Re: [PATCH 2/2] Don't rely on malloc to return NULL for zero size

2014-12-12 Thread Linus Torvalds
Ack. With our selfs corrected to ourselves in the comment :-) Linus On Dec 11, 2014 11:59 PM, Anton Lundin gla...@acc.umu.se wrote: We rely on samples being NULL if a dc have no samples. Its completely legal for malloc to return a valid pointer to nowhere for zero sized malloc, which you

bug hunt / trac day

2014-12-12 Thread Dirk Hohndel
Hi there, with 4.3 almost ready to go, can I ask for a favor? There are a TON of open bugs in trac. Many of them are for divecomputer support - those are hard to track down and usually only Jef can really answer them. But many other ones are things that need to be followed up on. Ask the

Re: bug hunt / trac day

2014-12-12 Thread Tomaz Canabrava
I'm trying hard to speed up the profile here. when we are creating a new dive, if we just move the handler around for a while, CPU goes to 100%. managed to reduce the amount of calls to QPainterPath::addText by 1/4, wich helped a lot, but it seems that Qt calls that internally so I'm now trying to

Re: bug hunt / trac day

2014-12-12 Thread Dirk Hohndel
On Fri, Dec 12, 2014 at 04:53:07PM +, Pedro Neves wrote: Hi all: On 12/12/2014 03:25 PM, Dirk Hohndel wrote: Play with printing (we seem to only check on that prior to a release). I found printing a bit strange. The profile is still rendered somewhat blurred and the spacing between

two patches

2014-12-12 Thread Tomaz Canabrava
Dirk, Take a look at 002 carefully, since I changed and added a few things to event manipulation. the idea of 0002 is to add / remove / rename an Bookmark without triggering an (expesive) redraw of the profile, so we create just the Event item and put it on screen. The issue is that the depth of

commit c76cb59

2014-12-12 Thread Joakim Bygdell
Dirk, I suggest that you reverse commit c76cb59. After your patch the line of the temperature graph sometimes gets drawn through the depth notation of the deepest point of the profile thereby obscuring the text. /Jocke ___ subsurface mailing list

Re: two patches

2014-12-12 Thread Dirk Hohndel
On Fri, Dec 12, 2014 at 03:53:28PM -0200, Tomaz Canabrava wrote: From cc64959903353c1d1cb4b182d50246d7fd1cd701 Mon Sep 17 00:00:00 2001 From: Tomaz Canabrava tomaz.canabr...@intel.com Date: Fri, 12 Dec 2014 14:06:04 -0200 Subject: [PATCH 2/2] Fixes the Bookmark Add/Rename/Remove witout

Re: bug hunt / trac day

2014-12-12 Thread Dirk Hohndel
On Fri, Dec 12, 2014 at 05:53:42PM +, Pedro Neves wrote: Of course... Debian/Sid. Qt 4.8.6, compiled from source on my system... Ah. Yes. Qt 4.x is known to have printing issues. That's why I went through all this work to have the majority of the binaries we provide (Win64, Mac, Ubuntu,

Re: commit c76cb59

2014-12-12 Thread Dirk Hohndel
On Fri, Dec 12, 2014 at 07:46:29PM +0100, Joakim Bygdell wrote: Dirk, I suggest that you reverse commit c76cb59. After your patch the line of the temperature graph sometimes gets drawn through the depth notation of the deepest point of the profile thereby obscuring the text. Can you provide

Re: two patches

2014-12-12 Thread Tomaz Canabrava
So I'm puzzled how often we are adding bookmark events and how that affects performance... or is this really just a first of a series and the others are more performance relevant? First of a series. diff --git a/dive.c b/dive.c index b318c4b..4fb6717 100644 --- a/dive.c +++

