I am starting to get busy characterizing and fixing Seaside3.1.1 issue #21[1]. The issue involves UTF8 performance and a fair amount of the UTF8 support is located in the Grease project.
Sooo, I am inclined to start using the Github-based Grease repository ... the packages are all synchronized with the packages on smalltalk hub (thanks Johan!). The upshot will be that Seaside3.1.1 will still use ConfigurationOfGrease, but just as Johan has done for ConfigurationOfSeaside3, the configuration will directly reference the github repository.
BTW, Johan, I notice that the reference in ConfgurationOfSeaside# that you did not point to a specific SHA, which means that folks loading Seaside3.1.1.1 could get different versions. I think it should be frozen at 9c53917253e0512a8a3ffd569b6d20fea8c5179e... what do you think?
Dale _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Hi Dale,
I defined a tag '3.1.1-gs31' for Seaside 3.1.1 in ConfigurationOfSeaside3-JohanBrichau.145 spec for: #'gs3.x' do: [ spec baseline: 'Seaside3' with: [ spec repository: 'github://glassdb/Seaside31:3.1.1-gs31/repository' ]. ]. So, I think that 3.1.1 is always loading that specific release, no? Note that I tried explicitly writing the ConfigurationOfSeaside such that Grease for Gemstone was fetched from the github repository, but I encountered some troubles. I think we exchanged some mails about that either on this list or the metacello group, but I cannot put my hands on them for now. I will keep looking. cheers Johan On 04 Jun 2014, at 20:38, Dale Henrichs <[hidden email]> wrote: > I am starting to get busy characterizing and fixing Seaside3.1.1 issue #21[1]. > > The issue involves UTF8 performance and a fair amount of the UTF8 support is located in the Grease project. > > Sooo, I am inclined to start using the Github-based Grease repository ... the packages are all synchronized with the packages on smalltalk hub (thanks Johan!). > > The upshot will be that Seaside3.1.1 will still use ConfigurationOfGrease, but just as Johan has done for ConfigurationOfSeaside3, the configuration will directly reference the github repository. > > BTW, Johan, I notice that the reference in ConfgurationOfSeaside# that you did not point to a specific SHA, which means that folks loading Seaside3.1.1.1 could get different versions. I think it should be frozen at 9c53917253e0512a8a3ffd569b6d20fea8c5179e... what do you think? > > Dale > > [1] https://github.com/glassdb/Seaside31/issues/21 > _______________________________________________ > Glass mailing list > [hidden email] > http://lists.gemtalksystems.com/mailman/listinfo/glass _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Johan, Yes, the tag is exactly the right thing to do ... I didn't look close enough at the url:) It will help if you find the thread so that I can refresh my memory ...
The good news is that you intend to head in the direction of using the github repo for grease as well ... I've branched both Grease and Seaside31 for my work on the utf8 issue ... also I intend to do the utf8 work for GemStone3.2 only ... the handling of Unicode* and *String instances is much more consistent (and correct) in 3.2 ...
Dale On Wed, Jun 4, 2014 at 2:10 PM, Johan Brichau <[hidden email]> wrote: Hi Dale, _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
In reply to this post by Dale Henrichs-3
On Wed, Jun 4, 2014 at 8:38 PM, Dale Henrichs
<[hidden email]> wrote: > I am starting to get busy characterizing and fixing Seaside3.1.1 issue > #21[1]. Producing a byte array is "the right thing to do ™". That Seaside currently expects is an unfortunate Squeak-ism that slipped into the design. I believe this has already popped up once (maybe for VW) but I can't remember what came out of this. There are a couple of things to consider: * Most strings that we render are HTML safe ASCII. The fastest way to render these is to do a quick test (we use CharacterSet on Pharo) and then copy them directly into the response buffer. You can make you own subclass of WAResponse to facilitate this. * All UTF-8 related tests should be in a custom packages. * URL rendering is also affected since we need to convert convert it to a byte array before escaping it and sending it through the HTML pipeline (and encoding it a second time) Cheers Philippe _______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Thanks, Philippe. I thought that this time around I was going to have to bite the bullet and convert the code to work on ByteArrays when dealing with encoded Strings, but (unfortunately?) I found a primitive-based scheme for fast conversion of an instance of Utf8 into a String (without decoding) and I can leave well-enough alone for the time being:)
Dale On Thu, Jun 5, 2014 at 1:22 AM, Philippe Marschall <[hidden email]> wrote:
_______________________________________________ Glass mailing list [hidden email] http://lists.gemtalksystems.com/mailman/listinfo/glass |
Free forum by Nabble | Edit this page |