[impression-bugs] Compile Contents vs too much memory
- From: Jim Nagel <jim-imp@xxxxxxxxxxxxxxxx>
- To: impression-bugs@xxxxxxxxxxxxx
- Date: Thu, 24 Oct 2013 11:40:33 +0100
The "Utilities > Compile contents" routine has exhibited funny
behaviour for years, and I might at last have discovered something.
In doing the contents list for the cover of Archive 23:7, with
Publisher Plus (5.13) running on the Armini (Ro 5.19 with Aemulor),
the utility did its usual thing of aborting partway through the job,
leaving a single long line of text in the target page (which is, as
usual, a new chapter at the top of the document).
Typing a Return anywhere in that line often causes the whole page to
format itself correctly. But why it should do this and why the
routine stops at page X (which might be different upon re-running the
routine, might even sometimes be the whole document) remains a
mystery.
Armini eventually crashed after repeated attempts to compile contents.
Tried it on the RiscPC instead of the Armini. Similar behaviour,
though RiscPC seems to get to an X greater than the Armini's X.
Then a bright idea: Amnesia! This is a little utility by Nemo that
simply grabs a huge dynamic area in order to reduce the amount of ram
available to applications -- often useful for old apps that were
written in the days of max 4M ram.
With Amnesia, the contents compiled perfectly on the RiscPC.
(I think I let Amnesia grab too much ram at first -- dunno its logic,
no way of configuring it -- but I got round that by making a ramdisc
of a couple of megabytes before invoking Amnesia, then quitting the
ramdisc.)
No luck on the Armini, though -- could be that Amnesia doesn't work
with Ro 5. Haven't yet tried on the Iyonix (5.18 with Aemulor 2.32).
Could be that the compile-contents problem manifests only with very
large files (like the Archive document, 7.8M, 18 internal chapters,
115 stories, 115 pictures) -- but then of course large documents are
the ones where a Contents list is wanted!
Can do further tests or let you have sample documents if you wish,
Richard, but I rather suspect it's something in the code that is
wrongly presuming less memory to be available. Maybe Aemulor is a
further complication.
--
Jim Nagel www.archivemag.co.uk
Abbey Press 32 Norbins Rd (01458) 83 3603
Glastonbury BA6 9JG pocket 0797 415 3861
See you at the London show? www.riscoslondonshow.co.uk Oct 26
Other related posts:
- » [impression-bugs] Compile Contents vs too much memory - Jim Nagel
