[Libreoffice-commits] core.git: Changes to 'private/matus/fastparser'

[Matúš Kukan]

New branch 'private/matus/fastparser' available with the following commits:
commit 2394b86994de9e4d4e6028603c596bb020457f31
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Wed Oct 16 19:58:36 2013 +0200

    fastparser: don't use multithreading for small documents
    
    Determined by XInputStream::available().
    
    Change-Id: I450f4796d9c072b395393582bfc3e1e7768e243b

commit 1469c6753569db2a1930c572372a33ba1ab6f477
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 15 14:43:24 2013 +0200

    fastparser: don't create temporary Events; use references to event list
    
    Change-Id: I1e12fbeeb90d6020d0566d05fc0318082e1da5fc

commit ed828774af8dd3918b4b6d7981095444ec8a82b4
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 15 11:39:04 2013 +0200

    fastparser: fix leak by storing rtl::Reference instead of pointer
    
    Change-Id: I003dc48fed9029c9af430d4d2d361425d351ff54

commit 32a080a1bd85b4c0386b4db7630ece0a60b5ac0d
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 15 10:32:55 2013 +0200

    fastparser: re-use event lists if possible
    
    Change-Id: I53800abaca51d42d7d44a98fb271de7df7f90f58

commit 283027d661ccc5d0c2fb44d1b76cd355d586e067
Author: Michael Meeks <michael.meeks at collabora.com>
Date:   Fri Oct 11 14:09:52 2013 +0100

    fastparser: re-work locking, add high & low watermarks, change sizes etc.
    
    Change-Id: I7fe1435addc6dce5a74a8411f7825cea331a5b3f

commit f80cc8bb1a3f1dc1743f555a1fea21cb85979550
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 8 23:28:32 2013 +0200

    FastSaxParser: implementation using two threads
    
    Instead of calling methods directly, generate EventList - vector of
    Events, where arguments for the callee are stored.

commit 91ffcb4f97ce553dc046b25ac91bba2f50a276a4
Author: Michael Meeks <michael.meeks at collabora.com>
Date:   Fri Oct 11 21:46:45 2013 +0100

    fastparser: cache default namespace token for ooxml.
    
    Change-Id: Iee98ec92380d6d0404ab236e062ddbc2378cda43

commit c6a4ada8e3ebd3c6444c743fa89cf7108acac9ac
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Wed Oct 16 11:15:31 2013 +0200

    fastparser: don't use OStrings in attribute list, just simple char buffer
    
    Change-Id: I4879563fae3b85c68bbd1c4b260f9833848f4bda

commit ffa64a3188ab033a09f0d632b53b290f105af887
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Fri Oct 11 13:11:44 2013 +0200

    FastSaxParser: remove duplicated OString version methods
    
    Also remove AttributeData and make the two for-cycles faster.
    
    Change-Id: If0343992173bb333eee39c33bcb86acaa5a3d288

commit d665e72b5347ab3f2ef5c159a420aee3bf7b629f
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 8 15:17:01 2013 +0200

    FastSaxParser: isolate calls to XFastContextHandler in Entity's methods
    
    Unfortunately, we have to store maNamespace in two stacks.
    One for future parser thread to compute tokens and one for
    main thread calling XFastContextHandler.
    
    Now we are prepared to implement multithreading.
    
    Change-Id: I421b55afa2e9ea80142e6068c7a515d31913ba69

commit 361292aa8fef70ef3346922091209a82c95f497b
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 8 12:42:21 2013 +0200

    FastSaxParser: store mnNamespaceCount in another stack
    
    This is preparation work for multithreading.
    mnNamespaceCount will be handled in parser thread and the rest in main
    thread.
    
    Change-Id: I571026ea499f6876b8dafb4e1bdc56d1add649e5

commit 0be2ea5dfc3e31c5eb1d7b8ba4b6c546fde56f2b
Author: Matúš Kukan <matus.kukan at gmail.com>
Date:   Tue Oct 15 09:29:27 2013 +0200

    FastAttributeList: use vectors instead of map; the size is small
    
    Change-Id: If0ea36155d8ab3f5c91c2aafd6932fabeadadd41

commit 0525b6af7749eaef1f71b19f37629c64a04c096c
Author: Michael Meeks <michael.meeks at collabora.com>
Date:   Fri Oct 11 13:51:47 2013 +0100

    fastparser: don't allocate uno::Sequences when we don't need to.
    
    Change-Id: Ic2fff8cabbc077b6fc9dabffd2c6fcf555152b11

commit 624706280b5593c0ac3ed84b667e0ea80b69440e
Author: Michael Meeks <michael.meeks at collabora.com>
Date:   Fri Oct 11 22:51:44 2013 +0100

    oox: special-case single-character a-z token mapping.
    
    Change-Id: I58a810cc6062d5b42558dd5c0f37426a8a210f40