"Xtext is normally using a small number of XMI files".
Correction. Xtext/Xtend has an indexer that indexes far more than
is necessary and so builds far more than is necessary far more
often than is necessary. Faster XML parsing no doubt helps. Less
indexing/building would help even more.
Hi
It may depend on your workbench.xmi size. For me a 750kB file
could well benefit from a faster parser, but parsing is only one
of many startup activities. I see an about 4s improvement in 28s
loading from a RAM disk. Less clear whether many 1-5kB files
benefit since EMF has a parser pool to try to avoid churning
parsers.
Xtext is normally using a small number of XMI files of modest
size so I wouldn't expect a significant overall speed up.
However if you're using SysML with many large XMI files
involved, that could be a very different matter.
Regards
Ed Willink
On 24/01/2019 15:26, Mickael Istria
wrote:
Hi all,
Can some of you please give it a try as well? Just
add the suggested flag as -vmargs in eclipse.ini and
compare startup time or other "measurable" operations
with/without it.
If the results are generally positive, we should
think of a strategy to pre-set this flag (and some other
"factory" flags that have influence so early) in the
IDE.
Cheers
--
_______________________________________________
ide-dev mailing list
ide-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/ide-dev
_______________________________________________
ide-dev mailing list
ide-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://www.eclipse.org/mailman/listinfo/ide-dev