Re: commit c76cb59

2014-12-12 Thread Tomaz Canabrava
People, setZValue( -10 ) or something, in the offending item that's overlapping the one that shouldn't be overlapped. On Fri, Dec 12, 2014 at 5:00 PM, Dirk Hohndel d...@hohndel.org wrote: On Fri, Dec 12, 2014 at 07:46:29PM +0100, Joakim Bygdell wrote: Dirk, I suggest that you reverse commit

Re: two patches

2014-12-12 Thread Dirk Hohndel
On Fri, Dec 12, 2014 at 05:03:54PM -0200, Tomaz Canabrava wrote: So I'm puzzled how often we are adding bookmark events and how that affects performance... or is this really just a first of a series and the others are more performance relevant? First of a series. OK Dear C expert... I

Re: two patches

2014-12-12 Thread Tomaz Canabrava
On Fri, Dec 12, 2014 at 5:38 PM, Dirk Hohndel d...@hohndel.org wrote: On Fri, Dec 12, 2014 at 05:03:54PM -0200, Tomaz Canabrava wrote: So I'm puzzled how often we are adding bookmark events and how that affects performance... or is this really just a first of a series and the others

[PATCH 3/7] Use generated Ids for site lookup in UDDF export.

2014-12-12 Thread Martin Long
As with buddies, sites could contain characters which are not valid in Ids. Also, it is very possible to have duplicate site names. This uses an XSL generated id to prevent any issues. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 12 ++-- 1 file changed,

[PATCH 5/7] Various minor fixes to UDDF export

2014-12-12 Thread Martin Long
Removed underscore from buddy name elements Added 'mix' prefix to gas ids. Ids must not start with a number. Added mandatory profile data with empty tags. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 20 +++- 1 file changed, 11 insertions(+), 9

[PATCH 4/7] Fix divecomputer part of UDDF export

2014-12-12 Thread Martin Long
XPath was incorrect for parsing divecomputers into the equipment section, meaning they dont get inserted. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/xslt/uddf-export.xslt b/xslt/uddf-export.xslt

[PATCH 7/7] Fix bug in tankpressurebegin in UDDF export

2014-12-12 Thread Martin Long
There was a bug when the first sample doesn't contain pressure info. This fixes that by selecting the first with pressure info. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 29 + 1 file changed, 13 insertions(+), 16 deletions(-) diff

[PATCH 1/7] Update UDDF export generator header to include madatory tags.

2014-12-12 Thread Martin Long
Add name at generator level, and add additional homepage in contact tag. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 7 +-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/xslt/uddf-export.xslt b/xslt/uddf-export.xslt index 2f7ba70..7cdb2b2

[PATCH 2/7] Use generated ids for buddies in UDDF export.

2014-12-12 Thread Martin Long
This is instead of using their names, which may contain illegal characters. Signed-off-by: Martin Long mar...@longhome.co.uk --- xslt/uddf-export.xslt | 45 - 1 file changed, 24 insertions(+), 21 deletions(-) diff --git a/xslt/uddf-export.xslt

UDDF export - nearly there.

2014-12-12 Thread Long, Martin
So another 7 patches. This makes it almost 100% UDDF 3.2 compliant. I say almost. There are a couple of places where the schema doesn't agree with the documentation, where I believe the schema. I believe there are also some bugs in the schema too. In both cases I'll raise these with the authors.

Re: UDDF export - nearly there.

2014-12-12 Thread Dirk Hohndel
On Dec 12, 2014, at 4:43 PM, Long, Martin mar...@longhome.co.uk wrote: So another 7 patches. This makes it almost 100% UDDF 3.2 compliant. I say almost. There are a couple of places where the schema doesn't agree with the documentation, where I believe the schema. I believe there are

Re: [PATCH 4/7] Fix divecomputer part of UDDF export

2014-12-12 Thread Miika Turkia
On Sat, Dec 13, 2014 at 2:34 AM, Martin Long mar...@longhome.co.uk wrote: XPath was incorrect for parsing divecomputers into the equipment section, meaning they dont get inserted. Can you elaborate on this one as on my divelog all the dive computers are listed on the settings section and the