Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 11 and week 12* I worked last couple of weeks on converting some of the contexts in sc/ as before. I spent most of my time looking to improve some performance in the codes we've written by avoiding unnecessary string allocations. Also, I did some refactoring of those codes to get better readability. *Patches:* https://gerrit.libreoffice.org/40791 https://gerrit.libreoffice.org/40846 https://gerrit.libreoffice.org/41063 https://gerrit.libreoffice.org/41364 https://gerrit.libreoffice.org/40928 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hello. I'm just a user, but very interested in the progress of LibreOffice as a platform. I sneak a bit the development stuff, trying to understand something and still keep my hope to someday learn computer programming. I'm going to try to be concise, but I can't resist to my need to satisfy my curiosity. Is this going to make LibreOffice faster in some way? Is it related in some way to compatibility? What about making file reading and writing more multithreaded? Kind regards. 2017-08-08 19:38 GMT+02:00 Mohammed Abdul Azeem: > Hi, > > Week 9 and week 10 > > I continued to work on converting legacy contexts in sc/ module to fast > contexts. We've converted almost all the legacy contexts in sc/ module, > except for few which are dependent on the contexts in xmloff. We would need > to convert contexts across all the other modules, before we start converting > the ones in xmloff. Now, I'm working on avoiding allocation of OUString for > attribute values wherever possible. > > Patches: > https://gerrit.libreoffice.org/40320 > https://gerrit.libreoffice.org/40326 > https://gerrit.libreoffice.org/40540 > https://gerrit.libreoffice.org/40553 > https://gerrit.libreoffice.org/40561 > > Regards, > Azeem > > ___ > LibreOffice mailing list > LibreOffice@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/libreoffice > ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 9 and week 10* I continued to work on converting legacy contexts in sc/ module to fast contexts. We've converted almost all the legacy contexts in sc/ module, except for few which are dependent on the contexts in xmloff. We would need to convert contexts across all the other modules, before we start converting the ones in xmloff. Now, I'm working on avoiding allocation of OUString for attribute values wherever possible. *Patches*: https://gerrit.libreoffice.org/40320 https://gerrit.libreoffice.org/40326 https://gerrit.libreoffice.org/40540 https://gerrit.libreoffice.org/40553 https://gerrit.libreoffice.org/40561 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 7 and week 8* We've continued to convert legacy contexts in sc/ module to fast contexts. I also fixed a bug tdf#108835 (Writer was freezing when adding autocorrect entry, due to synchronisation problems in package/ code). *Patches*: https://gerrit.libreoffice.org/39723 https://gerrit.libreoffice.org/39927 https://gerrit.libreoffice.org/40012 https://gerrit.libreoffice.org/40160 https://gerrit.libreoffice.org/40246 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 5 and week 6* We continued to work on converting legacy contexts to new contexts, along with some refactoring of code related to FastAttributeIter to make the code using them more readable. Besides, Michael helped me in solving the bug tdf#108821 (https://gerrit.libreoffice.org/39374). *Patches*: https://gerrit.libreoffice.org/39380 https://gerrit.libreoffice.org/39370 https://gerrit.libreoffice.org/39397 https://gerrit.libreoffice.org/39665 https://gerrit.libreoffice.org/39730 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 4* We worked on making FastAttributeIter of sax_fastparser::FastAttributeList cleaner and better, including some aesthetic changes. Majority of my time this week went in chasing tdf#108643 bug, which I realized was happening due to a change by Caolán McNamara and was later fixed by him, I believe. Also I've started working on adapting contexts from sc/source/filter/xml to work with fast parser, few at a time. *Patches*: https://gerrit.libreoffice.org/38881 https://gerrit.libreoffice.org/38964 https://gerrit.libreoffice.org/39275 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
Re: [GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 3* We've started using BufferedStream ( size < 10k ) or BufferedThreadedStream for unzipping instead of UnbufferedStream ( It resulted in a failure of some unit tests, which was fixed in https://gerrit.libreoffice.org/#/c/38661/ ). Also, I've started to adopt the string based contexts in sc/source/filter/xml/ to token based contexts. Also, we thought it's best to implement processingInstruction callback event for XFastDocumentHandler API, which makes it easier for this to take legacy APIs place. Also, I've added some static methods in SvXMLImport, that reverse maps fast tokens to Namespace and element name strings. *Patches*: https://gerrit.libreoffice.org/38661 https://gerrit.libreoffice.org/38708 https://gerrit.libreoffice.org/38784 Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice
[GSoC] Migrating from legacy Parser to FastParser - Weekly Report
Hi, *Week 1 and week 2* Michael Meeks and I have been working on moving the unzipping into new thread for over a month now. It took a couple of weeks to structure it right and then a couple more weeks to debug our way through the bugs. Meanwhile we discovered that there were some synchronisation problems in package/ code which was causing intermittent stream open failure and we've fixed it now. *Patches*: https://gerrit.libreoffice.org/#/c/38135/ https://gerrit.libreoffice.org/#/c/38563/ Regards, Azeem ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/libreoffice