Scrap the last message, java/javadoc crashed!
Steve.
-----Original Message-----
From: Rankin, SE (Stephen) [mailto:[log in to unmask]]
Sent: 16 June 2003 14:18
To: [log in to unmask]
Subject: Javadoc speed
It would appear that it takes some time to get through the breakiterator
errors when compiling the javadocs. That last line of the output really
takes some time, wrote this and it is still doing it.
Steve
[javadoc]
/export/scratch/ussc/stardev-rh9/sources/starjava/source/splat/src/main/uk/a
c/starlink/splat/plot/PlotControl.java:1157: warning - The first sentence is
interpreted to be:
[javadoc] "Get the physical coordinates limits of the current view of the
Plot
[javadoc] (i.e."
[javadoc] This sentence is different from what would be generated using
-breakiterator:
[javadoc] "Get the physical coordinates limits of the current view of the
Plot
[javadoc] (i.e. what you can see)."
[javadoc]
/export/scratch/ussc/stardev-rh9/sources/starjava/source/splat/src/main/uk/a
c/starlink/splat/data/SpecData.java:420: warning - The first sentence is
interpreted to be:
[javadoc] "Get references to spectrum X data (i.e."
[javadoc] This sentence is different from what would be generated using
-breakiterator:
[javadoc] "Get references to spectrum X data (i.e. the coordinates as a
[javadoc] single array)."
[javadoc]
/export/scratch/ussc/stardev-rh9/sources/starjava/source/splat/src/main/uk/a
c/starlink/splat/data/SpecData.java:431: warning - The first sentence is
interpreted to be:
[javadoc] "Get references to spectrum Y data (i.e."
[javadoc] This sentence is different from what would be generated using
-breakiterator:
[javadoc] "Get references to spectrum Y data (i.e. the data values)."
[javadoc]
/export/scratch/ussc/stardev-rh9/sources/starjava/source/splat/src/main/uk/a
c/starlink/splat/data/SpecData.java:442: warning - The first sentence is
interpreted to be:
[javadoc] "Get references to spectrum Y data errors (i.e."
[javadoc] This sentence is different from what would be generated using
-breakiterator:
[javadoc] "Get references to spectrum Y data errors (i.e. the data
errors)."
-----Original Message-----
From: Rankin, SE (Stephen) [mailto:[log in to unmask]]
Sent: 16 June 2003 10:38
To: [log in to unmask]
Subject: Re: USSC218
Mark,
The nightly build has been using JDK 1.4.1_02 from the beginning, and on
April 17th the javadocs-full target reached 5 or more minutes and has been
the same since. Only just noticed this as you mentioned it, but the time
taken to build the javadocs has been steadily increasing over time,
javadocs-full.log
rev 1.11 - 55s
rev 1.12 - 1 min 50s
rev 1.14 - 2 min 33s
rev 1.15 - 3 min 54s
rev 1.18 - 5 min 23s
When I looked at the javadocs they appeared to be OK. The new JDK may do
more cross referencing; perhaps there are some flags on javadoc which may
speed things up?
Steve.
-----Original Message-----
From: Mark Taylor [mailto:[log in to unmask]]
Sent: 13 June 2003 16:59
To: [log in to unmask]
Subject: Re: USSC218
On Fri, 13 Jun 2003, Rankin, SE (Stephen) wrote:
> All,
>
> I have now branched USSC218. I will test and wrap it for the CD over the
> weekend, fingers crossed everything should be OK. Also, I have tried to
> enable Watches on the CVS repository so that you should be able to do
> something like
One thing I've noticed is that the export-full-docs target runs about
four times slower under java 1.4.1 than 1.4.0 (6 minutes vs 1min 20 for me).
I haven't investigated this in much detail, e.g. I don't know if
the output is much different, or whether it's any particular package(s)
which are causing it. This is a bit of a pain, but not a serious
problem.
PS: what is the significance of USSC218?
--
Mark Taylor Starlink Programmer Physics, Bristol University, UK
[log in to unmask] 0117 928 8776 http://www.star.bris.ac.uk/~mbt/
|