th/13th?
I can try, but since I am the only one who seems to run into this
problem and I have a pretty vintage OS-X box (32bit PowerPC, OS-X 10.4)
at this point, I don't attach much importance to it. Also, remember that
this only effects the xcairo and aqt drivers. I guess if I could figure
out how to pass command line arguments using gdb I'd be willing to try
that and see what happens.
> (2) plend/qt segfaults on some platforms. This one is still pretty open
> ended, but Alban appears to be actively investigating it so he might
> well benefit by a week's delay in the release.
> (3) plend/xcairo segfaults on all (?) platforms. Certainly, I see this
> issue on my platform. Hez, are you working on this one?
I thought that all the cairo drivers had this issue? I'd be really
surprised if we could solve this one in a week though given how long it
has been with us. My recollection is that problem goes away if we don't
use dynamic drivers or if we don't render any text, suggesting some
mixture of Pango, dynamics drivers and loading/unloading libltdl resources.
> (4) locale/qt issue which have not been fixed in the best way, yet, IMO.
> I believe we can find the right solution before September 12th/13th.
> We all can only work on PLplot part time so a day of actual work may easily
> stretch to several days of real time. So (4) is very likely to benefit from
> a week's delay in the release, and some or all of the rest of the issues
> might benefit as well depending on how much time our various developers have
> to track down and fix these issues in the extra time provided by a week's
> delay in the release. I don't think we should delay the release by much more
> than a week though since if a week's part-time effort doesn't solve these
> issues, then throwing more time at it may not help much either.
> Hazen, would releasing during the weekend of September 12/13th be convenient
> for you? If not or if you or others here don't feel a week's delay in the
> release would help to solve the above issues for our best devices, then
> despite those issues I would be willing to go along with the scheduled
> release this weekend (September 5th/6th).
Either weekend is fine with me.