On Mar 23, 2011, at 1:46 PM, Carl Friedberg wrote:

> I have these issues which I will research and report back to the group (all 
> with VMS 8.3 on alpha):
> 
> (1) my batch script generates 2 batch logs with the same name. One from 
> configure; the other from the main build and test script. Craig made some 
> changes to configure.com which should have fixed this;  I will dig in and 
> report back.

I'm not sure I recall that specifically.  Configure.com reopens SYS$OUTPUT and 
there were some changes to how that's done a year or two ago, so maybe that's 
covered.  Do report if not.

> (2) the tests hang on (all flavors of 5-12)
> 
> ext/Devel-SelfStubber/t/Devel-SelfStubber

I've never seen that, and I don't recall hearing it reported before.  Does it 
hang when run outside of the test suite?  If so, running it under the Perl 
debugger and seeing where exactly it hangs might be useful.

> I will try again with that test removed from manifest. (is that the 
> recommended way to skip a test?)
> 

Well, the easiest way is to edit the test script and put 

BEGIN { print "1..0 # SKIP: dang thing don't work\n"; exit 0; }

somewhere near the top of it.

> (3) I am still building with -Uuseperlio because of previous issues (since 
> 5-10) with extraneous line endings inserted when processing long lines ( 
> Martin Z. and Craig have worked that; apparently there's an extra line ending 
> every 4048 characters ). 

It was more that (in output to record-oriented files only) the implicit flush 
every 4K when an internal buffer filled up would give you line breaks in odd 
places.  What's fixed in 5.12.3 is that it now does an implicit flush whenever 
it encounters a newline and is writing to a record-oriented file.  This is a 
complete fix if you never have lines longer than 4K (4096 bytes).  In 5.14.0 
(due next month), the internal buffer size is also increased, so you should be 
good for lines up to 32K, which I believe is the RMS limit for record-oriented 
text files.


________________________________________
Craig A. Berry
mailto:craigbe...@mac.com

"... getting out of a sonnet is much more
 difficult than getting in."
                 Brad Leithauser

Reply via email to